Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2007-11-18

00:00:35amiconnHow does that look, a fade from e.g. backlight level 3 to off? Not like fading at all, I guess
00:01:04kugelI don't know. I have backlight at level 7
00:01:05marazoblib: yeah, it should. Your original 0x10000000 patch + the same patch applied to the ata code
00:01:33oblibmaraz, can you double check? Just type 'svn diff' in the root svn directory
00:01:50marazsure
00:02:51kugelamiconn: Let's say you commit this patch. Would it hurt? It's just a small patch, it doesn't increase the binary size much.
00:02:57marazoblib: yeah, only those lines
00:03:28Lloreankugel: Are you just going to ignore the other problems with it?
00:03:42 Join mirak [0] (n=mirak@ip-21.net-89-2-71.rev.numericable.fr)
00:03:43 Join _novocaine_ [0] (n=kevin@ool-43548ef6.dyn.optonline.net)
00:03:44kugelNo, I'm just asking
00:03:46amiconntierra: amiconn.dyndns.org/rockbox-mini-lcd1.zip">http://amiconn.dyndns.org/rockbox-mini-lcd1.zip Hope you don't mind that this contains the .mod patch, plus 2 more test plugins than the one we'll need...
00:04:11Lloreankugel: Again, why don't you just go and find the people who maintain or work on the patch, and try to talk them into improving the problems other Rockbox core devs have commented on it.
00:04:37amiconnAfter installing it, go to System->Debug->View I/O Ports, and tell me the value for LCD1_CONTROL
00:04:45tierrano problem, I've got my original backed up
00:05:37 Quit J3TC- (Read error: 104 (Connection reset by peer))
00:06:09 Join J3TC- [0] (n=jetc123@pool-71-125-77-210.nwrknj.east.verizon.net)
00:06:41kugelamiconn: Was that directed to me?
00:06:50tierraamiconn: LCD1_CONTROL = 684
00:06:55tierrakugel: na
00:06:57kugelok
00:07:05amiconn684, not 4687?
00:07:09amiconnInteresting...
00:07:25tierrayeah
00:07:47marazoblib: maraz.be/rockbox-ipodnano-r15654M-071117-IDE0patch.diff.txt">http://maraz.be/rockbox-ipodnano-r15654M-071117-IDE0patch.diff.txt
00:07:48*amiconn makes notes
00:08:09 Quit zanzabar ("tribe..http://www.tribe.org.uk/t9i.zip")
00:08:43amiconnIt makes some sense though. There are only some bits changes which have no observable effect
00:08:55amiconnAnyway, I'll take this into account for the next tests
00:09:10amiconnNow, I need to know the lcd update speed at 30 and 80 MHz, as a reference
00:09:31amiconnFirst, go to plugins->applications, and run test_fps
00:09:40 Quit mf0102 ("Verlassend")
00:09:56oblibmaraz, can you test the .zip file I just posted to the bug report and let me know if it works?
00:09:57amiconnThis should display 2 fps values (full screen and 1/4 screen fps) and the current cpu frequency
00:10:13tierra1/1 = 64.5
00:10:22tierra1/4 = 225.5
00:10:29tierra30 MHz
00:11:07amiconnOkay. Now go to System->Debug->CPU Frequency, and turn the wheel one step counter-clockwise, so that the boost count becomes 1 and the cpu clock becomes 80MHz
00:11:22amiconnThen go back and re-run the test_fps plugin
00:12:12tierra1/1 = 172.0
00:12:19tierra1/4 = 601.5
00:12:25tierra@ 80 MHz
00:12:31amiconnok.
00:12:52kugelWhat are you testing?
00:13:00amiconnYou can un-boost the cpu now. I need to prepare the next test build (just a new rockbox.ipod this time)
00:13:40tierrak
00:14:40marazoblib: seems to be working okay
00:15:05*amiconn wonders why the mini 1st gen is slower than the (full size) 2nd gen ipod wrt lcd update, even though it has the same bridge and less pixels
00:15:14 Join Presto1 [0] (n=root@166-82-130-132.quickclick.ctc.net)
00:15:33Presto1sup
00:15:45oblibamiconn, what are you going to do about committing this fix. maraz's bug is now resolved
00:15:49amiconntierra: amiconn.dyndns.org/rockbox.ipod">http://amiconn.dyndns.org/rockbox.ipod
00:16:09amiconnJust put that on your mini into the .rockbox dir and reboot
00:16:14marazoblib: it'll probably fix FS #7510 completely
00:16:24 Part Presto1
00:16:38amiconnThen re-check LCD1_CONTROL (should be 0x00000484 now), and re-run the speed tests at 30 and 80 MHz
00:16:42kugelSo guys, you wouldn't commit a lcd fade on e200, even if it's thread-safe and noone complaints? Just because e200 has only 12 brightness levels?
00:17:10tierrayep, 484 now
00:17:19tierraone min on tests
00:17:22Lloreankugel: Once again, you seem to be jumping to conclusions.
00:17:33kugelAgain, I'm asking
00:17:34 Quit freqmod_nx (Read error: 104 (Connection reset by peer))
00:17:41oblibkugel, why don't you do your own patch?
00:17:48 Join freqmod_nx [0] (i=freqmod@dhcp208-90.ed.ntnu.no)
00:17:49LloreanIt can't be committed now, either way.
00:18:01LloreanSo why not work on fixing it until someone closes it as "Rejected" or accepts it.
00:18:20tierra30MHz: 1/1 = 94.5, 1/4 = 328.5
00:18:33LloreanOnce we're *sure* we won't commit a patch, it gets rejected. The fact that it's still there pretty clearly says "it's not final that it won't get included"
00:18:44 Quit freqmod_nx (Client Quit)
00:18:50kugelI'm asking in general, if the 12 levels in brightness prevent *any* lcd fade solution from getting committed
00:18:56 Join freqmod_nx [0] (i=freqmod@dhcp208-90.ed.ntnu.no)
00:19:05 Quit mirak ("Ex-Chat")
00:19:14tierra80MHz: 1/1 = 252.5, 1/4 = 877.0
00:19:22amiconnkugel: Is that fade configurable?
00:19:24 Quit freqmod_nx (Read error: 104 (Connection reset by peer))
00:19:30 Join mirak [0] (n=mirak@ip-21.net-89-2-71.rev.numericable.fr)
00:19:34 Join freqmod_nx [0] (i=freqmod@dhcp208-90.ed.ntnu.no)
00:19:44kugelI'm not talking about any specific fade right now
00:20:05kugel(the patch I mentioned early isn't configurable)
00:21:14amiconntierra: Ok, thanks: Next one uploaded: amiconn.dyndns.org/rockbox2.ipod">http://amiconn.dyndns.org/rockbox2.ipod
00:21:38amiconnRename after download (the changed name is just to make sure we won't see nasty browser caching issues)
00:21:52 Quit Arathis ("Bye, bye")
00:21:57amiconnLCD1_CONTROL should be 0x284 now, and it should be faster again
00:22:01oblibmaraz, Well I'm done. Glad to have that fixed.
00:22:09marazMe too.
00:22:32 Quit oblib ("Leaving")
00:22:35tierraLCD1_CONTROL = 284
00:23:11tierra30MHz: 1/1 = 114.5, 1/4 = 400.0
00:23:14 Join krazykit [0] (n=krazykit@adsl-76-240-200-149.dsl.ipltin.sbcglobal.net)
00:23:52tierra80MHz: 1/1 = 304.5, 1/4 = 1067.5
00:24:14amiconnwow, already quite nice speedup :)
00:24:22tierrahehe
00:24:42 Quit mirak (Read error: 104 (Connection reset by peer))
00:25:05kugelCan someone answer my question?
00:25:36amiconnBut there's one more level: amiconn.dyndns.org/rockbox3.ipod">http://amiconn.dyndns.org/rockbox3.ipod
00:25:55amiconnLCD1_CONTROL should be just 0x84 now, and it should be faster again
00:27:01amiconnDo the same tests, and after that keep the CPU at 80MHz and try various graphical intensive plugins, looking for (possible, but not expected so far) graphical glitches
00:27:39tierraLCD1 = 84 indeed
00:27:46amiconnThis would be an indication that we're driving the lcd controller too fast (it's not dangerous as a test)
00:28:04Lloreankugel: I think I rather clearly told you that the fact that the task isn't yet rejected says rather clearly "it hasn't been decided yet"
00:28:21tierra30MHz: 1/1 = 215.5, 1/4 = 741.0
00:28:33scorchejees..
00:28:46amiconn!
00:28:58kugelWasn't clear to me, but thanks. I just asked since amiconn said "drop the idea, there are only 12 levels".
00:29:03amiconnmini 1st gen sees about the same speedup as c200
00:29:12tierra80MHz: 1/1 = 575.0, 1/4 = 1978.5
00:29:27Lloreankugel: I said "Once we're *sure* we won't commit a patch, it gets rejected. The fact that it's still there pretty clearly says "it's not final that it won't get included"", what part of that was particularly confusing?
00:29:57amiconnThat means while the bridge is the same, it doesn't behave exactly the same on PP5002 and PP5020
00:30:33amiconntierra: Any pixel glitches at 80MHz, or does everything look like it should?
00:30:34tierrano glitches in screen so far as I can see
00:30:34kugelI won't answer that, I stated what confused me
00:30:44amiconnniice :)
00:30:46Lloreankugel: No, you really didn't.
00:31:12kugel"I just asked since amiconn said "drop the idea, there are only 12 levels"."
00:31:18Lloreankugel: You claimed afterward that you weren't talking about any specific patch, but rather the idea that 12 steps would get a patch rejected. I think it's rather clear it *didn't* get that patch rejected, right?
00:31:45amiconntierra: Thank you very much for testing. Looks like this great speedup can be committed safely for all targets using that lcd bridge.
00:32:02amiconnYou could try mpegplayer btw. Audio should not stutter with that patch
00:32:11Lloreankugel: Your question was "Will only having 12 steps keep a patch from inclusion" and I responded by saying "The patch that only has 12 steps hasn't been rejected, and we reject them if they definitely won't be included."
00:32:23tierrathe mandlebrot demo does some flicker on certain shades... not sure if that's normal
00:32:25kugelI got it.
00:32:32kugelYou can stop yelling at me now
00:32:42scorchetierra: mandlebrot is more CPU than graphics intensive
00:32:47amiconntierra: It is, due to the way the grayscale lib works
00:33:09tierraugh, how do I exit mandlebrot?
00:33:24amiconnSelect+Menu, like many plugins
00:33:44amiconnNothing to worry about, the little flicker is the little price to pay for being able to display more shades of grey than the lcd normally allows
00:34:05tierrahehe, anyway, everything else seems great
00:34:22linuxstbamiconn: Did you measure the framerate you get on your mini with mpegplayer?
00:34:28amiconnI have an idea for an all-new grayscale lib with even more shades
00:35:02amiconnlinuxstb: I can't, as the targets which use the grayscale lib cannot display the fps in parallel. And mpegplayer has no statistics screen...
00:35:27linuxstbHmm....
00:36:03linuxstbHow does it look though? Is it anywhere near full speed?
00:36:14amiconnfaster...
00:36:31linuxstbThat makes sense - no yuv conversion to do...
00:36:50amiconnIt already was quite fast without those controller speedups, but the audio stuttered
00:37:27 Quit midkay ("Leaving")
00:37:36amiconn..because both the grayscale lib and audio decoding run on the same core, and the grayscale lib draws quite some cpu power with its timer interrupt if the lcd update is slow
00:38:17 Join oblib [0] (n=oblib@111-231-8-204.erd.cust.wirelessbeehive.com)
00:38:21amiconnWith this speedup, I can even remove the mandatory boost on PP in the grayscale lib...
00:39:12linuxstbWith all your recent work, Rockbox is starting to look quite respectable on PP... :)
00:39:18 Quit davina (Remote closed the connection)
00:39:33amiconnThe mini 1st gen sees a 3.33x speedup :)
00:40:08amiconnYeah. Now if I knew what that type0 lcd controller in the color is...
00:40:12oblibDoes anyone know why the Nano doesn't have CHARGING_CONFIG set yet? I did a patch months ago (6940) that enabled car mode, and it hasn't been committed yet. I find it useful, just thought I'd share...
00:40:55amiconnI'll do a (very) few more tests on pixelma's c200 tomorrow, then commit that speedup
00:41:25tierraamiconn: sorry, just noticed your comment on mpegplayer, was there something you wanted me to try there?
00:41:36amiconnThe target with the fastest (fps-wise) lcd update is still the M5 though
00:41:50amiconntierra: Did you try it some time before?
00:41:59tierrana, never touched it
00:42:04amiconnokay
00:42:11 Quit oblib (Client Quit)
00:42:30amiconnIf you did, you would probably have noticed that audio stutters. With the patched build it no loner should
00:42:48amiconn*longer
00:43:02tierraah, good to know either way :)
00:43:02 Quit scorche (" HydraIRC -> http://www.hydrairc.com <- The professional IRC Client :D")
00:46:58linuxstbtierra: You can download a sample video here - http://download.rockbox.org/mpeg/ (choose one that matches your LCD resolution)
00:49:36Bagderhttp://daniel.haxx.se/blog/2007/11/18/free-to-use-compiler-from-ti/
00:50:49eigmaBadger: I've installed it on my server.. not using it yet, but hopefully will soon :)
00:51:06BagderI've installed it as well just to have a look ;-)
00:51:30eigmathink it's worth integrating into the rockbox build system?
00:52:40BagderI'm not sure...
00:52:49linuxstbBagder: Do you understand the license?
00:52:51*eigma will be back in 10 minutes
00:52:56 Quit merbanan (Remote closed the connection)
00:53:11Bagderlinuxstb: it seems to be pretty clear about OK to use for open source projects
00:53:24Bagderso yes, it looks fine to me
00:53:40linuxstbIt does? That was the strange part for me, especially their definition of open source.
00:54:47Bagderwhat in particular makes you think that?
00:58:08linuxstbI'm not saying we can't use it, just that I found the license text odd. e.g. from 2c - "Open Source Software" means any software licensed under terms requiring that other software combined or distributed with such software: (i) be disclosed or distributed in source code form; (ii) be licensed on terms inconsistent with the terms of this Agreement.
00:58:51Bagderright, I don't understand the (ii) point...
00:59:29linuxstbIt's also the "distributed with" - I don't know any open source licenses that try to impose limitations on what you can distribute with it.
01:00
01:00:01 Join atc200es-away [0] (n=Delorean@70-101-164-164.dsl2.mon.ny.frontiernet.net)
01:00:15Bagderyeah, but this license is for distributing this package, as I understand it. Thus it gets distributed with something else.
01:00:19atc200es-awayhas there been any recent change in the code that would affect wps files?
01:00:37 Quit Robin0800 (" HydraIRC -> http://www.hydrairc.com <- \o/")
01:00:54 Nick atc200es-away is now known as delorean90 (n=Delorean@70-101-164-164.dsl2.mon.ny.frontiernet.net)
01:01:13linuxstbBagder: That's also what I'm not sure about - that only seems to give us the right to distribute it with open source software, not distribute it by itself...
01:01:31 Quit PaulPosition ()
01:01:37linuxstbAnd although the blog mentioned distributing the original installer, the license doesn't...
01:01:41Bagderyou're right... we have to make a little open source code to include if we want to distribute it ;-)
01:02:16*DerPapst suggest hello world :)
01:02:37delorean90i have a custom made theme i use, its even hosted on rockbox-themes.org but recently i did a format because of hard drive issues, i instal the lastest build put on the theme and it wont load it will only show the default wps and its quite nerve racking, i have checked my wps file made sure everything is ok, i cant seem to explain it....
01:02:48linuxstbThe other clause was related to only using it to develop code to run on TI devices....
01:03:13linuxstb"You may use the Licensed Materials to generate applications that execute solely and exclusively with systems containing processing devices manufactured by or for TI."
01:03:34linuxstbBut I guess that's OK...
01:03:36Bagderwell, c54x is likely to be TI's devices
01:03:47marazdelorean90: which ipod do you have?
01:03:53Bagderbut it means it can't be emulated
01:03:58Bagderwithout breaking the license
01:03:59delorean90i actually have an iriver h10 20gb
01:04:05marazoh, okay
01:04:20delorean90i have used rockbox for a year and this has never happened to me
01:05:02delorean90hmm, just been very frustrating, first i couldnt copy the latest build, seems my hard drive had currupted files, now i cant get it back the way it was
01:05:03PaulJamdelorean90: how old was your previous bild? and was it the official build or some custom one?
01:05:10DerPapstdelorean90: depends on your definition of recent. since a week albumart is in SVN and some kind of margin patch
01:05:21delorean90i always use offical build, hmm id say a week or 2 old
01:05:49delorean90would that effect a wps that dosent use patches?
01:05:57linuxstbWhat theme is it?
01:06:03Bagder"You also agree that you will use your best efforts to prevent your employees and contractors from unlocking, decompiling, reverse engineering, disassembling, modifying or translating the Licensed Materials."
01:06:34linuxstbI'll tell my employees that...
01:06:36delorean90Nordic, which i belive has an issue in the .cfg file that needs to be corrected, which i have already done on on my dap
01:06:37PaulJamhmm, a theme that workes 2 weeks ago with the official build should work with the cuttent too...
01:06:51 Quit courtc (Read error: 110 (Connection timed out))
01:06:59 Join Rob222241 [0] (n=Miranda@p54B15C76.dip.t-dialin.net)
01:07:06delorean90yes, its very confusing
01:07:15PaulJameither i can't type or my keyboard is broken.
01:07:37linuxstbBagder: "translating" ? That seems a bit harsh, even for TI...
01:07:38delorean90ill try putting on another theme from the wpsgallery and see if that works
01:08:00Bagderhehe, indeed
01:09:04linuxstbSo if a non-English speaker asked you what the compiler error message meant, you couldn't tell them...
01:09:09***Saving seen data "./dancer.seen"
01:10:02Bagderlinuxstb: translate could even be just rephrasing something from english into different english...
01:10:41Bagderbut I figure they really are more into translating as in disassemblying etc
01:12:35 Quit atsea- (Remote closed the connection)
01:13:00linuxstbBut going back to distribution, I guess putting it on the Rockbox download server might count as distribution with open source software.
01:13:19 Join daurnimator [0] (n=daurnima@unaffiliated/daurnimator)
01:13:25PaulJamdelorean90: there is an error in the wps, line 17: %?ps<%xdn>%ar%?mh<%xd|%xdI>
01:13:33Bagderlinuxstb: I agree
01:13:35linuxstbAlthough a tarball would be much friendlier than the gui installer...
01:13:51linuxstbs/gui/binary/
01:13:55Bagderyes
01:14:06PaulJamdelorean90: the %xd near the end seems to cause an error.
01:14:08Bagdernot a single file has the target path in them after install anyway
01:14:09 Join saratoga [0] (i=9803c6dd@gateway/web/cgi-irc/labb.contactor.se/x-8d1087384375ceae)
01:14:14Bagderso I don't think any was modified
01:14:15eigmaI hate licenses so much.. all this wording and nitpicking is completely against the spirit of productivity
01:14:58amiconnlinuxstb: Hmm, am I remembering correctly that this TI toolchain would also allow building stuff for archos jbmm, AV1xx etc?
01:15:07BagderI guess a tarball would be considered a translation...
01:15:13 Quit Rondom ("Ex-Chat")
01:15:16Bagderamiconn: yes
01:15:40delorean90@pauljam, in the nordic wps?
01:15:58linuxstbBagder: Also, the original blog article mentioned that TI requested that only the binary installer be distributed.
01:16:08amiconnjbmm could be fun still. Relatively slow arm core, plus ti dsp, plus mas...
01:16:18amiconni.e. mixedcodec
01:16:24Bagderlinuxstb: yes, but that seems consistent with the license, only spelled out more clearly
01:17:05PaulJamdelorean90: yes. i think the whole part %ar%?mh<%xd|%xdI> can be removed, there are no images for HOLD anyway.
01:17:20delorean90oh ok thank you ill try that
01:17:33 Quit Buschel_ ()
01:18:11Bagderoh there's source in there too
01:18:17delorean90i wonder why this issue hasnt arisen before though....
01:18:17 Join Xerion [0] (n=xerion@cp198589-d.landg1.lb.home.nl)
01:19:26PaulJamdon't know, maybe the error checking became stricter.
01:19:32delorean90oh wow that did it thanks so much
01:19:43Bagderhah
01:19:48Bagder"Edison Design Group C++ Runtime
01:19:52BagderCopyright 1993-2004 Edison Design Group, Inc.
01:19:59Bagder:-)
01:20:01delorean90 is there a way can i contact rockbox-themes.org to make the changes necesary
01:20:11 Quit lee-qid ("aufwiederbyebientotsayonara")
01:20:38Bagder(from lib/rts.src in the TI package)
01:20:52delorean90or should i just submit a new one and ask in the conformation email to delte the old one
01:20:54Bagderit's an ar archive with 1600 KB source
01:21:00 Join linuxstb_ [0] (n=chatzill@i-83-67-212-170.freedom2surf.net)
01:21:17PaulJamdelorean90: well, the theme is released under creative commons, so i think you are free to send a fixed version to rockbox-themes.
01:21:25 Join atsea- [0] (i=atsea-@gateway/tor/x-e2eac2ae8beec2a0)
01:21:53 Quit linuxstb (Nick collision from services.)
01:21:56 Nick linuxstb_ is now known as linuxstb (n=chatzill@i-83-67-212-170.freedom2surf.net)
01:22:04delorean90oh ok, yea, i have noticed alot of sloppy mistakes on my part that could use cleaning up, thank you for your help, i would have never figured out the issues without it
01:24:22 Quit Rob2222 (Read error: 110 (Connection timed out))
01:27:06 Part _novocaine_ ("Konversation terminated!")
01:27:19amiconnlinuxstb: Different Q - what's the current status of the logikdax port?
01:28:43linuxstbI'm currently investigating the NAND driver, but have stopped to work on viewports.
01:28:55saratogaso did TI release any documentation for their compiler?
01:28:57amiconnAH. Buttons and lcd do work?
01:29:07linuxstbYes.
01:29:10Bagdersaratoga: there are a few man pages and so, but not a lot
01:29:40*amiconn is asking because he wants to see the grayscale lib running on th elogikdax once the port is basically working
01:29:43saratogai've never programmed a DSP before, how is it different then a CPU?
01:29:56linuxstbamiconn: I've got a bootloader build working, and a toolchain which appends the bootloader to the original firmware, which can then be installed on the device (dual-booting works)
01:29:58amiconnIt will become easier once I get this all-new grayscale lib done
01:30:06saratogajust faster FMACs or do you have to deal with the hardware more?
01:30:31linuxstbamiconn: I guess I'll need interrupts and the timers working first though...
01:30:49amiconnThere will be way less bit shuffling than now, i.e. faster updates of moving contents, with vastly reduced graininess, and more shades
01:30:57eigmagenerally speaking, math is faster and general purpose stuff is slower.. for example, there's a SQUR (square) instruction
01:31:14amiconnOh, and half the ram requirements of the current version
01:31:30 Quit ender` (" PCMCIA: People Can't Memorize Computer Industry Acronyms")
01:31:34saratogain terms of writing code for it, is there anything extra you have to do?
01:31:45amiconneigma: Does it also have hardware loops?
01:31:47Bagdersaratoga: I don't think so
01:31:58Bagderthe code included is perfectly normal C code
01:32:03linuxstbamiconn: BTW, I don't think lcd speed will be a problem - I have a loop in my bootloader which does 2 full-screen lcd updates and an ADC read, and it's running at over 1000 iterations a second...
01:32:16amiconn:)
01:32:27eigmaamiconn: the C54x in particular has a RPT instruction that can prefix a certain class of other instructions, like moves, adds, macs, etc.
01:32:36jhMikeSamiconn: I don't know if preglow asked already but he informed me some work was done on voicebox but there was something with drag/drop? What's the state of that effort?
01:33:07amiconnVoiceBox can be used in 2 modes: from a gui (.hta), and with drag-and-drop onto the script
01:33:30jhMikeSso what else needs doing there now?
01:33:44amiconnThe latter isn't working yet if the drop target is the script itself
01:34:06 Quit delorean90 ()
01:34:16amiconnDrag-and-drop mode should work when the drop target is a link that sets the options
01:34:54amiconnI have to find a nice way to integrate better .ini reading in a way that works both directly in the .wsf and in the .hta
01:34:55jhMikeSeigma: like rep movs on x86?
01:35:56eigmajhMikeS: very similar; C54x can repeat a much broader set of instructions than x86
01:36:11amiconnThe MAS core does have hardware loops (looping through more than one instruction
01:36:53saratogaare hardware loops that useful? most of the cores we run have such shallow pipelines anyway
01:36:54amiconnOne special 'feaute' (dunno if other DSPs also have that) is that the pipeline never stalls, although instructions need more than one pipeline stage
01:37:05amiconns/feaute/feature/
01:37:50*jhMikeS hasn't seen any MAS programming docs
01:38:08amiconnSo if you use an accumulator value too early (before the previos instruction that sets it has finished), you'll be using the old value...
01:38:20amiconnjhMikeS: There are some leaked ones
01:38:26jhMikeShow much of a load is mp3 on MAS %-age wise?
01:38:45 Quit BigBambi (Remote closed the connection)
01:40:08amiconnI don't know
01:40:29*jhMikeS was wondering if there were sufficient resources to decode two mp3 streams "at once". well, send channel data with the frames anyway...crazy thought.
01:40:42eigmaamiconn: wow. from what I've read, I've never seen anything like that on the C54x..
01:41:10 Join Aware [0] (n=bill@CPE-58-161-148-11.nsw.bigpond.net.au)
01:41:30 Join pandrew [0] (n=andrew@86.122.168.62)
01:43:06jhMikeSamiconn: how did adding a new voice format break drag-drop? there didn't appear to be much more than setting the encoder path and options.
01:43:44amiconnIt now needs an option to select the encoder
01:44:00 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
01:44:26LloreanIs there some way for the script to know its own filename?
01:44:46LloreanAnd if not, any reason why we can't provide voicebox_mp3 and voicebox_spx scripts, even if they're 99.9% identical?
01:45:10JdGordonwoohoo! awesome news on the ti compiler
01:45:31jhMikeSMaybe read that propertly from the script tag. it can be assigned an object id
01:46:01jhMikeSJdGordon: they provided the source code too? or a win version?
01:46:03LloreanI was thinking we either provide one script, and tell the user to either modify an obvious, early line, or rename it (if it can know its own name) or just provide two, hard coded.
01:46:40JdGordonjhMikeS: havnt checked the download yet.. i tihnk its only binaries
01:47:37 Quit kugel (Remote closed the connection)
01:47:39jhMikeSI'm downloading that right now actually
01:49:30eigmajhMikeS: re: decoding two mp3 streams on the MAS, it would be interesting to optimize the mixing by taking advantage of that
01:49:51*amiconn goes to sleep
01:51:24jhMikeSeigma: I was thinking for voice+playback on hwcodec
01:53:22jhMikeSthough memory is a bit tight for that on those devices
01:53:52eigmaTTS on the DSP?
01:54:27 Quit moos ("Rockbox rules the DAP world")
01:54:29jhMikeSif you could fit it there
01:54:55eigmacan you give me a general idea of how tts works in Rockbox?
01:55:04linuxstbIt doesn't...
01:55:09eigma:D
01:55:18eigmasadly, that is the answer to so many similar questions
01:55:19*jhMikeS lost the typing race
01:55:25eigmahaha
01:55:48 Nick fxb is now known as fxb__ (n=felixbru@h1252615.stratoserver.net)
01:55:56LloreanRight now voice is just pre-recorded clips
01:56:07LloreanTTS would be nice, but the closest effort is, afaik, still license incompatible
01:56:12eigmaI see why "voice" is a better term than "TTS" :)
01:57:20eigmaI have the DSP and ARM interrupting each other btw.. I'm working on some debugging code so the DSP can send messages all the way to my screen, then I'm actually gonna dive into the McBSP
01:57:39 Quit z35 (Connection timed out)
01:58:05 Join z35 [0] (n=z@149.123.33.65.cfl.res.rr.com)
01:58:09 Join XavierGr [0] (n=xavier@ppp101-12.adsl.forthnet.gr)
01:58:12LloreanI read "McBSD" and for a second thought McDonalds was really branching out. :-P
01:58:16jhMikeSso it can show "Codec Failure"? :p
01:58:20eigmalol
01:58:26eigmathanks for the vote of confidence, guys ;)
01:58:39*jhMikeS just pokes fun
01:58:51jhMikeSfriendly jibe
01:59:27*eigma knows :)
02:00
02:00:55 Quit donutman25 ("ChatZilla 0.9.79 [Firefox 2.0.0.9/2007102514]")
02:02:45jhMikeSis any memory directly mapped to both cores?
02:03:08eigmayes, that
02:03:12eigmathat's how I load code into it
02:03:22eigmaI have to go though.. I'll talk to you guys later
02:03:29jhMikeSah...wanted to ask one more
02:03:32 Part mokkurkalve_
02:04:00jhMikeSso you think the scheduler could actually sync threads between them? it sounds like all the kernel objects could work.
02:06:54 Quit Xerion (Read error: 110 (Connection timed out))
02:09:01eigmaah, I don't think it's worth it
02:09:31eigmafor one thing, I don't think the DSP would benefit from preemptive multitasking. also, they're very different architectures
02:09:52eigmafor now, I'm going with the "traditional" approach
02:10:02 Join mickleby [0] (n=chatzill@adsl-69-107-93-82.dsl.pltn13.pacbell.net)
02:10:11eigmaone thread of execution on the DSP, and message passing between the two cores
02:10:19 Quit DerPapst ("So Long And Thanks For All The Fish!")
02:10:26 Quit spiorf (Remote closed the connection)
02:10:39 Quit eigma ()
02:13:30 Quit Nico_P (Remote closed the connection)
02:13:46 Quit mickleby (Client Quit)
02:14:44 Join oblib [0] (n=oblib@111-231-8-204.erd.cust.wirelessbeehive.com)
02:15:09oblibI'm just learning about WPS's. Is there any way to have multiple sized fonts in one WPS?
02:15:44LloreanNot currently without a patch
02:16:19oblibIs there a working patch?
02:16:52LloreanI think so
02:16:56linuxstbSearch for "multi-font"
02:17:36linuxstbWithout the hyphen...
02:18:23 Join me [0] (n=me@bas12-toronto63-1088795095.dsl.bell.ca)
02:19:42 Join mokkurkalve [0] (n=eivind@084202223083.customer.alfanett.no)
02:20:09obliblooks like that allows one font per area? I'm looking for multiple fonts in the WPS screen. Am I missing seeing that in multifont?
02:20:17mewhat image format/settings are needed for WPS? If I open up someone's theme image in gimp, and edit something and save it, all the purple (supposed to be transparent) stops being transparent and becomes just purple..
02:20:56linuxstboblib: I guess you need the "custom line" patch as well.
02:20:59Lloreanme: 24-bit bitmap
02:21:09methanks I'll give that a try :)
02:21:49oblib FS #5900 - Custom WPS Line with x/y/width coordinates and font face/color settings? Do I need both, or just 5900?
02:21:56obliboh ,nm
02:21:59linuxstbI think both, but check the comments.
02:22:03oblibI see dependency
02:23:47oblibWhat is "viewports" mentioned on that flyspray page?
02:24:12 Quit jhulst (Connection timed out)
02:25:43linuxstboblib: This gives some info - http://www.davechapman.f2s.com/rockbox/viewports.txt
02:26:06linuxstbI'm still working out the details, then I plan on implementing them.
02:27:26oblibwow, very ambitious. keep up the good work
02:28:10 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
02:30:08 Quit oblib ("Leaving")
02:30:18 Quit atsea- (Remote closed the connection)
02:33:19 Quit Thundercloud (Read error: 104 (Connection reset by peer))
02:35:43 Quit linuxstb ("ChatZilla 0.9.79 [Firefox 2.0.0.9/2007102514]")
02:39:15 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
02:39:26preglowgheh
02:39:48linuxstbhrmf
02:40:01 Join atsea- [0] (i=atsea-@gateway/tor/x-937c0e036717a033)
02:40:23preglowany viewports fun going on? :>
02:41:18 Quit me (Remote closed the connection)
02:41:47linuxstbNo fun yet, just thinking...
02:41:59 Quit XavierGr ()
02:42:12linuxstbIt's going to be cool though...
02:42:34preglowyeah, my brief read of the design doc definitely gave me the impression
02:43:28 Join woodensoul [0] (n=woodenso@72.86.92.80)
02:43:42 Join smooveJ [0] (i=4b0715e7@gateway/web/cgi-irc/labb.contactor.se/x-3d282090a56ab6eb)
02:43:50woodensoulWhat's the current state of the scrolling margins situation?
02:44:04smooveJHas anyone here figured out how to do smart playlists on Rockbox?
02:44:34preglowhow do playlists get smart?
02:44:52Shaid5 years at uni?
02:44:54linuxstbwoodensoul: The left-margin part of the patch has been committed to SVN
02:45:17linuxstbsmooveJ: I think you want to investigate database searches and customising tagnavi.config
02:45:18preglowi had some playlists accompany me for one year at uni, they still couldn't cope with basic arithmetic
02:45:20woodensoulThanks linuxstb, what about bmp resize?
02:45:36preglowand man did they suck at circuit design
02:45:39smooveJtagnavi.config? Thanks - i'll look into it
02:45:53linuxstbwoodensoul: Not committed, but I expect it will be in the not too distant future, when it's implemented nicely.
02:45:54smooveJsmart playlists are the only thing keeping me tied to iTunes
02:46:05Shaidnot too bad at musical theory though, preglow
02:46:22preglowthey were stuck with my choice of music, so they didn't do too well at that either :>
02:46:32Shaidhaha
02:46:45woodensoulOK, then the next question is related. I'm looking into putting the Cobalt them on a friend's iPod, but the wiki entry doesn't say what the album art dimension are that is uses. How can I find this out?
02:47:43linuxstbOpen up the .wps file in a text editor, and look for the %Cl line - that will look like %Cl|x|y|width|height|
02:47:53woodensoulthanks, got it open now
02:48:34woodensoul100x100, thanks again :)
02:49:02smooveJThanks LINUXSTB - tagnavi is exactly what I was looking for
02:50:13linuxstbsmooveJ: Am I right in thinking you can only create "smart playlists" in itunes, and not on the ipod itself (with Apple's firmware) ?
02:50:43Lloreanlinuxstb: As far as I know, at one point at least smart playlists were handled (updated, etc) only upon sync
02:53:44PaulJamlinuxstb: with rockbox you can't really create smart playlists on the device either, unless you want to edit the tagnavi_custom.config with the text_editor.
02:54:27linuxstbPaulJam: Yes, I didn't mean to imply Rockbox could do it, just curious if Apple's firmware could, and if so, what the UI looked like...
02:54:41LloreanPaulJam: But Rockbox's tagnavi filters update every time a song's played.
02:55:39LloreanAs opposed to every time the player is synced.
02:55:49PaulJamlinuxstb: sorry, i missunderstood/misinterpreted.
02:56:12linuxstbAnother job that rbutil could do - GUI tagnavi.config editor...
02:56:23PaulJamLlorean: yeah, i didn't think of that
02:56:30LloreanThough it's not so useful _yet_
02:57:13LloreanIt could be with options like "the 100 top rated songs" or "the 50 most listened to songs" or "the 75 songs that have been the longest since played"
02:57:59smooveJThat is correct. It's not possible to create a smart playlist on the fly using the Apple firmware.
02:58:49 Quit smooveJ ("CGI:IRC (EOF)")
02:59:48 Quit Calcipher ("—I-n-v-i-s-i-o-n— 2.0 Build 3515 with A Pack Fix By www.ircmadeasy.com")
03:00
03:02:41woodensoulAnyone know the current state of a Sansa charging within Rockbox?
03:02:56 Join donutman25 [0] (n=chatzill@65.75.87.48)
03:03:52 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
03:09:11***Saving seen data "./dancer.seen"
03:13:13 Quit nicktastic ("leaving")
03:14:56saratogai don't think it works, though apparently its known how to enable it
03:18:31 Join preet [0] (n=preet@70.48.23.163)
03:18:40 Quit PaulJam (".")
03:22:35preetHey all. I wanted to set up the dev environment (Linux). Downloaded the source, tried to run 'rockboxdev.sh', but got the following error: ROCKBOXDEV: patch is req'd for this script to work. Please install and re-run the script.
03:22:50 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
03:23:11 Join miepchen^schlaf [0] (n=hihi@p54BF4652.dip.t-dialin.net)
03:23:13preetI'm following the Building a Cross Compiler guide on the tWiki, but I don't see anything about this... any help would be appreciated
03:23:45Shaidsounds like you don't have 'patch' installed
03:23:52ShaidI'd go look up a version of it for your distro
03:28:36preetI'm running ubuntu, and the only patch I see for it is for 'rbutil'... I don't know what that is exactly, and can't figure out if it has anything to do with my prob
03:28:54Lloreanpreet: apt-get install build-essential
03:29:30 Join andrelop [0] (n=andrelop@200-171-6-135.dsl.telesp.net.br)
03:29:59 Part pandrew
03:30:37andrelopcould someone please help me on converting a 80GB 5,5 generation iPod to FAT32 so I can run RockBox on it ?
03:31:27andrelopI'm following the instructions at http://www.rockbox.org/twiki/bin/view/Main/IpodConversionToFAT32 but mformat on Linux isn't accepting all the options
03:33:26andrelopI'm trying to use the command : mformat -t 2428 -h 255 -s 63 -S 2048 -M 2048 -F a:
03:33:29 Join TMM [0] (n=hp@ip565b35da.direct-adsl.nl)
03:33:35ttuttle"It's a known issue."
03:33:46andrelopbut the only command I can use is : mformat -t 2428 -h 255 -s 63 -M 2048 -F a:
03:33:50ttuttlepfft
03:34:04daurnimatorwow
03:34:07daurnimatorthe sim works well
03:34:35andrelopttuttle: know issue ? is there a place I can read something about it ?
03:35:52ttuttleandrelop: oh, sorry, i was in the wrong channel
03:35:58 Quit andrelop (Read error: 104 (Connection reset by peer))
03:35:59ttuttleandrelop: i was talking to some friends about a cell phone problem
03:37:08preet@Llorean: install seems to be working now... thanks a lot! If you don't mind, what exactly was build-essential? Some C libraries?
03:37:22 Join XavierGr [0] (n=xavier@ppp101-12.adsl.forthnet.gr)
03:38:17Lloreanpreet: patch, probably diff, the standard set of c/c++ compilers, make, and others. Or at least, any you don't already have.
03:38:22LloreanAs far as I know, at least
03:42:37daurnimatormmm
03:42:53daurnimatoris it wrong to use rockbox as the media player on your computer? :p
03:43:03daurnimatorthe sim works so well
03:43:16daurnimatorand a symlink to my music drive works a treat :)
03:43:47maleI actually haven't tried the sim.
03:43:56maleIs it a fixed res window, or does it scale?
03:45:44LloreanFixed
03:45:52LloreanIt's meant to simulate the UI, not be used as a standalone player.
03:46:03LloreanWhich is why "Rockbox as an application" gets discussed from time to time
03:46:43daurnimatorit should/could
03:46:53 Join andrelop [0] (n=andrelop@200-171-6-135.dsl.telesp.net.br)
03:47:07maleYeah, but it's not as thought there's a shortage of media players on the desktop.
03:47:10LloreanThe simulator shouldn't resize...
03:47:31LloreanThe whole point is to have a per-pixel representation of what you see on the player's screen.
03:47:31maleIt could at least scale.
03:47:37LloreanIt can be run at 2x scale, I believe.
03:47:49daurnimatormale: there sure is a shortage
03:47:53daurnimatorof good ones
03:47:56male2x of my Nano screen is still fucking tiny.
03:48:04Lloreanmale: Please, watch your language in here.
03:48:30maleI was watching it. You should hear what I *WANTED* to say.
03:48:36LloreanWell, watch it more.
03:48:41LloreanI'm not asking for a debate, here.
03:48:50 Join webguest53 [0] (i=45ffd379@gateway/web/cgi-irc/labb.contactor.se/x-ea7dc14dc7ace5cf)
03:49:05daurnimatorLlorean: my reference was to that we SHOULD make a rockbox media player for pc
03:49:20maleSorry, in the future I'll bear in mind that #rockbox is PG-13.
03:49:39Lloreandaurnimator: I believe Rockbox as an App is in the "ideas for Summer of Code '08" already, I think.
03:49:44daurnimatoraren't they still allowed to say f*ck once per hour on PG13 movies? :p
03:49:52LloreanIt's something I'd quite like to see. Both as a PC-side player, and for porting it to PDA-style portable devices.
03:50:06Lloreandaurnimator: Masking it doesn't really make the word go away. =/
03:50:10daurnimatormeh, I'll start the project now
03:50:21daurnimatorLlorean: well, want me to unmask it in future? :p
03:50:25LloreanNo.
03:50:27LloreanI want you to stop.
03:50:29LloreanAs I said earlier.
03:50:31maleNo ***ing ****.
03:50:55maleSorry, I'll stop being a ****faced ****er.
03:51:04maleYeah, this is SO much better...
03:51:35daurnimatorhey, the stars could be hiding anything!
03:51:37daurnimator:P
03:51:49maleI really said snowballs and marshmallows.
03:51:50 Quit webguest53 (Client Quit)
03:52:09LloreanOr you could just stop being juvenile entirely before someone loses their already short patience today
03:52:18daurnimatorinfact, I thought ****er was "earlier" until I looked closer :p
03:52:23daurnimatorLlorean: I'll be good :)
03:52:25maleI thought I was being too adult...
03:52:37maleNow too juvenile. Tough crowd.
03:53:06daurnimatormale: I advise you to stop now, before someone gets hurt
03:53:24maleI'm already hurt. Cursing makes me feel like a big man...
03:53:30maleNow I have nothing.
03:53:32saratogai do want to see someone work on the app idea
03:53:33 Nick [omni] is now known as [inmo] (n=omni@bestII.com)
03:53:37 Nick [inmo] is now known as [omni] (n=omni@bestII.com)
03:53:42saratogaor better yet make it an app for Android
03:53:43Lloreansaratoga: I think many of us do, for varying reasons.
03:53:53daurnimatorandroid?
03:53:57maleYeah, but what about things like external control?
03:54:02saratogahopefully we'll get somone this year who wants to do it for a GSOC
03:54:03maleI control my music player with my WM.
03:54:03 Nick [omni] is now known as [inmo] (n=omni@bestII.com)
03:54:10maleI don't think I could do that with the RB sim.
03:54:22Lloreanmale: That's the whole point to "porting" it to an app
03:54:23daurnimatorI'm just thinking it through not
03:54:28LloreanRather than simply telling people to use the sim
03:54:33saratogaif someone comes up with a c to java bridge for android, i'll also propose android as a GSOC project
03:54:36daurnimators/not/now
03:54:43saratogawe'd get google funding for sure
03:55:26 Nick [inmo] is now known as [omni] (n=omni@bestII.com)
03:55:28Awarethat would be great if it ends up being any good
03:56:52Awareandroid, i mean
04:00
04:03:36daurnimatoryeah, the vid looks awesome :)
04:10:41 Quit andrelop ("Ex-Chat")
04:16:53saratogadigging through the android group posting, it does sound like they'll eventually be a way to call c code from their java environment
04:17:09saratogaso i guess you could write a media app and then call rockbox for playback, buffering, etc
04:17:43saratogathough i don't know how hard it would be to switch from SDL to the Google API
04:18:05JdGordonsholdnt be too hard...
04:18:08saratogaon the upside, having actual documentation would be a plus
04:18:09ShaidI though java supported calling C code via the 'native' uh, thing.
04:18:13JdGordonjust reimplement the sim layer...
04:18:14saratogaits not real java
04:18:25saratogaits some kind of google java
04:18:34saratogathough its quite similar, and the syntax is the same
04:18:42Shaidoh, right, one of THOSE
04:18:53saratogathey've apparently stripped it down for speed, and added a native GUI toolkit
04:18:58saratogaand its open sourced
04:19:19saratogathough maybe java is these days, i'm out of touch
04:19:42ShaidI haven't really kept up with Java either
04:19:42thegeek_I wonder how "open" the hardware itself will be
04:19:58thegeek_if they just want open source built ontop of "their" libs it's not that great
04:20:02thegeek_if the entire platform is open
04:20:05thegeek_it would be awsome
04:20:21thegeek_if the java libs/api sucks, just replace it;P
04:20:27saratogai'd guess it will be a lot like these days, where the hardware is still covered by the manufactorer's NDAs
04:20:35thegeek_perhaps not
04:20:39thegeek_linux kernel and all that
04:20:40saratogahowever since you'd have the entire OS source, that will make it a lot easier
04:20:50thegeek_ofcourse the drivers and such will be probably be closed
04:20:57saratogano need to reverse engineer things to get code running
04:21:24thegeek_;)
04:21:25 Nick mokkurkalve is now known as mokkurkalve__ (n=eivind@084202223083.customer.alfanett.no)
04:21:26saratogawill certainly be interesting if this takes off for phones and mp3 players
04:21:31thegeek_mmhm
04:21:57thegeek_I'm almost tempted to buy a windows mobile-based phone, just to run software on it
04:21:59thegeek_but meh
04:22:13thegeek_windows mobile is horribly slow
04:22:30thegeek_and on top of that it seems like microsoft is encouraging devs to use .net
04:22:32thegeek_teh horror
04:22:49saratogadoesn't .net compact compile to native code anyway?
04:22:51preetDoes anyone know if the mi4 files on this site: http://daniel.haxx.se/sansa/mi4.html are decrypted or not?
04:23:05thegeek_saratoga: oh it does?
04:23:13saratogajust object dump one and see if you get code
04:23:18thegeek_I just know I had to install the .net framework on my pda
04:23:22saratogai'd assume so
04:25:55 Nick mokkurkalve__ is now known as mokkurkalve (n=eivind@084202223083.customer.alfanett.no)
04:27:41preetDo I need to be an a specific dir to use arm-elf-objdump? I get 'command not found'
04:29:53preetbtw what i typed in exactly: arm-elf-objdump -D −−target binary -marm PP5020.mi4
04:30:09saratogayou need to have built the arm compiler and added it to your path
04:30:24preetrockboxdev.sh doesn't do that?
04:30:32saratogait builds it but does not add it to your path
04:30:48saratogatheres a command to do that in linux/cygwin build instructions
04:30:51saratogai don't remember it offhand
04:30:59preetalright, i'll check it out, thanks!
04:35:10saratogaLlorean: whos the best person to ask about the Sim?
04:36:46Lloreansaratoga: I honestly don't know if there is one. Maybe look for whoever did the big commit to bring the Win32/Linux sims to just SDL?
04:37:46saratogaok
04:38:02saratogai just want to put the guy who did that rockbox as an app work in contact with someone who knows about the sim
04:38:21saratogai'm still hopeful that we can eventually get him to merge his work
04:38:38LloreanYeah, it would be quite nice to get that merged
04:39:59 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:40:07 Join miepchen^schlaf [0] (n=hihi@p54BF7BA1.dip.t-dialin.net)
04:43:27 Quit saratoga ("CGI:IRC (EOF)")
04:44:46 Join psycho_maniac [0] (i=psycho_m@ppp128.hk.centurytel.net)
04:45:01preetI can't find this command to set the path after executing rockboxdev.sh...I found something about editing /etc/profile under the cygwin installation, but the stuff it says to edit isn't in my profile file...
04:46:26psycho_maniacawesome the viewports idea is on gsoc 08
04:49:41 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
04:50:37 Join webguest28 [0] (i=20896222@gateway/web/cgi-irc/labb.contactor.se/x-124770089b4ed17e)
04:53:25webguest28Can anyone help me out? I just got rockbox to work on my Sansa e260 and got everything set-up the way I'd like it, but I can;t get the absolute best sound from the eq parameters (eg. Hertz frequency, decibel levels, bass and treble levels, etc.) I listen to mostly rock and metal and I'd like to achieve an EQ set that brings out the most for that style of music. Anyone got any suggestions on the best fine twea
04:54:25hcswebguest28: your rant was cut off after "any suggestions on the best fine twea"
04:54:53webguest28Anyone got any suggestions on the best fine tweaking for the whole EQ gammut???
04:54:55webguest28Anyone got any suggestions on the best fine tweaking for the whole EQ gammut???
04:55:05psycho_maniacthey have saved eq settings in rockbox. ie: rap, hip-hop, rock. did you try those?
04:56:42webguest28yeah, but they all sounded too similar..(Not enough or too much bass, treble, hertz frequencies not bringing out the depth of music, either too much "tinny" sound or too muffeled.
04:57:26LloreanThen buy new earphones.
04:57:49webguest28Using a pair of Shure SE530-PTH
04:57:58Awarethen that's not the issue
04:58:02LloreanGenerally speaking, an EQ is to compensate for weaknesses in your hardware, as music is already mastered to where it's intended to be listened.
04:58:36Awarethere would probably be something on head-fi but since it's down...
04:58:41LloreanIf you feel you need to use an EQ to "bring the most out of your music", then you're talking about "adjusting the music to your personal tastes" which really isn't something someone else can do for you, since it depends entirely on how you like the original noise distorted.
04:59:10webguest28From what I've read on several forums is that the chip that the sansa e200 series uses is pretty weak when it comes to sound output. That's why I installed Rockbox since it has a wide range of sound adjustment options.
05:00
05:01:12webguest28I'm just trying to get a general idea of what a good setting would be to emphasize the music for styles of rock, metal, etc. as far as what the best starting point would be to set the decibel rates, the hertz frequency rates, etc. for the five EQ bands.
05:01:36daurnimatordon't ask us then
05:01:37daurnimator:p
05:01:46daurnimatortry looking in a pc app with an eq
05:01:54daurnimatorand see where the eq sliders go when you change genres
05:02:03daurnimatorthen exagerate that on your player
05:03:11psycho_maniacspeaking of eqs. where did the ones that come in rockbox come from?
05:04:08webguest28Ok. I'll give that a shot. Any good suggestions on a freeware/shareware app. that I can use for the EQ parameters? I used to be able to use EAX that I installed with a soundblaster x-fi card, but it fried a while back on my tower, so now i'm stuck to a laptop.
05:05:17webguest28The ones in rockbox were in a sub-folder in the software directory that is labeled "eqs" next to the "icons" and "docs" folders. All spelled like they are in the quotation marks
05:05:51psycho_maniacyes i know that. i want to know who decided to put them in the source. Was there any dicussion
05:07:43webguest28Not sure of that. Whomever created them and put them in the software did an ok job, but they all sound too similar to each other when adjusting for different music genres. Not a lot of extremes when it came to peak sound quality.
05:07:51Soap_I don't know off hand, but a quick search of the SVN commit logs would tell you, psycho_maniac.
05:08:23 Quit Soap (Nick collision from services.)
05:08:26 Nick Soap_ is now known as Soap (n=Soap@rockbox/staff/soap)
05:08:35Lloreanwebguest28: And EQ *should* never be extreme...
05:08:52 Join Soap_ [0] (n=Soap@rockbox/staff/soap)
05:09:12***Saving seen data "./dancer.seen"
05:13:11 Quit webguest28 ("CGI:IRC (Ping timeout)")
05:13:55Awarei don't believe in the use of eq whatsoever
05:14:53SoapSomeone needs to (and I'll do it next week if nobody beats me to it) go through (at least) the WPS page for the iPod Video and bring all the links back home. There are many links to offsite hosts (for images at least - I wouldn't be surprised if also for the WPSs)
05:18:32 Quit Aware ()
05:20:21psycho_maniacSoap: explain more what you want to do
05:21:21Soaprehost all offsite images (themes if need be) on the wiki.
05:25:24 Join mdevz [0] (i=mdevz@cpe-065-184-173-244.ec.res.rr.com)
05:25:25SoapJust start by doing a search for .com and .us.
05:25:39Soapyou'll notice how-fricking-many of the images are stored on photobucket.
05:25:58Soapand imageshack
05:26:52psycho_maniacoh yeah i noticed that uploading some wps's the other day
05:26:53SoapI'm curious if that is part of the reason the page takes six months to load. Regardless, I don't think Rockbox should be hosting images on photobucket.
05:27:05Soapspeaking of uploading some WPSs the other day.
05:27:16Soap1 - Do we have clear license to modify all those?
05:27:54Soap2 - pay attention to capitalization. I changed your link for blackglass-AA. unix is case sensitive.
05:28:03Soapdouble check the others.
05:28:09 Join AJG120882 [0] (i=20a5e43e@gateway/web/cgi-irc/labb.contactor.se/x-e0989a4792e38ffb)
05:28:29psycho_maniacoh. sorry. thanks
05:28:56SoapI only looked at that one - because I had seen in the logs an earlier blackglass-AA you were having trouble with - and was curious if you had figured out that %s and dynamic content doesn't go hand in hand.
05:29:24SoapYou were never online when I was around to point that out to you. ;)
05:29:36psycho_maniaci was probably either sleeping or working
05:29:41AJG120882Got disconnected earlier before I could ger an answer to this ?. From general opinons, what is the best overall mp3 player to purchase to replace a sansa e260? (eg. best sound, battery life, storage compacity, etc)
05:29:50psycho_maniacits only 1030pm here.
05:31:03LloreanAJG120882: Sansa e280.
05:31:04psycho_maniacAJG120882: have you checked the wiki?
05:31:10SoapAJG120882, http://www.rockbox.org/twiki/bin/view/Main/BuyersGuide
05:31:10AJG120882Sound being the most important feature, then storage, then battery run time per charge
05:31:28Soapwhat don't you like about the Sansa sound?
05:33:09AJG120882Checked Wiki, CNet reviews, Wired reviews and they all come up with mixed answers. I was thinking about an e280 to replace the 260, but sandisk support said that the e series won't support SDHC cards. Only up to 2GB Micro SD. And the new 16GB View is good, but no user-replaceable battery
05:34:38SoapRockbox supports SDHC
05:38:52 Quit AJG120882 ("CGI:IRC (EOF)")
05:39:43 Quit XavierGr (Read error: 110 (Connection timed out))
05:39:50 Join AJG120882 [0] (i=20a5e43e@gateway/web/cgi-irc/labb.contactor.se/x-d0b0ac877a291995)
05:41:33 Join me [0] (n=me@bas12-toronto63-1088795095.dsl.bell.ca)
05:42:07meis it possible to have differnt fonts fot the wps alone, while the rest of the theme has another font?
05:42:12LloreanNo.
05:42:23medarn :(
05:42:55AJG120882Better to get an e280 or the 16GB View that just came out a couple weeks ago?
05:43:09ttuttleAJG120882: the e250 I have is nice.
05:43:23LloreanAJG120882: Rockbox doesn't support the view.
05:43:45 Join jake_ [0] (n=jake@user-0c9afn5.cable.mindspring.com)
05:44:19AJG120882I like my e260, thinking about the 280 as long as it'll work with rockbox and with a 8gb Micro SDHC card that i have.
05:45:05 Join wbandit [0] (n=dustincr@c-68-47-133-166.hsd1.ar.comcast.net)
05:45:29 Quit wbandit (Client Quit)
05:45:51 Quit AJG120882 ("CGI:IRC (EOF)")
05:48:16 Join sarixe [0] (n=sarixe@ool-435403e9.dyn.optonline.net)
05:48:27 Quit preet ("Leaving")
05:48:34 Join AJG120882 [0] (i=20a5128d@gateway/web/cgi-irc/labb.contactor.se/x-f3c8b285214c5b3f)
05:50:40 Part radinp
05:54:57 Part AJG120882
05:57:21 Quit me (Remote closed the connection)
06:00
06:00:19 Join Traveler4 [0] (n=traveler@CPE-69-23-137-242.wi.res.rr.com)
06:00:47Traveler4is rockbox able to play compressed video?
06:01:10LloreanTraveler4: This and many other secrets are held in the cleverly hidden tome known as "The Manual"
06:04:34 Join courtc [0] (n=court@c-24-99-230-218.hsd1.ga.comcast.net)
06:05:58DogBoylol
06:06:17 Quit grndslm (No route to host)
06:06:30psycho_maniackinda late: is this the reason why the wps page on the wiki is so slow? other sites host the images? what about r-t?
06:08:13SoapI don't know if that is why it is so slow.
06:09:04psycho_maniaci just know that r-t.org is A LOT faster and i browse that first because its faster
06:11:32 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
06:11:53 Join miepchen^schlaf [0] (n=hihi@p54BF7BA1.dip.t-dialin.net)
06:22:22psycho_maniacis there a patch that will load a cfg file for each folder you load?
06:24:08LloreanI don't think so, but the tracker does have a search function
06:25:09psycho_maniaci know and i already searched. i tried folder and folders. anything else i could try?
06:27:20 Join Tdog [0] (n=Tdog@75-104-110-242.cust.wildblue.net)
06:27:39Tdoghey guys
06:28:20psycho_maniachello
06:32:39 Quit Traveler4 ("Java user signed off")
06:37:38 Part Tdog
06:38:58 Quit joshin (Read error: 113 (No route to host))
06:48:27psycho_maniacseems that 7738 might get committed soon :D
06:56:11 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
07:00
07:04:39ttuttle'night.
07:04:40 Part ttuttle
07:07:35 Quit mdevz (Read error: 110 (Connection timed out))
07:09:14***Saving seen data "./dancer.seen"
07:10:29 Quit linuxstb (Read error: 110 (Connection timed out))
07:20:01 Join webguest84 [0] (i=7d63919d@gateway/web/cgi-irc/labb.contactor.se/x-206ff07c3164b029)
07:22:21 Join mdevz [0] (i=mdevz@cpe-065-184-173-244.ec.res.rr.com)
07:23:55 Quit webguest84 (Client Quit)
07:30:26 Join midkay [0] (n=midkay@rockbox/developer/midkay)
07:37:01 Quit woodensoul ()
07:37:58 Join Workaphobia [0] (n=Jonathan@magneton-35.dynamic.rpi.edu)
07:47:47 Quit homielowe (Read error: 110 (Connection timed out))
07:51:40 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
07:52:02 Join miepchen^schlaf [0] (n=hihi@p54BF7BA1.dip.t-dialin.net)
07:56:56psycho_maniachow do i make a zip file for a simulator? i built the simulator and am now able to run it. what directory do i go in to run the make zip command?
08:00
08:01:19psycho_maniacholy smokes. 8mbs for a zip file for a simulator. does that sound right?
08:04:06 Join Tetris-Block [0] (n=End@bosko-78.dynamic2.rpi.edu)
08:10:44 Join weezerle_moep [0] (n=weezerle@dslb-088-072-030-169.pools.arcor-ip.net)
08:11:26 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
08:12:46 Quit weezerle (Read error: 110 (Connection timed out))
08:14:42 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
08:15:03 Join miepchen^schlaf [0] (n=hihi@p54BF7BA1.dip.t-dialin.net)
08:18:14 Join Crash91 [0] (i=d98bd903@gateway/web/cgi-irc/labb.contactor.se/x-5f6af47d83bab8c3)
08:19:06 Join Rob2222 [0] (n=Miranda@p54B169DF.dip.t-dialin.net)
08:19:09Crash91hi
08:19:20psycho_maniachello
08:23:03 Quit Crash91 (Client Quit)
08:28:44 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
08:29:05 Join miepchen^schlaf [0] (n=hihi@p54BF7BA1.dip.t-dialin.net)
08:34:13 Join homielowe [0] (n=chatzill@d207-81-67-190.bchsia.telus.net)
08:36:33 Quit Rob222241 (Read error: 110 (Connection timed out))
08:48:49 Join desowin [0] (n=desowin@hdp186.internetdsl.tpnet.pl)
08:52:02 Quit tuplanolla (Read error: 104 (Connection reset by peer))
08:52:05 Join tuplanolla [0] (n=jani@a80-186-126-4.elisa-laajakaista.fi)
09:00
09:01:44 Quit sarixe ("(EE) Failed to load "quit" module")
09:04:20 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
09:08:58 Join Xan [0] (n=mathuran@125.99.145.157)
09:09:17***Saving seen data "./dancer.seen"
09:11:39XanHello...Can some1 teach me how to use the recorder iin the Ipod Nano 1st gen using rockbox
09:12:09Xandoes the ipod really have recording capabilities?
09:13:36hcsXan: yes, according to the device char
09:13:40hcs*chart
09:13:46homieloweXan: the Nano can only record via line-in ( through the dock connector) don't attempt to pump music throught the headphone jack, it could possibly damage your player.
09:16:45homieloweThis is a link to where you can buy your own parts and fashion yourself a cable http://home.swipnet.se/ridax/connector.htm .
09:24:49 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
09:25:10 Join miepchen^schlaf [0] (n=hihi@p54BF7BA1.dip.t-dialin.net)
09:31:38 Part homielowe
09:32:47 Join homielowe [0] (n=chatzill@d207-81-67-190.bchsia.telus.net)
09:34:20 Quit Xan (Read error: 110 (Connection timed out))
09:40:53 Join stevenm [0] (n=stevenm@infranelson.student.umd.edu)
09:40:58stevenmHello people
09:41:42stevenmHey n1s, if you're interested, I re-did findDelta in sequencer.c last night. FRACTSIZE got bumped to 12 bits also
09:42:07stevenmn1s, I changed the order of operations (all mult before all div). This gives a more accurate result, but requires more intermediate storage space, hence the long long
09:42:18stevenmn1s, have you got any idea how to optimize this ?
09:48:58 Quit desowin (Read error: 113 (No route to host))
09:54:42 Quit psycho_maniac (" HydraIRC -> http://www.hydrairc.com <- It'll be on slashdot one day...")
10:00
10:00:16 Join webguest14 [0] (i=430296f9@gateway/web/cgi-irc/labb.contactor.se/x-35e07e793add418c)
10:02:47 Part Workaphobia
10:02:50 Part webguest14
10:04:26 Join woodensoul [0] (n=woodenso@72.86.92.80)
10:11:20 Quit ivan` (Read error: 110 (Connection timed out))
10:12:11woodensoulI'm trying to fix the AABox theme for the Gigabeat...
10:12:21woodensoulThis line has me confused: %Cl|0|0|sc240|sb240|
10:12:31woodensoulIt's for the album art I assume.
10:13:35 Join ivan` [0] (n=ivan`@adsl-71-142-236-175.dsl.scrm01.pacbell.net)
10:13:55 Join Thundercloud [0] (n=thunderc@resnet31.nat.lancs.ac.uk)
10:14:22 Quit amiconn (" mrf")
10:14:54 Join bertrik [0] (n=Bertrik_@121-021-045-062.dynamic.caiway.nl)
10:18:56 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
10:25:03 Join ender` [0] (i=krneki@84-255-206-8.static.t-2.net)
10:25:29 Join root_ [0] (n=root@210-84-19-213.dyn.iinet.net.au)
10:26:40root_anyone here on debian?
10:26:47root_oops
10:26:51 Quit root_ (Client Quit)
10:27:23 Join ddalton [0] (n=daniel@210-84-19-213.dyn.iinet.net.au)
10:27:28ddaltonok anyone here on debian?
10:33:12 Quit BigBambi (Remote closed the connection)
10:35:18 Quit ddalton ("Lost terminal")
10:36:41 Join Agrajag__ [0] (n=Agrajag@c-68-84-64-192.hsd1.nj.comcast.net)
10:36:55 Join davina [0] (n=davina@cpc1-sout6-0-0-cust616.sotn.cable.ntl.com)
10:37:26Agrajag__Guys, new to Rockbox. All working great. Now a new build is out. What's the best way to get this new build onto the iPod Photo? Is there a util? Drop the whole zip? Just the exe?
10:38:05hcsAgrajag__: just "drop the whole zip", as you say, overwriting all the old files
10:38:11scorcheextract the zip to the root of the device, as the manual says
10:38:17Agrajag__That won't wipe all the old settings?
10:38:29scorcheno
10:38:52Agrajag__Great, thanks guys. Anyone know if this latest build supports FLAC embedded album art?
10:38:56scorcheyou can always save your settings to another file if you wish
10:39:14Agrajag__Probably a good idea. Just a bit overwhelmed just out of the gate learning it all.
10:39:16scorcheit doesnt, which you would have seen from the commit messages
10:39:32Agrajag__What's a commit message?
10:39:48scorcheas well, new builds come out multiple times a day, so i wouldnt recommend upgrading everytime a new one is out ;)
10:40:02Agrajag__Then I might way as my build is from the 10th.
10:40:06Agrajag__wait
10:40:17Agrajag__Can't imagine that much changed by the 17th.
10:40:19woodensoulSpeaking of album art, I'm trying to fix the AABox theme that uses album art and so far the only problem I can't figure out is why a rectangular section of the album art remains black.
10:40:28scorchehttp://www.rockbox.org/recent.shtml#svn
10:40:37scorchea commit message is the "what" on there
10:40:41Agrajag__Bookmarked, thx
10:40:51 Join ddalton [0] (n=daniel@210-84-19-213.dyn.iinet.net.au)
10:42:00Agrajag__Any recommendations for a noob of what I could be doing with Rockbox that I might not know out of the gate? Currently I have 1600 320k MP3's on it. That's about it. Later I want to consider moving FLAC files onto it (I work for a high-end earphone company and use this for demos on show floors).
10:42:07woodensoulCan someone take a look at the code?
10:42:41 Join Isolinear [0] (n=A@c-76-105-254-119.hsd1.or.comcast.net)
10:42:53Agrajag__Oh, and is there a way to change "Database" to "Media" or something simple that other uses would equate with music, etc?
10:43:11scorcheAgrajag__: if you want to know what you can do, you might consider reading the manual, as it lists all you can do
10:43:26scorchethere is if you recompile, but only then
10:43:30Agrajag__I rifled through it already but assumed it might be a bit out of date.
10:43:38scorchewhy did you assume that?
10:43:57Agrajag__That's the way it normally is with 90% of software out there. Nothing against RB at all.
10:44:18Agrajag__You get the manual but then the community is doing all sorts of beyond-the-manual things.
10:44:30Agrajag__Nice to know this one is complete.
10:44:53scorchewe might lag a few days behind on new features, but we have a few people very dedicated to our fine manual and are quite proud of them :)
10:45:14Agrajag__It was pretty damned useful though I couldn't get auto-install to work. I ended up just doing it manually.
10:45:30Agrajag__You guys should be proud of that manual. It's great.
10:46:39scorchecan you join #rockbox-community for a sec?
10:46:50Agrajag__Who?
10:46:52Lloreanscorche, Agrajag__: Actually, if you edit the language files, you can change any of the menu text without having to recompile.
10:47:04 Quit ompaul (Client Quit)
10:47:04woodensoulAnyone have an idea on my album art question?
10:47:09scorcheLlorean: doesnt it have to be recompiled into a lng though?
10:47:13LloreanYeah
10:47:33LloreanBut it doesn't require setting up a full dev environment, at least to the extent of getting cross-compilers and stuff
10:47:39LloreanActually, wait...
10:47:43LloreanLangV2 does at the moment. =/
10:47:49*scorche nods
10:47:50LloreanNeeds a GCC, doesn't it?
10:47:56scorchei think so
10:48:04Agrajag__Too much of a hassle then. Don't sweat it. It'd be a nice feature later for the non-techies.
10:48:39LloreanAgrajag__: Part of the reason Database is named as it is, is that we don't want to suggest to people that it's the *only* way to play music.
10:49:00LloreanA name like "Music" or "Media" might suggest to people that it's how they're supposed to access it, rather than simply an alternative means to filetree based access
10:49:03Agrajag__The trouble I ran into was when my wife got hold of it and had NO idea where to start. <grin>
10:49:10amiconnAnd 'media' wouldn't be correct. It's 'audio'
10:49:24LloreanIt's very definitely "audio" at the moment, yes.
10:49:39LloreanAnd probably should stay that way
10:49:47Agrajag__Ah, I thought I read that this would let the Photo play video.
10:49:57LloreanYou can play video on the Photo
10:50:01LloreanBut the database doesn't index Videos.
10:50:16LloreanYour songs, pictures, videos, text files, etc, are all accessible via the Files browser
10:50:21Agrajag__I just hope I can play lossless now without it breaking up like it did with the default firmware.
10:51:11woodensoulWhat does toppos (top position) refer to in the progress bar syntax? %pb|height|leftpos|rightpos|toppos|
10:51:54 Join Buschel [0] (n=AndreeBu@p54A3E456.dip.t-dialin.net)
10:51:55Lloreanwoodensoul: Coordinate to draw at?
10:52:02 Quit stevenm (Read error: 110 (Connection timed out))
10:52:46woodensoulthat doesn't help
10:53:17woodensoulleft position is obvious, height is obvious, right position is how far from the right side it ends, top.... ?
10:53:34amiconnLlorean: I don't think videos can be tagged at all, so it would probably stay an audio database even on a PMP
10:56:03Lloreanamiconn: MP4 videos can, at least.
10:56:08LloreanMPEG2, probably not
10:56:17amiconnDidn't know that...
10:56:20Lloreanwoodensoul: How far from the top of the screen...
10:56:50amiconnAnd photos can't be tagged either (unless you count EXIF as a tagging format)
10:58:23 Join pixelma [0] (n=Marianne@rockbox/staff/pixelma)
10:58:56 Quit ddalton ("Lost terminal")
10:59:23Buschelint the last days iu played around a lot with DEV_EN1, CACHE_PRIORITY(DEV_EN3?) and GPO32_ENABLE on iPod Video. I switched off all the bits in DEV_EN1/EN3 which did not show any effect -> no change in power consumption. I switched off/on IDE_ENABLE together with HDD power off -> no change in power consumption. Now I switched off all bits in GPO32_ENABLE which could be removed -> power consumption goes up. This dam* PP-registers :/
11:00
11:00:48amiconnHow do you monitor power consumption?
11:00:56woodensoulThanks Llorean. I'm trying to figure out why a huge rectangular section of the album art is just black in this WPS I'm fixing.
11:01:45Buschelamiconn: running a battery bench and in paralell reading the battery percents during the test and use Excel.
11:01:56Buschelamiconn: quite a pain
11:02:04amiconnHmm. Not very precise, and it's time consuming
11:02:19amiconnUsing an ammeter would be both more accurate and way faster
11:02:53amiconnUnfortunately it's not that simple to use an ammeter for current measurement on G5
11:02:59Buschelamiconn: I don't have any. The main issue with this measurement is that the test take at least 4h
11:04:20Buschelnevertheless there are results which show no/few effect when changing DEV_EN1/EN3 and (bad) effect when playing around with GPO32-ENABLE
11:05:32Buschelamiconn: btw, do you know how often the CPU switches between boost/normal in a standard use case like mp3-playback?
11:05:54Buschelamiconn: not in percent, but in switches/sec
11:06:32 Quit nanok (Read error: 113 (No route to host))
11:07:45 Join Robin0800 [0] (n=Robin080@cpc1-brig8-0-0-cust97.brig.cable.ntl.com)
11:09:20***Saving seen data "./dancer.seen"
11:09:37woodensoulWhy does the Wiki speak of resizing the album art bitmap if that is not committed yet?
11:10:26amiconnBuschel: CPU boosting is completely on-demand. so it depends on the actual tracks and formats how often it changes frequency
11:10:27bertrikBuschel: on my sansa it depends a bit on the particular song but it's on the order of about once every two seconds
11:11:01amiconnCan be as low as never (for flac files). The upper limit is about 1..2 times per second
11:14:02Buschelamiconn: ahh, ok. I never knew about this limitation −− good solution.
11:14:12amiconnThis is no hard limit
11:14:24amiconnIt's just what results from playback behaviour
11:15:43BuschelHmm, so it's done if a buffer watermark is reached. Is this checked each tick or so?
11:16:11 Join MethoS- [0] (n=clemens@pD955E71F.dip.t-dialin.net)
11:17:04 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
11:17:25 Join miepchen^schlaf [0] (n=hihi@p54BF7BA1.dip.t-dialin.net)
11:19:36 Join Redbreva [0] (n=chatzill@host86-144-108-57.range86-144.btcentralplus.com)
11:21:27woodensoulWhat is the height of the progress bar?
11:21:35woodensoulerr status bar
11:22:34 Quit MethoS- (Remote closed the connection)
11:25:35 Join MethoS- [0] (n=clemens@pD955E71F.dip.t-dialin.net)
11:28:37 Quit amiconn (" HydraIRC -> http://www.hydrairc.com <- Like it? Visit #hydrairc on EFNet")
11:30:08 Join Bagder_ [0] (n=daniel@1-1-5-26a.hud.sth.bostream.se)
11:30:16 Quit MethoS- (Remote closed the connection)
11:30:27 Join moos [0] (i=moos@m147.net81-66-159.noos.fr)
11:30:28 Join MethoS- [0] (n=clemens@pD955E71F.dip.t-dialin.net)
11:32:18 Quit Robin0800 (Read error: 110 (Connection timed out))
11:46:18 Quit Bagder (Read error: 110 (Connection timed out))
11:46:39 Quit atsea- (Dead socket)
11:51:11 Part Agrajag__
11:53:38 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
11:53:59 Join miepchen^schlaf [0] (n=hihi@p54BF7BA1.dip.t-dialin.net)
11:54:12 Quit weezerle_moep ("...und tschüss!")
11:56:11 Part pixelma
11:58:07 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
11:58:40 Join Robin0800 [0] (n=Robin080@cpc1-brig8-0-0-cust97.brig.cable.ntl.com)
12:00
12:00:53 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
12:03:17 Join Lear [0] (i=chatzill@rockbox/developer/lear)
12:06:04 Join Falen [0] (n=chatzill@81-237-250-8-no147.tbcn.telia.com)
12:06:17 Quit bertrik ("bye")
12:08:50 Join merbanan [0] (n=banan@83.233.243.232)
12:11:58Falenwarning: function declaration isn't a prototype
12:12:17FalenThat error shows up on all the function prototypes I got
12:13:11Falenhttp://pastebin.com/m10435413
12:13:59 Join linuxstb__ [0] (n=linuxstb@i-83-67-212-170.freedom2surf.net)
12:14:09 Quit linuxstb_ (Nick collision from services.)
12:14:13 Nick linuxstb__ is now known as linuxstb (n=linuxstb@i-83-67-212-170.freedom2surf.net)
12:14:33linuxstbFalen: You need to write them as "void func(void);"
12:15:40FalenOk
12:16:07FalenYeah. That got the warnings away, thank you
12:17:26moosFalen: out of curiosity, what plugin are you working on?
12:17:43FalenA basic intepreter
12:18:29moosworking from the scratch or porting something?
12:18:32FalenIt's simple, but the idea is to port over most of the rockbox functions (drawing, files, ...)
12:18:40 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
12:18:48moosoh ok
12:21:06FalenIt's a port, but it's so modified you can't see the original code (from ISO standard), so I'd rather call it my own :)
12:21:08JdGordoncan someone with a radio target please try the patch attached to FS #8168 and let me know if it fixes the bug?
12:21:38moosFalen: hehe, good luck then :)
12:24:05moosJdGordon: don't have my iaudio at hand for test :(
12:24:28*JdGordon 's sansa has no radio and my brother has my h300 :( and i cant test that in the sim
12:24:37JdGordonLlorean: does your sansa have radio?
12:25:08JdGordonLlorean: also, feel like fixing 8169?
12:26:33moosJdGordon: how goes your icon code status bar code going?
12:26:38LloreanJdGordon: Yes, yes, busy right now though.
12:27:11JdGordonmoos: got fed up with it... its been pushed to the side for the itime being :p
12:27:33mooshehe :)
12:30:25 Quit jhMikeS (Read error: 104 (Connection reset by peer))
12:33:09linuxstbI'm not sure if it will work, but I had an idea for how we can customise the status bar as part of thinking about viewports. The idea is that each viewport defined by the WPS is given a type ("wps" or "status bar"). The wps_refresh() option would then take a parameter saying which kinds of viewports should be drawn, and only draw those. The existing status bar code would simply be replaced by a call to wps_refresh() telling it t
12:33:09linuxstbo only refresh viewports marked as "status bar". The normal WPS refresh would refresh both types of viewport.
12:35:03JdGordonbut the statusbar doesnt have naything to dso with the wps ddrawing...
12:35:22JdGordon_unlss_ we want to wps-ify it.. but even then it would still be seperate
12:35:23linuxstbWhy not?
12:35:28JdGordonjust it would use the parser
12:37:31linuxstbThis would also allow the theme designer to make the "status bar" a lot more than just a status bar, and to display WPS content with the menus...
12:38:10JdGordonyes but the statusbar doesnt have to be part of the wps to do that
12:39:05 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
12:46:47 Nick parafin|away is now known as parafin (i=parafin@paraf.in)
12:53:45 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
12:58:18 Quit amiconn (Nick collision from services.)
12:58:27 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
12:59:37 Join pixelma [0] (n=Marianne@rockbox/staff/pixelma)
13:00
13:03:32n1sstevenm: haven't looked at your changes at all, but generally long long is bad in performance critical code... will try to think of something seriously hung over now though :/
13:05:14 Join Daniel_S [0] (i=57b0dc0c@gateway/web/cgi-irc/labb.contactor.se/x-ccfd53e4bc8ccd17)
13:09:21***Saving seen data "./dancer.seen"
13:13:04 Join lee-qid [0] (n=liqid@p54964F37.dip.t-dialin.net)
13:13:15 Part Llorean
13:15:23 Quit martin_ (Remote closed the connection)
13:16:18 Join zicho [0] (n=martin@c-6a98e355.68-7-64736c14.cust.bredbandsbolaget.se)
13:16:21 Join Lord [0] (n=mathuran@125.99.145.157)
13:16:21 Quit qweru (Read error: 104 (Connection reset by peer))
13:19:12 Part Lord
13:19:23 Join Lord [0] (n=mathuran@125.99.145.157)
13:21:39 Join SkinInd95 [0] (n=travis@host-69-144-93-208.hln-mt.client.bresnan.net)
13:24:04 Join Riggzy [0] (n=Riggzy@78.150.118.50)
13:24:45RiggzyHmm, I seem to be having a track skipping problem on my 5G iPod
13:25:32RiggzyWhich I'm still debugging a little... Playing an MP3, and at a certain point the progress bar jumps to the end, and it skips to the next track
13:26:05RiggzyAny pointers as to what to check?
13:26:22moosRiggzy: that is a known bug
13:26:46Riggzymoos: What's causing it, if it's known?
13:27:51FalenAm I missing something?, http://pastebin.com/m4b031f53, this code draws all the lines like it should (the top is taken from stopwatch). But it draws them with a very large height between
13:27:51 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
13:28:12 Join miepchen^schlaf [0] (n=hihi@p54BF7BA1.dip.t-dialin.net)
13:28:47moosRiggzy: that was introduced with those recents metadatas on the buffer changes. A solution is still didn't find yet by the developer that worked/working on it
13:29:34moosRiggzy: best to be patient...
13:29:56linuxstbFalen: I think lcd_puts() takes x,y in character positions - you don't need to multiply by lineheight.
13:30:09LordHow can i record using my Ipod nano?
13:30:20Nico_PRiggzy: did you resume on that MP3 then skip back to the beginning of it?
13:30:44FalenAha, Ok
13:30:46Soap_lord - there are line-in pins in the dock connector.
13:31:20Lordand what if i dont have a dock
13:31:27preglowhaha
13:31:31moosNico_P: the bug appear every time you skip back and cause rebuffering
13:31:39preglowmy h120 just skipped to some weird-ass track when i tried to change track at the end of a song
13:31:41Soap_Lord: you can make a dock connector to line-in adapter if you want.
13:31:43jmspeexIn case anyone here's interested, it's now possible to build libspeex (encode+decode) without a single float instruction.
13:31:51FalenYeah, now it works!
13:31:52preglowjhMikeS: woop, will check out
13:31:59Nico_Pmoos: true.
13:31:59RiggzyNico_P: Just going to see if it's time based or at that specific pointin the track..
13:32:07jmspeexpreglow: I'd be interested to know how big that turns out on Coldfire/ARM
13:32:13preglowjmspeex: that was for you, yes...
13:32:17preglowmy tab completer sucks
13:32:23LordSoap: hmm.. i dont really understand you ..
13:32:35jmspeexhaha
13:32:45jmspeexpreglow: −−disable-float-api −−disable-vbr
13:32:49Nico_PRiggzy: did you do some seeking in the track?
13:32:50RiggzySigh, my disk is making nasty noises
13:33:09Soap_Lord: You can buy a dock connector (to plug into the butt of your Nano) and wire up a custom line-in cable to it.
13:33:23RiggzyNico_P: Nope, just tried that, looks like my disk is having some problems though x.x
13:33:29jmspeexpreglow: oops, forgot to push to svn. Done now.
13:33:31RiggzyAnd not just on Rockbox
13:33:36LordSoap_: Oh! ok...
13:33:49Soap_Lord: Basicly I'm saying the pins are there for line-in on the Nano - but nobody seems to have used it. So if you are going to use it - it will most likely take some hardware hacking on your part.
13:34:09LordSoap_: ok...
13:34:21LordSoap_: thanks mate..!
13:34:38Soap_one second, lodesi
13:34:41Soap_Lord even
13:34:42preglowjmspeex: i'd have to either sync rockbox svn with that first or try doing a speex svn build with cross compilers
13:34:53Lordyes...?
13:35:40Soap_http://home.swipnet.se/ridax/connector.htm <-if you want to wire your own.
13:35:57 Join davina_ [0] (n=davina@cpc1-sout6-0-0-cust616.sotn.cable.ntl.com)
13:36:16LordSoap_: wow! thanks!
13:36:30 Quit UncleRemus ("leaving")
13:36:39markunjmspeex: did you see my PMs?
13:36:49 Nick Bagder_ is now known as Bagder (n=daniel@1-1-5-26a.hud.sth.bostream.se)
13:36:54 Join Xerion [0] (i=xerion@cp198589-d.landg1.lb.home.nl)
13:37:12 Join Frazz [0] (n=Fraser@thelawsons.plus.com)
13:37:19 Quit SkinInd95 ("Ex-Chat")
13:37:31jmspeexmarkun: just saw it. Cool.
13:37:44LordSoap_: well i guess i better go buy the dock.. ium a noob with the ipod
13:37:49Lordhehe
13:38:00preglowjmspeex: configure isn't too happy to use my m68k and arm crosscompilers, so it'll have to wait after i sync
13:38:04 Join webguest36 [0] (i=50e0a118@gateway/web/cgi-irc/labb.contactor.se/x-820e1aa919656353)
13:38:11jmspeexmarkun: actually, I should have mentioned that my experience with Blackfin gcc is from quite a while ago. I'm glad it has improved.
13:38:26Soap_Lord: make sure any dock you buy has a line-in jack. Just because the Nano has the capability...I'm not sure the Apple dock has a line-in jack.
13:38:44pixelmaNico_P: just seen the bug with next song info (I think I read about it before either here or in the bug tracker). If there is no next track at the end of the playlist it shows the info of the current track - and this even though I have it in a %?Fn conditional...
13:38:53jmspeexpreglow: most of the changes I did are on the encoder BTW. In terms of decoder, it should be similar to what you already did.
13:38:57Lordohk...
13:38:57linuxstbSoap_: I'm not sure if there are any commercial docks with line-in..
13:39:25Nico_Ppixelma: oh yes, I had forgotten about that one... thanks for reminding me :)
13:39:33*Riggzy baps his iPod around a bit, runs it on official firmware for a bit
13:39:35preglowjmspeex: sure, but i can enable the encoder anyway
13:39:42preglowjust not use it
13:39:59preglowjmspeex: btw, we've integrated speex as a voice codec now, works great
13:40:05pixelmaNico_P: you're welcome ;)
13:40:09moosNico_P: time to come back to MoB bug hunting ;)
13:40:12Soap_They sell one for $40 @ the link I posted Lord earlier. I had posted the link, though, because you can buy unwired connectors for $2
13:40:15jmspeexpreglow: cool
13:40:16preglowlike expected, smaller filer, better quality
13:40:19Nico_Phehe :)
13:40:26 Join DerPapst [0] (n=DerPapst@p5B23EBFC.dip.t-dialin.net)
13:40:30jmspeexpreglow: got your "raw" decoder working?
13:40:36preglowjmspeex: we'll probably go svn with it today or tomorrow, need to coordinate a ton of stuff first
13:40:40preglowjmspeex: raw encoder and decoder, yeah
13:40:47Nico_Ppixelma: I have something almost ready to fix the problems with AA clearing
13:40:49preglowjmspeex: raw decoder isn't exactly much work...
13:41:04Lordlinuxstb: http://home.swipnet.se/ridax/connector.htm
13:41:26preglowjmspeex: btw, it'd be cool if you could have a quick glance through the encoder to see if there's anything silly, it's quite small
13:41:29preglowjmspeex: http://svn.rockbox.org/viewvc.cgi/trunk/tools/rbspeex/rbspeexenc.c?view=log
13:42:29linuxstbLord, Soap_: Yes, just saw that - I assumed (!) that you were talking about the "normal" kinds of docks, rather than that kind of supplier...
13:43:02 Join UncleRemus [0] (n=caj@81-234-174-183-no95.tbcn.telia.com)
13:43:31LordSoap_, linuxstb :ok then... thanks a ton mate!.. now i gotta go around play with my nano for a while... :)
13:44:57 Quit Lord ()
13:47:43jmspeexpreglow: I'm having a hard time following what you're doing with the resamler. Looks odd
13:48:19preglowjmspeex: really? was the simplest i could make it :)
13:49:02jmspeexOK, you're reading everything in one bit, right?
13:49:10preglowjmspeex: indeed
13:49:20preglowjmspeex: all our clips are small, so i did that for simplicity
13:49:50preglowjmspeex: i basically just have the resampler output blocks of frame_size a bunch of times
13:50:11 Quit davina (Read error: 110 (Connection timed out))
13:50:24FalenIs there sprintf function in rockbox, if so, where is it located?
13:50:43 Quit UncleRemus ("leaving")
13:50:48preglowFalen: there is snprintf
13:51:33FalenOk
13:51:59jmspeexpreglow: BTW, you say "Limit this or resampler will try to allocate it all on stack". What do you mean exactly.
13:52:00jmspeex?
13:52:38preglowjmspeex: just have a look at speex_resampler_process_int, it converts the entire input buffer to float before calling the float version
13:52:53 Quit Redbreva (Read error: 104 (Connection reset by peer))
13:53:07 Join Febs [0] (n=chatzill@207-172-204-33.c3-0.rdl-ubr4.trpr-rdl.pa.cable.rcn.com)
13:53:23jmspeexpreglow: sorry, forgot you weren't using fixed-point for that
13:53:23preglowjmspeex: not just as much as it needs of it
13:53:46preglowi could do a proper calc there and limit it only to what is needed, of course
13:53:50preglowas a matter of fact, i guess i should do that...
13:53:56jmspeexpreglow: actually, you can disable that behaviour by not defining VAR_ARRAYS
13:54:57 Join UncleRemus [0] (n=caj@81-234-174-183-no95.tbcn.telia.com)
13:54:59preglowjmspeex: i'm find with VAR_ARRAYS, i'll just do it properly on my end anyway
13:55:06preglowfind/fine
13:57:28 Join Casainho [0] (n=chatzill@87-196-37-73.net.novis.pt)
13:57:50CasainhoBagder: Hello! :-)
13:59:54 Join hunz_ [0] (n=hunz@dslb-088-064-071-194.pools.arcor-ip.net)
13:59:58hunz_hey there
14:00
14:04:00 Quit UncleRemus ("leaving")
14:04:01*amiconn wants to remind Nico_P of another bug
14:04:11Nico_Pyes?
14:04:21 Quit lee-qid ("aufwiederbyebientotsayonara")
14:04:24amiconnThe playlist index in the wps is still advancing too early
14:05:09 Quit jhMikeS (Nick collision from services.)
14:05:15 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
14:05:26 Quit Robin0800 (Read error: 110 (Connection timed out))
14:05:27Nico_Pamiconn: that's not new to MoB and last time I tried to fix it, auto dir change broke
14:05:42amiconnIirc it worked before MoB
14:05:59Nico_PI'm pretty sure it didn't... I'll check though
14:06:35Nico_Pplaylist_next() was always called in audio_check_nex_track (which is the codec track change)
14:06:38amiconnAs I find that one somewhat irritating, I think I would have noticed if it happened before
14:07:05 Quit moos ("Rockbox rules the DAP world")
14:07:12amiconnYeah, but for the metadata display, you are able to synchronise it with the pcm output. Why not for the playlist index?
14:08:02jhMikeSamiconn: any chance to provide a way to both align and pad graylib data to a specified boundary in a build as an option since you're reworking it? (by a #define)
14:08:26amiconn?
14:08:35Nico_Pamiconn: it breaks auto dir change. playlist_next is what handles it... if I call it on PCM track change, the codec won't decode the right track
14:08:45preglowjmspeex: but nothing more serious?
14:08:48jhMikeSamiconn: for use on COP in mpegplayer
14:09:47jmspeexpreglow: make sure the last frame is handled properly, but I'd say if it appears to work, it probably does
14:09:56jhMikeSI setup video_out_rockbox to do that. I don't need that stuff in IRAM, just cache aligned.
14:11:25jhMikeSpreglow: will check out what?
14:12:03amiconnYou can do that already
14:12:07Nico_Pamiconn: same thing on r15305 (i.e. MoB-1)
14:12:26amiconngray_init() allocates the necessary stuff in the buffer you pass it
14:12:29 Quit Buschel ()
14:12:41jhMikeSamiconn: but the _gray_info is a static
14:13:29amiconnah yes
14:13:41amiconnWhy do you want the graylib on cop, btw?
14:14:15 Join atsea- [0] (i=atsea-@gateway/tor/x-d49830169d637c59)
14:14:19amiconnThe new version will be a rewrite, introduced in parallel to the old one
14:14:43jhMikeSamiconn: because video is run on COP.
14:17:40amiconnmpegplayer works nicely as it is now...
14:18:49 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
14:18:51 Quit pondlife (Client Quit)
14:19:27jhMikeSonce you change anything, you can get suprise clobbering if data isn't explicitly aligned and padded. works now doesn't mean it will continue to do so if static data is allocated in a different order.
14:20:41preglowjhMikeS: meant to talk to jmspeex
14:20:52jhMikeSthis already happened to me...would abort when saving settings because data in configfile.c ended up in the way of the COP
14:20:54jhMikeSpreglow: ah
14:21:03preglowjmspeex: properly as in how?
14:21:15preglowjhMikeS: btw, do you skip the first few samples of all voice clips when decoding?
14:21:28jhMikeSpreglow: no
14:22:21jhMikeSwhy would I do that?
14:23:06preglowjhMikeS: because of encoder delauy
14:23:15preglowdelay, the first couple of samples are always zero
14:23:24preglowcheck out codecs/speex.c for how
14:24:13preglowjmspeex: btw, just feeding a narrowband file to the wb decoder should work fine, yeah?
14:25:16jmspeexyes
14:25:50jhMikeSSPEEX_GET_LOOKAHEAD?
14:26:21jmspeexbasically says how many zeros you need to discard.
14:27:38 Join BlindBricks [0] (i=519ee311@gateway/web/cgi-irc/labb.contactor.se/x-8f6c88320d86803a)
14:28:58BlindBricksSo if I install a different firmware file through ipod wizard
14:29:10BlindBrickscan I get rockbox again?
14:29:18BlindBricksI installed this guyes classic firmware
14:29:20BlindBricksand its nice
14:29:29BlindBricksbut I wondered if I can get rockbox to?
14:30:30jhMikeSjmspeex: but in our case process_header isn't needed since it's raw?
14:31:16jmspeexjhMikeS: the encoder's lookahead is how much zeros the encoder inserts and the decoder's lookahead is how much zeros the decoder inserts
14:31:26jmspeexnothing to do with headers or anything
14:31:34 Quit BlindBricks (Client Quit)
14:31:45jhMikeSok, good
14:31:48 Join UncleRemus [0] (n=caj@81-234-174-183-no95.tbcn.telia.com)
14:31:54jmspeexby "how much zeros", I basically mean the extra delay (on top of the frame size itself)
14:33:33amiconnHmm. Looks like the boost for PP in the graylib can't be removed
14:34:04jhMikeSjmspeex: this should not change unless the decode mode changes or do I need to check every clip?
14:34:07amiconn...because if I do, zooming and scrolling like mad in mandelbrot makes it hang
14:34:19preglowjhMikeS: unless the decode mode changes, no, but it costs nothing to check per clip
14:34:34preglowi am going to include a narrowband flag in the next rbspeexenc
14:34:37amiconnProbably the lcd bridge doesn't like the clock changing while it's active - and the graylib drives the lcd from an isr...
14:34:49preglowbut we're probably going to be using the wb mode anyway
14:34:55jmspeexpreglow, jhMikeS, note that there's *two* look-ahead to consider: encoder and decoder.
14:35:27jmspeexthe encoder's lookahead needs to be included in the file somehow. the decoder's is obtained directly with the _ctl() call.
14:35:42 Join bLIND-bRICKS [0] (i=519ee311@gateway/web/cgi-irc/labb.contactor.se/x-427cc1f10baddc64)
14:35:53jmspeexof course, that's *if* you want gapless.
14:36:20preglowwe most certainly do want that in our ordinary speex decoder
14:36:26preglowi haven't gotten around to doing that properly yet :/
14:36:33jhMikeSwell, clips are pasted together on the fly though I hadn't noticed a problem
14:36:35 Quit bLIND-bRICKS (Client Quit)
14:36:41 Join bLIND-bRICKS [0] (i=519ee311@gateway/web/cgi-irc/labb.contactor.se/x-aff79ca8547459c5)
14:36:51bLIND-bRICKSArgh my router kiled itself
14:36:59bLIND-bRICKSDid someone reply to me?
14:37:02bLIND-bRICKS;)
14:37:04preglowpills? shotgun?
14:37:16bLIND-bRICKSNo just British Telecom
14:37:19jmspeexpreglow: even more important, are you aware that seeking into a Speex file is a bit tricky and you actually need to start decoding a few frames back from where you're seeking ot?
14:37:32preglowjmspeex: not at all
14:37:36bLIND-bRICKSSo can I have rockbox working alongside a different ipod firmware file?
14:37:36preglowjmspeex: our seeking is dodgy at best, right now
14:37:57preglowi haven't really looked at it, i didn't code the original speex decoder wrapper, i've just been picking at it lately
14:38:13jmspeexpreglow: basically, Speex encodes frames based on the previous frames, so theoretically, you can't seek at all. In practice, if you start a few frames back, you're fine.
14:39:05jhMikeSjmspeex: what's the minimum compressed clip size I should consider valid?
14:39:26jmspeexwhat do you mean?
14:40:19jhMikeSif starting a new voice clip, how many bytes minimum should it be to not be rejected for playback?
14:40:45jmspeexwhy would you reject a small clip?
14:41:23jhMikeSI'd reject 0 bytes :) However I have it try to play anything >= 1 byte
14:41:46BigBambibLIND-bRICKS: Should be OK, yes
14:41:53jmspeexWell, technically, you can encode 20 ms of silence in just 5 bits :-)
14:42:11BigBambibLIND-bRICKS: Fiddle with the OF first, then install the rb bootloader
14:42:53jhMikeSjmspeex: I see, so that's valid then. thanks.
14:43:47bLIND-bRICKSSo I have to reinstall rockbox after instaliing the new OF
14:43:47pixelma[14:30:29] <BlindBricks> I installed this guyes classic firmware <- do you mean the Ipod Classic or was that just a descriptionß
14:43:57pixelmas/ß/?
14:44:00bLIND-bRICKSYes the ipod classic through ipod wizard
14:44:06BigBambibLIND-bRICKS: Just the bootloader
14:44:29BigBambipixelma: I think it is a classic like mod for previous models
14:45:06bLIND-bRICKSpixelma : yeah its the new classic firmware for ipod 5g
14:45:15pixelmaaha, thanks for clarifying.
14:45:22bLIND-bRICKSIts quite nice
14:45:30bLIND-bRICKSalthough they have not got cover flow yet
14:45:42*BigBambi doesn't care
14:46:12Soap_regardless, bLIND-bRICKS, that is simply a graphical chage of the Apple original firmware, and should not affect the installation of Rockbox
14:46:16jhMikeSpreglow: you did if(speex_bits_remaining() <= 8) but the codec checks for < 0. What's up?
14:46:17BigBambiI'm not too fond of ipods
14:46:21 Join lee-qid [0] (n=liqid@p54964F37.dip.t-dialin.net)
14:46:29 Quit Casainho ("ChatZilla 0.9.78.1 [Firefox 2.0.0.8/2007102213]")
14:46:37preglowjhMikeS: *shrug* :>
14:46:48jhMikeSwhich is right?
14:46:53preglowjhMikeS: i'm almost never as thorough as you, so if you find better ways of doing stuff, go ahead
14:46:56 Quit Soap (Nick collision from services.)
14:47:00 Nick Soap_ is now known as Soap (n=Soap@rockbox/staff/soap)
14:47:02preglowjhMikeS: both are probably good
14:47:03bLIND-bRICKSActually soap its a new firmware they managed to split the screen to show now playing etc
14:47:24 Join Soap_ [0] (n=Soap@rockbox/staff/soap)
14:47:24SoapbLIND-bRICKS: no - my original statement is correct.
14:47:25BigBambibLIND-bRICKS: what is the link?
14:47:40preglowjhMikeS: ours basically just cater to the fact that we might have seven bits left at the end of a decode and still be finished
14:47:45bLIND-bRICKShttp://driven-design.net/
14:47:47preglowjhMikeS: ours as in the voice decoder one
14:47:54bLIND-bRICKSThen click classic project
14:47:57SoapRockbox does not care what the origial firmware looks like.
14:48:00amiconnlinuxstb: Are you around?
14:48:01*n1s wonders why the buildsystem expects a build to take half an hour...
14:48:05bLIND-bRICKSOh
14:48:16bLIND-bRICKSI can't seem to get back rockbox thats all
14:48:55BigBambibLIND-bRICKS: From the site you posted "This is a modified iPod Video firmware, changed to look as close as possible to the new iPod Classics."
14:49:01Soapreinstall the bootloader, bLIND-bRICKS
14:49:07BigBambiIt is a mod, not a new firmware as Soap said
14:49:16bLIND-bRICKSOh okay thanks
14:50:19amiconnn1s: huh?
14:50:31bLIND-bRICKSAha now it works
14:50:33bLIND-bRICKSThanks guys
14:50:36jhMikeSpreglow: ok, so it's < 8 not <= 8
14:51:07n1samiconn: i get "Build expected to complete in 29m 6s, at 14:49:10"
14:51:15amiconnCheck your system clock...
14:51:17preglowjhMikeS: yeah
14:51:18 Nick bLIND-bRICKS is now known as Blind-Bricks (i=519ee311@gateway/web/cgi-irc/labb.contactor.se/x-aff79ca8547459c5)
14:51:34preglowjhMikeS: we'll never have eight bits left for a complete frame, though, so it will work
14:51:41amiconn"Build should have been done 2m 44s ago, at 14:49:10" here
14:52:02BigBambinls: given it is currentl XX:51, 51 mins + 29 isn't XX:49!
14:52:10n1samiconn: heh, right my clock is way off, I wonder why...
14:53:20jhMikeSpreglow: so we'll actually need our own mini clip header to tell encoder delays too?
14:53:24amiconnAnyone with a 4th Gen (grayscale or color this time) around?
14:53:32 Quit Blind-Bricks ("CGI:IRC (EOF)")
14:53:46 Quit donutman25 (Read error: 104 (Connection reset by peer))
14:54:33preglowjhMikeS: no, encoder delays are constant
14:54:39preglowhmm
14:54:51preglowjmspeex: if we decode nb clips with wb_mode, will the lookahead value still be right?
14:55:16preglowit should, shouldn't it? the delay appears in the pitch prediction stage anyway, and only nb does that?
14:55:38jmspeexpreglow: lookahead has nothing to do with pitch prediction.
14:55:50jmspeexThe value you'll get for wideband will be different.
14:56:12preglowok, so if we decode nb clips with wb_mode, the lookahead value will be wrong and we'll cut too much?
14:56:25jmspeexAt the decoder size, it won't matter at all because the wb decoder will still give you the right value regardless of how the file was encoder.
14:56:31preglowahh, right
14:56:38amiconnDoes that matter at all? How many samples we're talking about?
14:56:42preglownot many
14:57:00jmspeexOTOH, you'll need to change the encoder lookahead value to account for the fact you're decoding a narrowband file in wideband mode.
14:57:21jmspeexamiconn: we're talking in the order of 10-15 ms
14:57:40 Quit Seed (Nick collision from services.)
14:57:46 Join Seed [0] (i=ben@bzq-84-108-237-178.cablep.bezeqint.net)
14:58:05preglowjmspeex: so, what exactly would i need to do extra in rbspeexenc.c to compensate for that?
14:58:13*amiconn will commit blindly and wait for complaints this time
14:58:21preglowamiconn: commit what?
14:58:33amiconnIt's not a big change, just moved backlight init for 4th gen where it belongs
14:58:36preglowjhMikeS: actually, we _might_ want to add a small header to tell the decoder if the clip is nb or not
14:58:42amiconn(was in lcd init...)
14:58:49*jhMikeS worried about underrunning a fream and assumes nsamples - lookahead now.
14:58:49preglowjmspeex: hmm, or doesn't the first bit of any frame tell me if the clip is nb or not?
14:59:03jmspeexpreglow: you'd need to double the lookahead when converting from nb to wb
14:59:10 Join Rondom [0] (n=Rondom@p57A945DB.dip.t-dialin.net)
14:59:23jhMikeSnsamples = 320 always (static decoding buffer)
14:59:32jmspeexpreglow: technically, a nb stream is also a perfectly valid wb stream
14:59:53preglowjmspeex: yeah, i noticed the wb decoder just generates a zero upper band if the file is nb
15:00
15:00:26 Quit webguest36 ("CGI:IRC")
15:01:08jmspeexpreglow: and that's perfectly valid. You could have a VBR wb stream that switches to nb in some periods
15:01:13jhMikeSjmspeex: will all lookaheads ever add up to the first frame or more discarded?
15:01:18preglowjmspeex: oh, ok
15:01:24jmspeexjhMikeS: no
15:02:20jmspeexjhMikeS: well, not with the current implementation. You could have an implementation for which that's the case.
15:02:57 Quit Seed (Nick collision from services.)
15:03:01 Join Seedy [0] (i=ben@bzq-84-108-237-178.cablep.bezeqint.net)
15:04:38jhMikeSok, so better just plan for it and avoid crashing later if that changes :)
15:04:48preglownot a bad plan
15:04:49 Join Seed [0] (i=ben@bzq-84-108-237-178.cablep.bezeqint.net)
15:06:08preglowjmspeex: so to reliably decode a nb stream with an nb decoder, we'd have to add our own little header to tell if it's a true nb stream or not
15:06:36 Quit Frazz ("Leaving")
15:07:04jhMikeSmuch better than obscure bugs at cost of a MIN macro
15:09:04 Quit Riggzy ()
15:09:23***Saving seen data "./dancer.seen"
15:09:24 Join XavierGr [0] (n=xavier@ppp176-211.adsl.forthnet.gr)
15:14:23jhMikeSstill get "Loading, Loading" when loading a plugin...but consistently on all targets :)
15:14:48 Join handmadematters [0] (n=handmade@80.224.161.24.dyn.user.ono.com)
15:14:51 Join joshin [0] (n=josh@unaffiliated/joshin)
15:15:27 Join qweru [0] (n=kvirc@bb-87-80-66-156.ukonline.co.uk)
15:15:31 Join MethoS-- [0] (n=clemens@pD955C9E2.dip.t-dialin.net)
15:17:25 Join PaulJam [0] (i=PaulJam_@vpn-3026.gwdg.de)
15:21:05 Quit handmadematters ()
15:22:13Learn1s: he, you were quick on picking up that function...
15:23:09n1sLear: I just remembered that place in the playlist viewer, iirc it was my first patch accepted into cvs :)
15:25:52 Quit Seedy (Read error: 110 (Connection timed out))
15:26:08amiconnJdGordon: around?
15:26:19JdGordonbarely
15:26:42JdGordon... btw freeky... i opened the irc window just as you typed that!:p
15:27:24amiconnhehe :)
15:28:01amiconnDid you have an idea how to nicely add integer settings using a table?
15:28:27JdGordonno, I havnt thought about it properly since we discussed it
15:28:46*jhMikeS looks at the Rockbox Info screen and sees it's -2:26am :)
15:29:05JdGordon-2? wtf?
15:29:17 Quit MethoS- (Read error: 110 (Connection timed out))
15:29:23jhMikeSIt talks 10:26a correctly though
15:29:46 Quit Daniel_S ("CGI:IRC")
15:30:59JdGordonamiconn: just thought of this now... for the table we should generate the list on the fly so we can put the current value into the list which could save some hassle when users wonder why it selects 45 when the config says 33
15:31:00amiconnJdGordon: Hmm, would be nice if that gets added. Allows to get rid of duplicate tables, for better flexibility
15:31:47amiconnHmm, sounds like a plan... do you think that would be possible in an efficient manner?
15:31:50 Nick fxb__ is now known as fxb (n=felixbru@h1252615.stratoserver.net)
15:31:55JdGordonanother thing which might be better is having a button to change the step size? so start at step=1, press the button step=5, press again step=10.. etc
15:32:28n1shmm, i guess the imx32 chip in those new zunes is related to the imx31 in gigabeat s and old zunes http://www.rapidrepair.com/guides/zune2/IMG_0841.JPG
15:33:10JdGordonyeah, should be able to do it with no extra buffers actually
15:34:11JdGordonamiconn: if i remember ill have a go at this tomorow (today... monday...), its 1.30am and i need my sleep
15:34:41n1sand oh, I see the freescale logo now :)
15:35:14 Quit JdGordon ("Konversation terminated!")
15:37:10 Quit Lear ("ChatZilla 0.9.78.1 [Firefox 2.0.0.9/2007102514]")
15:37:18 Quit BigBambi (Remote closed the connection)
15:40:47 Join bluey [0] (n=bluey@dslb-088-073-082-041.pools.arcor-ip.net)
15:41:16 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
15:44:19preglowjhMikeS: cpu usage isn't that big a point right now, is it?
15:44:28preglowjhMikeS: don't think i'll bother with a specific nb decoding mode
15:45:10jhMikeSit's not an issue at all
15:45:58jhMikeSno IRAM output buffer on coldfire with pcm ll during playback is just fine
15:48:38 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
15:48:51jhMikeSpreglow: are we going to add a small header to give band info or anything else?
15:48:59 Join miepchen^schlaf [0] (n=hihi@p54BF7BA1.dip.t-dialin.net)
15:49:32preglowjhMikeS: our current scheme will decode narrowband just fine, so the only point would be to save some cycles and output buffer
15:49:35preglowjhMikeS: is it worth it?
15:50:23jhMikeSunless we always use nb, I'll still need the output buffer or it could be leased from the main buffer
15:50:36preglowwb is what we'll use by default, and do now
15:51:47jhMikeSyeah, so if wb mode decodes nb just fine...forget it. I'm still not clear on the lookahead issue for different encoding and whether the voice thread needs to care.
15:51:55preglowi'll figure that out
15:51:58preglowyou shouldn't need to care, i think
15:52:08preglowif i understood jmspeex correctly, that'd be an encoder issue
15:57:32 Quit amiconn (Nick collision from services.)
15:57:41 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
15:57:50preglowanyway, did you fix up the voicebox stuff?
15:59:51jhMikeSamiconn was the one that pretty much had that done so I thought he was going to finish up on that.
15:59:56preglowmokey
16:00
16:15:40 Quit pixelma (Read error: 110 (Connection timed out))
16:23:08jhMikeSIs it really _that_ important it not lag a few days? Just put a notice on the wiki page. It's not even hosted on a rockbox site or in the svn tree is it.
16:24:00 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
16:27:48 Quit MethoS-- (Remote closed the connection)
16:28:25preglowamiconn: there?
16:28:40preglowjhMikeS: point is, there's not much point in us commiting this if people can't use it anyway
16:29:03preglowanyway, i'm hoping on having stuff ready to go today
16:29:09preglowi think i'm going to let rbutil lag
16:29:15preglowpeople seem to use voicebox more anyway
16:29:32jhMikeSmmkay
16:30:01preglowit really would be nice to get this out the door
16:30:46jhMikeSquite
16:31:56preglowso ok, the second voicebox is ok, we have someone stuff a notice on the front page and commit
16:32:06 Join moos [0] (i=moos@m147.net81-66-159.noos.fr)
16:33:01preglownot necessarily in that order
16:33:44jhMikeSit's on?
16:34:38*jhMikeS doesn't wanna get served :\
16:35:39preglowvoicebox, then commit
16:35:45 Join thgz [0] (n=thgz@dsl-lprgw5-fe5fdc00-169.dhcp.inet.fi)
16:36:04preglowlet's just hope amiconn is present today
16:36:05jhMikeSdonde esta' el voicvox, ya?
16:36:39 Quit jhulst ("Konversation terminated!")
16:36:46jhMikeShe was here...decided to slip out the door when the moment approached. :)
16:36:47*preglow doesn't do spanish :/
16:37:18*n1s actually does enough spanish to get that :)
16:39:16*jhMikeS wonders if amiconn doesn't have a custom voice building system and really has to fix all that first
16:39:28preglowjhMikeS: but yeah, you can't think of any good reason to add a header then, no?
16:39:40thgzhi
16:40:08jhMikeSpreglow: not at this point
16:41:27thgzWhenever I try to create a voice file using cygwin, I get an error which reads: "libspeex.a: could not read symbols: Archive has no index; run ranlib to add one"
16:42:07jhMikeSis the sim perhaps?
16:42:09jhMikeS*in
16:42:37preglowehm
16:43:01jhMikeSwere symbols made private by default? *just a musing*
16:44:08thgzDoes anyone know how to get this working?
16:44:15preglowthgz: i'm on it
16:44:21jhMikeSthgz: what build?
16:44:44thgzI'm using the latest from svn
16:44:58jhMikeSthgz: not sure if you said but is it a sim build?
16:44:59n1sthgz: 64 bit?
16:45:24thgzn1s: No it isn't
16:45:28preglowfound the bug
16:45:54preglow$(AR) is defined to be target ar
16:46:01thgzjhMikeS: I don't know.
16:46:05preglowshould i just replace $(AR) with "ar" in rbspeexenc makefile?
16:46:57thgzI can try
16:46:58jhMikeSsounds like ranlib would only be a sim thing...bah
16:47:07preglowjhMikeS: you don't need ranlib, ar can do that part for you
16:47:10preglowthgz: that does work
16:47:17preglowi just wonder if it's the proper way to do it
16:47:30preglowi'll just commit that for now, it fixes it
16:47:32jhMikeSwhat's ar mean?
16:47:46thgzPreglow: Can you explain where is the file that I should open?
16:47:48jhMikeSarchive?
16:48:02Bagder"ar - create, modify, and extract from archives"
16:48:26preglowthgz: i've fixed it in svn now
16:48:37thgzokay
16:48:49thgzI download it again.
16:48:49preglowthgz: so just do "make clean" in tools/rbspeex
16:48:56preglowthgz: then go at it like usual
16:49:16thgzok
16:50:15thgzShould I re-download the src?
16:50:20jhMikeSthere's too many libspeex makes that now so I have no idea which one somebody might be referring to
16:51:16jhMikeSespecially when still half awake
16:51:37preglowthe rbspeexenc one
16:52:09jhMikeSI figured that out...but really slowly :P
16:52:50 Join hunz [0] (n=hunz@dslb-084-056-009-123.pools.arcor-ip.net)
16:53:02preglowheh
16:53:41thgzIt start compiling the file
16:54:04thgz*started
16:57:05 Join Riggzy [0] (n=Riggzy@78.150.118.50)
16:57:24*jhMikeS still wonders why "Yes" gets played as a silence clip instead of "Yes"
16:57:46preglowi wouldn't know, but i sure did notice
16:57:58thgzgreat
16:58:03RiggzyIs there a known bug with the sleep timer? Specifically on video ipods if that's a factor
16:58:03thgzIt compiled successfully!
16:58:06thgzThanks much!
16:59:21 Part bluey ("Leaving")
17:00
17:00:25 Part thgz
17:01:23 Join jurrie__ [0] (n=jurrie@adsl-068-209-041-021.sip.asm.bellsouth.net)
17:01:29RiggzyThe sleep timer gets to 0:00 then goes into negatives... x.x
17:04:37PaulJamRiggzy: does it continue to count negative for a longer time or does it shut down shortly after crosing 0:00?
17:05:25 Quit hunz_ (Read error: 104 (Connection reset by peer))
17:05:45*Riggzy peeps at the screen, "Going on 12 minutes into negatives."
17:06:08RiggzyDoesn't look like anything's stopping it
17:06:09 Join Robin0800 [0] (n=Robin080@cpc1-brig8-0-0-cust97.brig.cable.ntl.com)
17:06:16preglowwell, you could
17:06:18RiggzyHowever, it's currently charging
17:06:43PaulJamthis is certainly not normal
17:07:09RiggzyLet's see what happens when it's not charging *pop*
17:07:21RiggzyAh, as soon as I unplug it, it shuts down x.x There's the bug
17:07:24PaulJami think i have notices something similar on my h300, but the shutdown was only delayed by a few seconds
17:08:15RiggzyIt seems that if I'm charging the iPod (over USB with a computer) it's not going into sleep
17:08:49 Quit Nico_P (Remote closed the connection)
17:08:56PaulJamRiggzy: this is still a bug, while charging playback should stop and display be turned of (at least on my playeer it behaves that way when on the charger).
17:09:26***Saving seen data "./dancer.seen"
17:09:28RiggzyI read something about USB support with the iPod being a bit flakey still
17:09:38 Quit jurrie_ (Read error: 113 (No route to host))
17:09:41RiggzyWhat're you running it on Paul?
17:09:50PaulJamlet me see what happens on my devich when charging from USB
17:10:05PaulJami use iriver h300
17:10:14RiggzyAh right
17:10:14 Join FunkyELF [0] (n=funkyelf@105.149.243.24.cfl.res.rr.com)
17:10:46RiggzyWell, a track just skipped, guess I'll just have to wait for this bug fix like all the other current iPod ones
17:11:27FunkyELFdoes rockbox work on the new iPod classic? do I just follow the video ipod docs?
17:11:39PaulJamRiggzy: is there already a bugreport on flyspray? if not, you should open one (if you want the bug to be fixed).
17:11:40n1sFunkyELF: no, doesn't work
17:11:51 Quit lee-qid ("aufwiederbyebientotsayonara")
17:12:09FunkyELFn1s, doesn't work yet you mean? will it work eventually?
17:12:15RiggzyI was trying to find the Rockbox bug report site *chuckles
17:12:29n1sFunkyELF: I mean it doesn't work, i can't predict the future.
17:12:32preglownot really hard to find, is it
17:12:39RiggzyIt's rather disheartening to google "rockbox bugs" and get a random mail archive
17:12:57preglowRiggzy: if you can't find the bug report page, you're not looking
17:12:58preglowwww.rockbox.org
17:12:59preglowit's there
17:13:22RiggzyI'm silly and use Google instead of webpages sometimes *facepalm*
17:13:34preglowoften works, sometimes doesn't
17:13:43preglowi'm a bit like that myself
17:16:38PaulJamah interesting, i can confirm that the sleeptimer is broken when charging from USB.
17:17:27RiggzyEenteresting
17:17:36RiggzyCan't see anything specific on the tracker..
17:19:10 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
17:19:18 Join SolidSnake [0] (i=FakeNam@dialup-4.227.124.192.Dial1.Dallas1.Level3.net)
17:19:27 Join petur [0] (n=petur@rockbox/developer/petur)
17:19:28SolidSnakehmm
17:19:34SolidSnakeIs anyone alive
17:19:41SolidSnakeI have some questions
17:19:45BigBambiJust ask
17:19:46Nico_PI am
17:19:46SolidSnakeBeyond the Standard FAQ
17:19:48SolidSnakeok
17:20:01BigBambiSolidSnake: Beyond the manual too?
17:20:02SolidSnakeI have a Turly F602 MP3 Player
17:20:06SolidSnakeI know it won't work
17:20:20SolidSnakebut is there a possibily that it might?
17:20:27SolidSnakeBigBambi:
17:20:28SolidSnakeNo
17:20:31SolidSnakeok
17:20:35SolidSnakeI'll Read first
17:20:47BigBambiSolidSnake: Check out the NewPorts wiki page
17:20:54n1sSolidSnake: if you port rockbox to it sure there's a possibility
17:21:22SolidSnakeok
17:21:24SolidSnakeThank You
17:21:29SolidSnakeAnother thing
17:21:38SolidSnakeMy Brother has a Sansa E200R
17:21:42SolidSnakeIt SHOULD work
17:22:01SolidSnakebut if update it, will it still work on Windows 98
17:22:05SolidSnakeW/O Driver
17:22:07SolidSnake?
17:23:08n1sSolidSnake: you have to use sandisk firmware for usb so it will work the same as before
17:23:53SolidSnakeok
17:24:03 Join lazka [0] (n=lazka@83-65-235-215.dynamic.xdsl-line.inode.at)
17:24:15 Join japc [0] (n=japc@bl8-7-67.dsl.telepac.pt)
17:26:30SolidSnakeIs there any other Opensource/Closedsource MP3 Firmware Projects?
17:26:50DerPapstiPodLinux
17:26:58DerPapstand some others too afaik
17:27:13SolidSnakehmm
17:27:14n1sand a couple others like archopen but that is OT here
17:27:27moosSolidSnake: please, we speak rockbox here!
17:28:09SolidSnakeSorry
17:28:13SolidSnakeI was just asking
17:28:39moosno problemo
17:30:22SolidSnakeWow
17:30:25SolidSnakethis is great
17:30:41SolidSnakeI can't wait until someone ports it to Truly Line of MP3 PLayers
17:31:13moosSolidSnake: just buy one supported device ;)
17:31:25SolidSnakelol
17:31:27SolidSnakeNice
17:31:47SolidSnakeand throw away my $119 Turly FT602?!?
17:31:50SolidSnake;)
17:32:05DerPapstor sell it ;)
17:32:24SolidSnakeYeah, maybe Ebay it
17:32:42SolidSnakeOr I can Learn Programming and Port Rockbox Myself
17:32:45 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
17:33:18markunSolidSnake: or just be happy with the current features of your player. What´s missing from it=
17:33:52moosRockbox !
17:33:54moos:)
17:34:04mooshi markun btw
17:34:08SolidSnakewell it Has a skinning engine
17:34:11 Quit linuxstb (Nick collision from services.)
17:34:13 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
17:34:17SolidSnakebut Truly has only 3 Skins for it
17:34:20linuxstbamiconn: I'm around now.
17:34:35SolidSnakeAnd I can't make my Own becasue it has this .res extension that nothing can open
17:35:00amiconnlinuxstb: I committed the stuff meanwhile. Would be nice if you could do a quick check of latest svn on your Color
17:35:11markunmoos: hi!
17:35:28amiconnI don't really expect problems though. I just moved the backlight init where it belongs - out of lcd_init_device()
17:35:40markunSolidSnake: that sounds like a great reason to port rockbox to it! :P
17:36:11SolidSnakehey thats not the only problem
17:36:15SolidSnakeIt Crashes
17:36:30SolidSnakeYou expect something like that to stay in Windows OS
17:36:46SolidSnakeWhen you scan the playlist sometimes it just freezes up
17:36:53linuxstbamiconn: No problem. So I'm just checking the backlight is still working?
17:37:03markunSolidSnake: rockbox also crashed 2 times on me this weekend..
17:37:06SolidSnakeThere are some serious problems with it...
17:37:11SolidSnakemarkun why?
17:37:24markunSolidSnake: because it contains bugs?
17:37:30SolidSnakeI know
17:37:44 Quit Riggzy (Remote closed the connection)
17:37:45*linuxstb looks at the front page, and wonders what "oops, forgot a few" means...
17:37:47SolidSnakebut what were you doing that casued the crash, or did it just freeze on its own
17:38:03amiconnlinuxstb: Yes, basically. It would also be nice if someone with a 4th Gen (grayscale or color) would investigate the PP502x PWM for backlight brightness...
17:38:35 Join webguest70 [0] (i=4cc18e54@gateway/web/cgi-irc/labb.contactor.se/x-e6460af4e9284fb9)
17:38:40amiconnI tried it on my mini (maybe apple put another hidden feature like brightness on Nano), but that doesn't work
17:39:27n1slinuxstb: you're reading it out of context! :P
17:39:40linuxstbn1s: I think that's my point ;)
17:39:49 Join grujilj [0] (i=4cc18e54@gateway/web/cgi-irc/labb.contactor.se/x-79eb044f9954dd98)
17:39:49 Quit webguest70 (Client Quit)
17:40:51 Quit grujilj (Client Quit)
17:41:34preglowamiconn: ahoy, how's voicebox coming along?
17:41:42preglowwe want this stuff commited, it seems :>
17:41:43amiconnStill the same state
17:42:07amiconnGuess I should put it on the wiki as-is, and let others who want this drag-and-drop working care of it
17:42:37preglowamiconn: sounds good to me
17:42:39preglowjhMikeS: ready?
17:42:42 Part SolidSnake
17:42:49jhMikeSaye
17:42:56jhMikeSmon capitan
17:43:01preglowamiconn: so i should just tell people not to expect drag and drop to work?
17:43:40jhMikeSplop a TODO note on the page detailing the situation?
17:44:40jhMikeSpreglow: exactly what is your part?
17:44:48preglowjhMikeS: just configure and voice.pl, really
17:44:53jhMikeSok
17:45:27jhMikeSI guess I get the bin delta blame :P
17:45:39preglow:D
17:46:52preglowi can live with that
17:47:48linuxstbAm I right in thinking voicebox has no license?
17:48:03preglowlinuxstb: not exactly desirable, but you are correct
17:48:08preglowi'd like it if voicebox was in svn
17:48:16preglowand most definitely gpl
17:48:21jhMikeSpreglow: give the countdown. ready to launch here.
17:48:27preglowjhMikeS: just waiting for amiconn
17:48:39jhMikeSwhat's he gotta do now? :(
17:48:40linuxstbHas anyone tried to contact the original authors?
17:48:56preglownot in the least
17:49:03preglowjhMikeS: just stuff voicebox in the wiki
17:49:33preglowBagder: it would also be cool if a notice in the new voice file format could be put on the front page news or something
17:49:39preglows/in/of/
17:50:40preglows/of/on/ :>
17:51:06linuxstbamiconn: No problems on my Color.
17:51:22amiconnThanks :)
17:52:04 Join PaulPosition [0] (n=noneofye@modemcable228.133-82-70.mc.videotron.ca)
17:52:11amiconnI'm just updating a few things in voicebox
17:52:24preglowsure
17:52:44amiconnThe supported extension list grew considerably since voicebox was made...
17:53:12*jhMikeS 's launch finger gets twitchy
17:53:20 Join mf0102 [0] (n=michi@85.127.180.92)
17:54:14linuxstbHmm, seems with dircache enabled, my Color freezes when booting with USB connected, but if I disable dircache, it works fine...
17:56:50amiconnpreglow: Are there recent changes in rbspeexenc which are needed? My rbspeexenc is dated 16-Nov-2007 22:03
17:58:37preglowamiconn: i don't think so, no
17:59:05preglowamiconn: maybe one thing, yes
17:59:10preglowamiconn: a read that was supposed to be fread
17:59:25amiconnOkay, I'll rebuild it
17:59:45 Quit idnar (Nick collision from services.)
17:59:48 Join idnar_ [0] (n=mithrand@unaffiliated/idnar)
18:00
18:00:17amiconnHmm. The only thing it did was making a libspeex.a ...
18:00:37linuxstbrbspeexenc is created directly in tools/
18:00:49amiconnI know
18:01:24amiconnBut the read->fread change happened before I've built rbspeexenc, so it should be fine
18:01:41 Nick idnar_ is now known as idnar (n=mithrand@unaffiliated/idnar)
18:01:57preglowok, then we're ok
18:03:11 Join Frazz [0] (n=Fraser@thelawsons.plus.com)
18:04:42Nico_Pwhat's voicebox?
18:07:01Nico_Phmm actually I think I found... didn't know about it
18:08:18preglowamiconn: it's up?
18:09:25amiconnYes, now including some notes
18:09:26 Join nanok [0] (n=nanok@194.145.183.75)
18:09:32amiconnhttp://www.rockbox.org/twiki/bin/view/Main/VoiceHowto#Automated_generation_of_talk_cli
18:10:26preglowamiconn: excellent, thanks a bunch
18:10:28preglowjhMikeS: then we go
18:10:34jhMikeSready?
18:10:50preglowsure
18:10:55jhMikeSone last change to tell me
18:10:59jhMikeSchance
18:11:05amiconnHow large will a typical swcodec voice file be now?
18:11:14preglowamiconn: dunno, mine is 8kk
18:11:16preglow800k <-
18:11:23amiconnNanO?
18:11:28preglowh120
18:11:30amiconnah
18:11:43linuxstbpreglow: Regarding your AR fix to the rbspeex Makefile, I think it might be nicer to deal with it the same way as CC - i.e. create a HOSTAR variable in configure, similar to HOSTCC.
18:11:58jhMikeSpreglow: ok? 2nd last chance.
18:12:09preglowjhMikeS: go <-
18:12:10preglow:P
18:12:16preglowlinuxstb: sure, but no one told me so at the time :P
18:12:19jhMikeSlaunched...
18:12:28*amiconn wonders if it will work right away :>
18:12:45preglowme too
18:12:46preglowheh
18:13:13linuxstbpreglow: I'll do it now.
18:13:48jhMikeS...and n1s has a build waiting to complete...>:[
18:15:31jhMikeSwell, good syncing anway :)
18:15:34amiconnn1s: backlight_timeout is expected to be present on all targets. backlight_timeout_plugged is *not*
18:16:05amiconnIt depends on CONFIG_CHARGING, i.e. Ondios don't have it
18:17:16amiconnThe reason why backlight_timeout exists for all targets is that the Ondio can be modded to have backlight, and we do not want the plugin api to be incompatible between modded and unmodded builds
18:20:36*Nico_P is thrilled to see all the deletions in playback.c
18:22:04 Join radinp [0] (n=pradin@psyduck-08.dynamic.rpi.edu)
18:22:14*preglow is thrilled to finally have this commited
18:22:42 Quit radinp ("Leaving.")
18:23:06jhMikeSbeh...got some warnings
18:23:12preglowi wonder how casting QStrings to char * will work
18:23:12amiconnDa big red delta...
18:23:23preglowamiconn: i'm going to work a bit on shrinking it
18:23:28 Join gamerdonkey [0] (n=gamerdon@adsl-76-223-195-214.dsl.emhril.sbcglobal.net)
18:23:42preglowbut all in all i don't think this delta is too bad, considering we now have an entire codec in the core
18:23:51amiconnYeah, I know
18:24:12preglowbut i expect to be able to shave another few kb off it
18:24:24amiconnjhMikeS: One extra warning for sim builds
18:25:47preglowjhMikeS: looking forward to that dual core spc codec, then :>
18:26:27*amiconn would prefer to see some dual core work happening for PP5002 :>
18:26:39jhMikeSyeah, working on that
18:26:57preglowwhat, pp5002 didn't get the dual core treatment some time ago?
18:26:59jhMikeSthat SPC codec is ready, just needs some CI entries
18:27:03gamerdonkeyhey, would an ipod 3G question be too annoying?
18:27:22preglowgamerdonkey: ask and find out, as long as its about rockbox, you won't be beat up
18:27:22jhMikeSpreglow: swp(b) is broken there too and that's really the issue
18:27:33jhMikeSbut it's no issue now
18:28:11amiconnCache handling is the other issue.. or did you already do the necessary re?
18:28:45jhMikeSnot yet. why is it running 1/2 speed? the setup or the HW?
18:28:54amiconnThat's a HW bug
18:29:14gamerdonkeywell, i just installed rockbox on the ipod and, despite the battery being fully charged and fine in the original apple software, it says the battery is drained and shuts down
18:29:16amiconnBut we need to find out the necessary bits for flushing/invalidating
18:29:19jhMikeSI suppose if 1/2 speed is faster than no cache, it's an improvement
18:29:23gamerdonkeyhave you ever heard of this before?
18:29:30preglowamiconn: are we absolutely sure about that, btw? is it possible it _could_ be a setting?
18:29:32amiconnjhMikeS: It clearly is
18:29:47 Nick Snake_ is now known as snake (n=Snake@adsl-68-76-119-91.dsl.bcvloh.ameritech.net)
18:30:27amiconnReally time critical functions need to go into iram on P5002. That has nothing to do with dual core though
18:36:21*amiconn tries to build a german voice using at&t
18:37:18*jhMikeS notices there's still that delayed delta bug in the table
18:37:34amiconnpreglow: It errors out...
18:38:13*gamerdonkey takes that as a no
18:38:18gamerdonkeythanks anyway, guys
18:38:18amiconnEh, you're passing absolute paths?
18:38:29amiconnThat's boud to break...
18:38:34amiconn*bound
18:39:32linuxstbgamerdonkey: Yes, that's a no. We don't have very many 3rd gen users though (and no developers I think at the moment...)
18:39:50Nico_PjhMikeS: now we can think about improving the codecs by giving them a message queue :)
18:40:51jhMikeSNico_P: yeah, no doubt.
18:40:55 Quit z35 ("Leaving")
18:41:34 Join z35 [0] (n=z@149.123.33.65.cfl.res.rr.com)
18:41:39gamerdonkeylinuxstb, yeah, that's understood
18:41:52preglowamiconn: what? were?
18:42:00preglowwhere, that is
18:42:31amiconnThe rbspeexenc call fails, because you're passing an absolute cygwin path to the vbscript
18:42:46preglowi haven't even touched the vbscript
18:43:03preglowi can't even run it
18:43:21Nico_Pwhy are some functions from voice_thread.h called mp3_* ?
18:43:56amiconnpreglow: No, but you added an absolute enocder path somewhere...
18:44:19preglowamiconn: i use a relative encoder path in configure, i think
18:45:02preglow ENC_CMD="$rootdir/tools/rbspeexenc"
18:45:07preglowi think that's the only directory handling i do
18:45:20preglowdoes voicebox even interact with those values? bear in mind i can't even run it and have no idea how it works
18:46:00 Quit gamerdonkey ("Leaving")
18:46:05amiconnI'm not talking about voicebox
18:46:15amiconnI'm talking about generating a voice file with sapi
18:46:35preglowwell, ok, the same applies there, pretty much
18:46:37preglowwhere does it break?
18:46:44amiconnAnd the Makefile generated by configure contains the absolute path
18:46:57preglowis $rootdir absolutele on cygwin?
18:47:01preglowabsolute....
18:47:09amiconn...which does work fine within cygwin, but the vbscript doesn't understand that path at all
18:47:46 Join pixelma [0] (n=Marianne@rockbox/staff/pixelma)
18:48:00amiconnI think the most compatible way would be to (temporarily) add the tools dir to the path, and then call rbspeexenc the same way as lame - without any path specifier
18:48:22amiconnHow could I add a path to the system path within a perl script?
18:48:30preglowhmm
18:49:21 Quit advcomp2019 ("Leaving")
18:49:23preglowlet's see
18:50:15amiconnThis would also simplify calling wavtrim and voicefont
18:50:22jhMikeSthis scheduler boost control would simply be better if a thread called say, cancel_cpu_boost() when it wants to unboost instead of it having to do with timed waits (which I see as irrelenvant to the decision). basically voice handles it that way now by sleep(0).
18:50:24preglow$ENV{PATH} .= ";newpath" ?
18:50:47billeniumBRB
18:50:52 Quit billenium ("Ex-Chat")
18:55:19preglowamiconn: i can't really remember how perl exports path to programs it spawns
18:56:01preglowamiconn: is this what the last ml post is talking about, btw?
18:57:59amiconnJust manipulating $ENV{'PATH'} works within the script
18:58:15amiconnNow I have to try whether cygwin manages the path passing properly
18:58:44amiconnThen we can just add the tools dir to the path temporarily, and not worry about the paths anymore
18:59:00preglowwhich would be great
19:00
19:00:41 Join AceNik [0] (n=AceNik@117.98.14.233)
19:00:57AceNikhey guys i cant cmpile on vmware anymore
19:01:08preglowwhat's wrong?
19:01:16amiconn$ENV{'PATH'} = dirname($0) . ':' . $ENV{'PATH'};
19:01:44AceNiksvnversin.sh & makeinfo.pl error 127 when givin make command
19:02:30preglowehh
19:02:57AceNikamiconn shpould i enter the above command
19:03:25amiconnno, that doesn't have to do with your prob
19:03:28 Join stevenm [0] (n=stevenm@129.2.200.154)
19:04:13AceNikthen i had recently updated the vmware player to versin 2.02 then copied the source back to the virtual drive fter that when i tried compiling i et this error
19:05:03 Join me [0] (n=me@bas12-toronto63-1088801894.dsl.bell.ca)
19:05:09lostlogicwoah... no more codec swapping!? I might faint.
19:05:51meHow do I clear the database? When I update it, it does not properly reflect the changes I made to the ID3 tags.
19:05:57 Quit stevenm (Client Quit)
19:08:39pixelmaNico_P: thanks for the fix, seems to work ok now. There is just a short delay when "restoring" the text after the bitmap was displayed but a lot better than ... nothing :)
19:09:26Nico_Ppixelma: you're welcome :) I know about the delay, but I don't think there's much I can do about it
19:09:30***Saving seen data "./dancer.seen"
19:09:49pixelmame: you could initialise it again...
19:10:13Nico_Plostlogic: isn't it awesome that playback.c is now only ~2600 lines? :D
19:10:22amiconnpreglow: PATH fiddling works fine... voice generation seems a bit slower than with lame (but quite okay)
19:10:29preglowamiconn: it slower than with lame
19:10:35preglowamiconn: you can drop -c to make it faster
19:10:44preglowamiconn: but cool anyway!
19:10:49preglowplease check that i works on target as well
19:10:55amiconnI'll test hwcodec as well, then commit
19:11:01preglowexcellent
19:11:17lostlogicNico_P: shock, awe.
19:11:20pixelmaNico_P: ah and the wrong next track info during the last track in the playlist is gone now too :)
19:11:27AceNikcan anyone help i get permission denied in svnversion.sh & makeinfo.pl then a error 127
19:11:32Nico_Ppixelma: yeah, fixed it too :)
19:11:34amiconnI can't drop all dirname($0) calls though, because the tools path is needed for some more things than just commands
19:11:40 Quit Thundercloud (Read error: 104 (Connection reset by peer))
19:11:56amiconnThat's all internal to voice.pl and works regardless of whether it runs on linux or cygwin
19:12:22amiconnWow, voice file size for mini G2 went 1151KB->736KB
19:12:30amiconn(same tts parameters)
19:12:48preglowamiconn: and it still sounds quite nico,imo
19:13:01linuxstbAceNik: Sounds like the file attributes are wrong - how did you download the source code?
19:13:39Nico_Plostlogic: playback.c was 3902 lines before MoB... 2625 now... this is pure awesomness IMO :)
19:13:56Nico_Pmaybe even legendary
19:14:21AceNiklinuxstb: i just copied the folder that i had earlier n recopied it back to the new vmware drive
19:14:38linuxstbAceNik: Ah, I've just read your earlier posts... Sounds like the copy broke the permissions....
19:14:51n1slostlogic: could you rebuild arm-elf-gcc on your buildserver with rockboxdev.sh so it gets the needed patches?
19:15:12linuxstbAceNik: If I was you, I would just delete the tools directory and do an "svn update"
19:15:21n1sor patch manually if that's your cup of tea
19:16:03amiconnn1s: Did you see my remark regarding backlight_timeout_plugged?
19:16:22AceNiklinuxstb:but i get an error in the apps directory
19:16:39AceNiklinuxstb; ill try ur method
19:17:49 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
19:18:35n1samiconn: ah, yes forgot about it bu I'll fix now
19:19:05 Quit DerPapst ("So Long And Thanks For All The Fish!")
19:20:12 Join DerPapst [0] (n=DerPapst@p5B23EBFC.dip.t-dialin.net)
19:20:30n1spreglow: jhMikeS, voice works great in sim here not tested on target, size dropped from 1.2MB to ~1.0
19:20:53preglowgood, good
19:21:22preglowamiconn: but yeah, do you think the ml error report is because of the sapi bug? the error code pasted tells me nothing...
19:21:47linuxstbpreglow: You're talking about the "error 141" ?
19:21:48amiconnVoice works fine on target btw (tested mini G2, german voice)
19:21:53preglowlinuxstb: aye
19:22:20amiconnpreglow: I would think so
19:22:22linuxstbGoogle is telling me that it's due to a broken pipe
19:23:15amiconnpreglow: This is a more verbose paste of the error I got: http://pastebin.ca/780877
19:23:23preglowi'll just ask
19:23:49preglowhmm
19:24:10lostlogicn1s: I ... thought I did...
19:24:13lostlogiclike just last week
19:25:04amiconnlinuxstb: Yes, because the vbscript errors out if it cannot find a command
19:25:24DerPapstare the buildservers already building new voices in speex?
19:26:01n1slostlogic: it's failing the mrobe builds because it doesn't have the correct lib...
19:26:04amiconnThat should happen automatically
19:26:25DerPapstamiconn: ok
19:28:38amiconnThat is, the next daily voices should work
19:29:17 Quit AceNik (Read error: 104 (Connection reset by peer))
19:29:45linuxstbDoes the new voice playback work noticably better than the previous implementation? (I never used it...)
19:29:59lostlogicn1s: I'm a retard. fixing momentarily.
19:30:06 Quit Falen (Read error: 104 (Connection reset by peer))
19:30:13n1sgreat :)
19:31:37pixelmaif Bagder was around, he could kick off a new voice build for today...
19:31:40amiconnenglish voice dropped from 1.2M to 830K
19:33:25amiconnpreglow: .talk clips are working fine too
19:33:37preglowamiconn: yeah, that had bugs, so i've tested that quite thoroughly
19:34:00preglowamiconn: however, each time i hover over a talk file, it says "talk mp3 clip"
19:34:01amiconnI still have to try whether hwcodec still works - can't do that now
19:34:11preglowis it supposed to do that even when speak file types is off?
19:34:12 Quit joshin (Read error: 113 (No route to host))
19:34:21amiconnno
19:34:27preglowit seems to do so here
19:34:28amiconnOr rather, I don't know
19:34:41preglowalso, "yes" is never pronounced, but that doesn't seem to be our fault
19:34:44amiconnI don't have 'view all' enabled
19:35:16amiconnhmmmm
19:35:25amiconn'yes' definitely worked before
19:35:41amiconnNow I get a click
19:35:43preglowthe bug was there before we touched the code, i'm 95% certain of it
19:35:51preglowother people mentioned it
19:35:55preglowhint: yes is the first clip in a file
19:37:52amiconnThere are several quirks now
19:38:57preglowwhich?
19:39:14amiconn(1) 'yes' doesn't work. (2) When hovering a .talk file, it says 'talk clip' first, and then the content of that clip
19:39:57amiconn(1) might have slipped in with a recent voicefont change
19:40:31amiconnHmm, or not...
19:41:16*pixelma tried on today's M5 build (r 15654) and can definitely hear "yes"...
19:41:27preglowis that so...
19:41:39amiconnIt definitely worked before
19:42:47jhMikeSthis "Yes" thing is something strange (getting a silence clip instead). can't be in the voice thread since that just plays whatever asked and the loader code seemed ok. I couldn't track down any core fault.
19:47:53 Join Lear [0] (i=chatzill@rockbox/developer/lear)
19:52:36 Join bertrik [0] (n=Bertrik_@121-021-045-062.dynamic.caiway.nl)
19:56:12 Join Thundercloud [0] (n=thunderc@resnet02.nat.lancs.ac.uk)
19:56:25 Quit amiconn (Nick collision from services.)
19:56:34 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
20:00
20:00:20 Quit Mouser_X (Read error: 110 (Connection timed out))
20:00:38 Join karashata [0] (n=karashat@pool3-048.adsl.user.start.ca)
20:06:06 Join AceNik [0] (n=AceNik@117.98.45.163)
20:06:25AceNiklinuxstb: thanks a lot it helped now i can compile perfect
20:06:56AceNikcan anyone tell me what does the reason for closin this patch mean ? #4988
20:08:28preglowjhMikeS: so, now what features does spc have on ipods?
20:09:15jhMikeSecho.
20:09:26jhMikeSsome asm optimizing
20:09:35preglowstill no gauss interpolation?
20:09:49jhMikeSyou'll never get that out of it
20:10:05*preglow wonders why that is so cpu hungry
20:10:16preglowi would have thought gaussian interpolation was pretty much just a table lookup
20:10:26jhMikeScf has a prayer but that darned ARM multipler
20:10:31SoapAceNik: IIRC the decision was that instead of bloating the core of Rockbox with that function, the plugins themselves should be modified to keep the backlight on if they so desire.
20:10:45jhMikeSmaybe I'll work on it, but it sounds much better now
20:10:54preglowi can imagine
20:11:06preglowand this thing is a nice way to see if dual core threading works nice,yes?
20:11:39AceNiksoap: ok i get that , but which plugin as such has been modified to keep that fuction working since then ? can you name a few of them
20:11:45jhMikeSwell, it has been run for hours on end so it pretty much confirms it
20:11:55SoapAceNik: That I can't say.
20:12:10 Join jpt9 [0] (n=chatzill@teal-blue-85.dynamic2.rpi.edu)
20:12:12jpt9hey.
20:12:25jpt9what does it mean if Rockbox locks up with an error of "Data abort at 40001314 (0)".
20:12:38jpt9I was trying to play an MP3 (selected it from the file browser).
20:13:23AceNiksoap: i dont think after this was declared, a few plugins miht hav kept up to keeping, backlight on but not others whih require, like i know mpegplayer as a fact did get this feature, but none of the other plugins
20:14:31 Quit pixelma (Read error: 110 (Connection timed out))
20:15:17jpt9and sometimes rockbox just refuses to play music; it seems to have trouble reading the MP3 file (it can't get the bitrate, for example).
20:15:25bertrikjpt9: something bad happened, I think it means that the processor tried to access some memory or peripheral but nobody answered
20:15:26jpt9i'm using a Sansa e250R.
20:15:29jpt9oh.
20:15:34jpt9i assume it's a bug?
20:15:42bertrikyes probably
20:15:54jpt9how stable is rockbox, exactly?
20:16:04AceNikjpt: which player model & rockbox revision are you referin to ?
20:16:12lostlogic'exactly', hehe
20:16:18jpt9an e250R with the latest version as of a few minutes ago.
20:16:18bertrikI've seen bugs but never experienced data aborts
20:16:22jpt9oh.
20:16:49*karashata has rarely run into bugs of any sort recently with normal use
20:17:10LearCould be a bad ID3 tag here (bad, as in causing problems for Rockbox, not necessarily bad in itself).
20:17:21AceNikjpt9: what kind of settings you usin in playback, like do you hav crossfade on with equilizer?
20:18:24jpt9crossfade is on...
20:18:40jpt9equalizer is on...
20:18:43jpt9custom settings...
20:18:44Soapcan we quickly verify one thing jpt9? Are you using a current, stock, build?
20:18:52jpt9yes.
20:18:59jpt9downloaded and installed using rockbox utility.
20:19:02AceNikjpt9: turn off the eualizer
20:19:26jpt9huh...
20:19:45jpt9it just started working again...
20:19:57jpt9would having the USB connection plugged in have anything to do with it?
20:20:01bertrikthose kind of bugs are the worst
20:20:08AceNikjpt9: things will be fine fromt here, im sure you using a little more resources than your player a handle, turn off the equalizer & use the bass & treble instead
20:20:39AceNikjpt9: no usb wont make a diff till your battery is low
20:20:49jpt9i discovered that if you plug the USB connector in at the right time (possibly after unplugging it), you can get rockbox to *not* display the USB screen...
20:20:56jpt9I'm not sure if this is supposed to happen...
20:21:11jpt9but i'm definitely making it not do something it normally does.
20:21:29jpt9is there any way to see how much RAM/CPU is left or something?
20:21:52bertriksystem/debug/buffering thread show memory buffers
20:22:04 Join billenium [0] (n=billeniu@c-69-249-243-110.hsd1.pa.comcast.net)
20:22:05jpt9ah.
20:22:28bertrikand shows business of processor too, as a percentage of the time that the cpu is boosted
20:22:47jpt9shiny.
20:23:21jpt9huh...
20:23:28jpt9rotating the wheel switches tracks at that screen.
20:23:58bertrikI think that's a feature, not a bug :)
20:24:03jpt9ah.
20:24:09jpt9oh yeah..
20:24:21jpt9can you add a feature to stop recording automatically when there's x MB of space left?
20:24:24 Quit atsea- (Remote closed the connection)
20:24:32jpt9i temporarily screwed up my Sansa when it ran out of space recording...
20:24:55jpt9(rockbox couldn't save any settings or launch any viewers/players −− basically, it couldn't do anything. including delete files)
20:25:56 Join xaNz [0] (n=xc@bas12-toronto63-1088795417.dsl.bell.ca)
20:26:01bertrikI guess you could sumbit a bug on the tracker, if there isn't one like that already
20:26:02jpt9oh yeah...
20:26:25jpt9i assume there aren't any undocumented features like being able to record from line-in by plugging the output of another device into the headphone jack?
20:27:11jpt9it's already in the tracker −− http://www.rockbox.org/tracker/task/5852
20:27:28 Join Orc [0] (n=Orc@83.149.21.140)
20:27:30bertrikno, AFAIK the headphone jack is hardwired as output
20:27:45jpt9damn... that'd be cool.
20:28:01AceNikguys if i patch files in the bootloader directory, does that mean need to rebuild a bootloader or a build will reflect the changes ?
20:28:07 Quit OlivierBorowski (Remote closed the connection)
20:28:56 Join OlivierBorowski [0] (n=OlivierB@ANancy-157-1-24-9.w81-50.abo.wanadoo.fr)
20:29:15*lostlogic wants his 27k back from the voice codec if it's not enabled at all
20:29:29 Quit HellDragon (Nick collision from services.)
20:29:31preglowi'm open to suggestions as to how that can be managed
20:29:43Orchi. I'll want import rockbox fonts on LinuxiPod? but i dont know as
20:29:45 Join HellDragon [0] (i=jd@unaffiliated/helldragon)
20:30:05AceNikwhat does trailing CRs from patch mean ?
20:30:17SoapCarriageReturns
20:30:23lostlogicpreglow: if the codec can run out of standard RAM, couldn't it be loaded and started in the 'main buffer' when voice is enabled?
20:30:53preglowcan we not do any further voice related hacks?
20:30:59preglow25kb is nothing
20:31:08preglowand i'll shrink it even further
20:31:40Orcanybody may help me?
20:31:45lostlogicpreglow: hehe, fine fine −− how about a build time #ifdef :-P (I _really_ like turning off things everyone else has on in my builds)
20:32:14preglowwell, sure, but i don't see us having a separate build for that
20:32:21 Quit merbanan ("Leaving")
20:32:55lostlogicpreglow: yeah, we don't have a separate no tagcache, no dircache, no priority scheduling build either :-D
20:33:06jpt9does the radio in the Sansa provide RDS? and can Rockbox read it?
20:33:07preglowheh, yeah
20:33:21preglowand yeah, the codec can run out of standard ram, but performance is very much enhanced by the use of a few kb of iram
20:33:25jpt9the spec sheet linked to from the sansa hardware page on the wiki suggests so...
20:33:28jpt9but i'm not sure...
20:33:43lostlogicmakes sense.
20:34:43lostlogicOrc: probably need to ask IPL people −− very few people (if anyone) in here uses it.
20:34:56moospreglow: hi, I guess this page need an update (nice works btw) ;) http://www.rockbox.org/twiki/bin/view/Main/MajorChanges
20:35:20bertrikjpt9: I would be surprised if it did RDS
20:35:21preglowmoos: yeah, that makes sense
20:35:41moosbig enhancement !
20:36:03moosfor those that use voice ui (I don't)
20:36:06jpt9is there any way to manually tune the radio (without presets)?
20:36:11preglowmoos: me neither :P
20:36:29moos: )
20:36:32*jhMikeS doesn't understand lostlogic's priority scheduling hangups
20:36:42Orcwhere I may find this people. on IPL forum nobody tell me as
20:37:02*jhMikeS will set out to make it impossible to remove it on SWCODEC :P
20:37:21Orci understand 2 things. font and codepage
20:38:24 Part AceNik
20:39:38lostlogicOrc: #ipodlinux
20:39:49Orcthx
20:39:53lostlogicjhMikeS: and I'll keep fixing the bugs you introduce by so doing
20:40:16lostlogicjhMikeS: I just don't see why I'd want it... things worked (and continue to work) just fine without it
20:41:29*Nico_P figures it's time to try voice
20:41:49*lostlogic too is more tempted now that it's not all codec-crazy-swapping
20:42:01jhMikeSlostlogic: then what's the rationale for it and how would I introduce bugs if things work fine either way?
20:42:20lostlogicjhMikeS: I just meant if you made it impossible to disable it, that would be a bug.
20:42:39jhMikeShehe...depends
20:42:55*preglow never uses voice and will live fine with it in core
20:43:23jhMikeSthe voice codec _could_ be loaded on the fly. only problem is the core IRAM usage for it.
20:43:36amiconnpreglow: Something's hosed on coldfire
20:44:05amiconnA current build + speex voice freezes within a few seconds when just moving around the main menu
20:44:10amiconn(on M5, that is)
20:44:13*jhMikeS has been testing repeatedly on coldfire
20:44:28*preglow too
20:44:39 Quit PaulPosition (Read error: 110 (Connection timed out))
20:45:03preglowi did manage to make it freeze in a menu once, but was never able to reproduce
20:45:14preglowi just figured it was a general rockbox bug since that does happen from time to time...
20:45:39amiconnFreezes after moving through 4..5 menu items, consistently
20:45:47preglowwith voice enabled?
20:45:51amiconnyes
20:45:55preglowtried without?
20:45:59amiconnSettings cleared
20:46:09amiconnThat means menu voice is enabled
20:46:35jhMikeScan speex as codec do it without difficulty?
20:46:39amiconnI can't disable it
20:47:02amiconnIt freezes before I can reach the setting
20:47:11preglowwell, i'll do another test
20:47:27jhMikeSI guess edit the config and see if it's ok when just playing .spx files
20:47:34preglowperhaps coldfire speex is just bad
20:47:49preglowi've had hangs with speex before, but i thought i fixed that with the memset commits
20:48:09amiconnI'll check H300
20:48:10jhMikeSit
20:48:19amiconnJust have to build voice files
20:48:30 Join atsea- [0] (i=atsea-@gateway/tor/x-64cfc7b52c4dfeb2)
20:48:49 Quit ompaul (Read error: 113 (No route to host))
20:48:55jhMikeS's a hard lockup when it happened to me. I put more thread safeguards in to be sure size > 0 before accepting it for decoding.
20:49:27jhMikeSjust in case the clip queue did something funny
20:50:28 Join The-Compiler [0] (n=The-Comp@150-114.0-85.cust.bluewin.ch)
20:50:41 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
20:51:37jpt9man... the fm receiver on the sansa sucks...
20:51:45jpt9it gets interference from my laptop.
20:51:57amiconnWithout voice, the M5 doesn't freeze
20:52:50amiconnEnabled directory voicing, entered browser -> freeze at 3rd dir
20:52:53 Quit DerPapst (Read error: 110 (Connection timed out))
20:52:58 Quit iamben (Read error: 110 (Connection timed out))
20:54:00preglowjmspeex: will the decode_lost functions ever be used in ordinary file decoding?
20:54:48preglowamiconn: you think we should try to contact this brian wolven and ask to put voicebox in svn?
20:54:50Nico_Pheh, this is actually quite nice
20:55:06preglowamiconn: just tried with voice, no freezes :/
20:56:03jhMikeSpreglow: hand him one of your voice files and see if it changes things *curious*
20:56:19 Quit amiconn (Nick collision from services.)
20:56:26 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
20:56:27preglowhrm
20:56:31*preglow is tired of ifdefs
20:56:50preglowi'm _almost_ tempted to just make a voice codec version of libspeex and sync it manually when things break :/
20:57:10 Quit Orc ("Leaving")
20:57:17amiconnjhMikeS: I doubt that the voice file is the problem. It's working nicely on mini G2
20:57:24 Join eigma [0] (i=eigma@216.48.162.210)
20:58:00jhMikeSamiconn: so something in the codec then?
20:58:53amiconnI would think so
20:59:00preglowwtf, i just had a core dump when generating voice clips with festival
20:59:05amiconnI wonder why it influences M5 but not H1x0
20:59:17amiconnCan't test on M5 anymore now
20:59:18*jhMikeS svn blame preglow :P
20:59:25amiconnWill try H300, and later als X5
20:59:30preglowi'm trying with another voice file now, just in case
20:59:44preglowwith vbr files, speex uses quite a number of different submodes, so you might get a bug on just some files
21:00
21:01:05 Join handmadematters [0] (n=handmade@80.224.161.24.dyn.user.ono.com)
21:01:11amiconnHmm. Could the mode depend on the actual voice?
21:01:35handmademattersHi all
21:01:37amiconnBecause, I put a male voice on the M5, while I use female voices on my targets
21:01:52preglowamiconn: maybe
21:02:25handmadematters if ( GPIOD_INPUT_VAL & 0x20 )
21:02:26handmadematters {
21:02:26handmadematters GPIOD_OUTPUT_VAL &=~ 0x40;
21:02:26DBUGEnqueued KICK handmadematters
21:02:26handmadematters udelay(250);
21:02:27handmadematters data = adc_scan(ADC_SCROLLPAD);
21:02:27***Alert Mode level 1
21:02:27handmadematters GPIOD_OUTPUT_VAL |= 0x40;
21:02:31handmademattersups
21:03:11handmademattersaehmm... how critical is it to run the above code at to positions at the same time?
21:03:17handmadematterstwo
21:04:03handmademattersI dont know exactly what this GPIOD stuff is doing with the hardware
21:04:49CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
21:04:49*linuxstb points out http://pastebin.ca to handmadematters
21:06:36bertrikgpiod controls the voltage on a set of pins on the controller, in this case the "GPIOD" set of pins
21:07:14bertrikit looks like a specific pin is set low, then an adc conversion is done, and the pin is set back to high
21:09:34***Saving seen data "./dancer.seen"
21:10:18 Join hannesd_ [0] (n=light@gate-hannes-tdsl.imos.net)
21:12:28***Alert Mode OFF
21:12:29 Nick fxb is now known as fxb__ (n=felixbru@h1252615.stratoserver.net)
21:12:49jpt9so the sansa's cpu is really dual-code?
21:12:56jpt9*** dual-core?
21:13:08preglowi love how festival pronounces "bass" like the fish
21:13:19amiconnumm
21:13:32amiconnOn H300 I get 'stkov voice' within seconds'
21:13:44amiconn-'
21:13:56preglowwell, we do ride the stack _really_ tightly
21:14:39preglowamiconn: try adjusting the staack in voice_thread.c
21:15:31amiconnweird... after a reboot it works??
21:15:46preglow...
21:15:47amiconnAh, no
21:15:55amiconnIt just didn't crash as early
21:15:59 Nick fxb__ is now known as fxb (n=felixbru@h1252615.stratoserver.net)
21:16:03bertrikI'm not a native speaker, but ploo-gins sounds pretty funny to me
21:16:09preglowploojins :P
21:17:42jhMikeSpreglow: could some sets of modes use more stack?
21:17:55preglowjhMikeS: jmspeex said no
21:18:03preglowa main mode always uses the same amount of stack
21:18:23jhMikeSit sould be dialed to about 93% or so...less than libmad
21:19:18jpt9so what's this about you guys changing voice file formats?
21:19:31preglowwhat you said
21:19:42jpt9will it matter at all for end-users?
21:20:16 Quit eigma ()
21:20:27 Join eigma [0] (i=eigma@216.48.162.210)
21:20:51preglowjpt9: well, yeah, they can't use their old files anymoer
21:20:59preglowand the new voice files will be smaller and sound better
21:22:02jpt9ah.
21:22:26jpt9in case it matters, I just updated Rockbox and the voice files, and voice doesn't work anymore.
21:23:03linuxstbThe voice files on the website aren't updated yet - you'll need to wait until tomorrow.
21:23:12jpt9ah.
21:23:13jpt9ok.
21:23:40*preglow reminds himself he still hasn't updated rbutil
21:24:13linuxstbpreglow: Have you added rbspeexenc functionality to rbutil?
21:24:22 Quit hannesd (Read error: 110 (Connection timed out))
21:24:23 Nick hannesd_ is now known as hannesd (n=light@gate-hannes-tdsl.imos.net)
21:25:13amiconnI increased the stack size to 0x800 (from 0x740) and it seems to fix the problem on H300
21:25:24amiconnBtw, stack usage is 92% with that increased stack
21:25:54 Quit The-Compiler ("Verlassend")
21:26:06J3TC-Hrmm
21:26:17J3TC-Is the scroll margin patch have been completely scrubbed off?
21:26:39J3TC-Like I know you guys implemented a part of it and I'm trying to see which one wasn't ;x
21:27:04PaulJamJ3TC-: some parts of it are now in the official build (only leftmarins)
21:27:16linuxstbIt's still on flyspray, but the task is closed, so you need to include closed tasks in your search.
21:27:38preglowlinuxstb: not at all
21:27:39PaulJamthe rightmargin part wasn't.
21:27:45J3TC-Yeah I'm searching by number.
21:28:01preglowlinuxstb: i started thinking about it yesterday, but then had to do other stuff, apparently
21:28:02J3TC-It's just the names have been changed and I do see the leftmargin tags here and there
21:28:58preglowlinuxstb: i was thinking that perhaps just calling the (renamed) rbspeexenc.c main() routine straight from rbutil would make for nice code-reuse, but i don't know how the libc file ops work together with qt
21:29:30amiconnThat means the stack overflows with just 0x740
21:30:01amiconn92% of 0x800 is 0x75c
21:30:05linuxstbpreglow: It guess it could be problematic - rbutil is unicode as well...
21:30:19preglowlinuxstb: exactly
21:30:39linuxstbpreglow: But Domonoky and bluebrother are the people to talk to I think.
21:31:11preglowlinuxstb: i know, need to find them
21:32:10amiconnI think it's just a problem that the stack is a little too tight
21:32:16*J3TC- tries to compile without the scrollmargins
21:32:28J3TC-I wonder how much that would affect the wps :3
21:32:50*amiconn thinks we should use at least 0x780, or even 0x800 to be safe
21:32:50J3TC-Too much stuff to weed through
21:33:30preglowamiconn: well, sure, if it actually crashes on some targets, then hell yeah
21:33:32 Join Guile`` [0] (n=Guile@78.113.57.106)
21:33:58amiconnWell, it crashes on H340. Not immediately, but after moving around for a while
21:34:11preglowthat's really weird, does the voice thread stack grow or what?
21:34:24amiconnMaybe the fact that I still have mp3 .talk clips on there makes things worse
21:34:36 Quit billenium ("Ex-Chat")
21:34:41preglowamiconn: ahh, yeah
21:34:55preglowamiconn: we really just try to play those right now, and that _might_ make speex do something stupid
21:35:09 Join handmadematters_ [0] (n=handmade@62.174.115.183.dyn.user.ono.com)
21:35:31preglowbut then again, speex isn't supposed to allocate stack based on input values, just mode constant values
21:35:40*preglow just shaved 2kb off rockbox.ipod
21:35:46 Quit mf0102 ("Verlassend")
21:36:21preglowjhMikeS: seems "yes" was broken by us,then
21:36:31jhMikeSwhat happened?
21:36:43preglowwell, i don't know, but several people confirm that the bug wasn't there before
21:37:25jhMikeSI know it wasn't there before...but what is going on I don't know. I wasn't able to spot a reason in the core. perhaps something in the file creation.
21:37:40eigmajhMikeS: has anyone looked into making debugf use the DCC (debug communications channel) on ARM architectures?
21:39:02jhMikeSeigma: btw, yesterday I wasn't talking about throwing a bunch of threads on the DSP, but simply using the regular kernel functions to use it which wouldn't be too hard from the sound of it.
21:39:24 Quit davina_ (Read error: 104 (Connection reset by peer))
21:39:26 Join davina [0] (n=davina@cpc1-sout6-0-0-cust616.sotn.cable.ntl.com)
21:40:34jhMikeSamiconn: crank up the stack a little and see if it hangs again?
21:41:28amiconnCheck back 15 minutes...
21:42:28 Quit nicktastic (Read error: 110 (Connection timed out))
21:43:18preglowjhMikeS: perhaps just add a check in the sim to see if speex_decode_int fails?
21:43:22*preglow back in fifteen
21:46:44 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
21:47:12 Join AceNik [0] (n=AceNik@117.98.45.163)
21:47:27 Quit n17ikh ()
21:48:00 Quit handmadematters (Read error: 110 (Connection timed out))
21:48:15AceNikguys please try & include this patch in the svn its really neat FS #7538 - Custom user splash screen.
21:48:58eigmajhMikeS: yeah, I dunno.. I guess we can worry about thread.c integration after things are actually working
21:49:11eigmajhMikeS: do you have an m:robe or plan on getting one?
21:49:42AceNikguys id like to know one thing hw come the ipod 5G hav this awesome colour contrast to them as compared to the h10[20GB], is it because the h10 has a bad lcd , & why are we only stuck at 24-bit bmp's for the h10, why not more?
21:50:23jhMikeSeigma: Think I'd like to. It's an interesting way to handle things.
21:51:01nanokAceNik: 24bit is more than you would ever be able to tell ;)
21:52:12eigmajhMikeS: sorry to repeat myself, but have you heard anything about the DCC? I have a patch that gives me debugf output over JTAG, was thinking about committing it
21:52:27AceNiknanok: are you sure i see a such a drastic difference in the ipod 5G wps's & the h10's
21:53:49jhMikeSeigma: what's DCC now? dualcore codec? :P
21:53:53nanokAceNik: as far as i know (btw, note that his is completely offtopic), iriver's are not about looks, but about sound and audiofile menaningfull features, and i think it's safe to say they kick ipod's ass out of the box. now, people who have used both can confirm or just the same say that i don't know what i am talking about
21:54:08eigmajhMikeS: lol.. Debug Communications Channel.. sideband channel for misc application data over JTAG
21:54:27 Quit animeloe ("Leaving")
21:54:31jhMikeShaven't head anything, no.
21:54:39eigmadoes anyone else have JTAG on an ARM target?
21:54:53nanokAceNik: that is probably not because of 24bit. your computer display is probably 24bit, and it is probably more than yout eye can discerne
21:55:02 Join Arathis [0] (n=doerk@p508A50AA.dip.t-dialin.net)
21:55:11amiconneigma: barrywardell has
21:55:30AceNiknanok: you seriously dont knwo what your talkin of, which player do you posses
21:55:36nanokAceNik: ofcourse, this is arguable, better displays seem to, indeed, make a difference, when they are much better (like extremely high quality photo prints)
21:56:06nanokbut that is a completely different matter i think. arguably, it is completely irrelevant for a DAP
21:56:26AceNiknanok : so its pretty simple to assume ipods have a better LCD, than the iriver
21:56:41AceNikbut still i would liek to know why 24-bit bmp's for the h10
21:56:44eigmaalso, I'm using debugf from multiple threads and I get data corruption because debugbuf is static.. any problems with protecting it with a spinlock?
21:56:52nanokAceNik: as i mentioned, i might be completelly off, and people who owned at least one of them might put me right. i, ofcourse, own neither of the two :)
21:57:08 Join n17ikh|Lappy [0] (n=n17ikh@c-76-23-98-11.hsd1.sc.comcast.net)
21:57:37nanokAceNik: anyway, i would suggest to move this to -comunity, we are disturbing the developers, and it is not very nice, nor is it efficient ;)
21:57:54nanokcommunity, sorry
22:00
22:00:07 Part AceNik
22:03:15jhMikeSpreglow: interesting results. even with a new voice file, I get a segfault if it tries saying "yes" in nb_decode
22:03:52jhMikeShmmm...seems to be a DEBUGF line
22:11:21 Quit me (Remote closed the connection)
22:13:02jhMikeSpreglow: It tries to call speex_notify in nb_decode but the DEBUGF actually bites the dust. odd.
22:13:16preglowjhMikeS: debugf lines aren't included in ordinary builds
22:13:28 Quit amiconn (" bbl")
22:14:11jhMikeSit's a sim build
22:14:40preglowjhMikeS: well, yeah, but the debugf can't be the error in target builds
22:14:47preglowprobably the same cause, though
22:15:16jhMikeSno...but what's up? it should notify that the frames are corrupt via DEBUGF. it tries, but that crashes.
22:15:32preglowso the frame is corrupt
22:15:37preglowcan you see which return path it takes?
22:15:44 Join ForgottenMemorie [0] (n=XwLiLxAz@cpe-075-177-170-087.nc.res.rr.com)
22:15:48preglowi guess this is my field, but it'll have to wait thirty minutes or so
22:15:54ForgottenMemoriesup
22:16:12ForgottenMemorieI uhhhhh need some help right now... >.>
22:16:17preglowthen you must ask it
22:17:03ForgottenMemorieI was messing around with the Rockbox application on my ipod then I plug it back into my computer and it won't detect so then I unplug it and now it won't turn on...
22:17:28pregloweh
22:17:31preglowwhat does the display say?
22:18:08ForgottenMemorieummm wat display...?
22:18:28 Quit jpt9 (Read error: 110 (Connection timed out))
22:18:29preglowthe ipod display
22:19:01ForgottenMemorieit's.. blank...?
22:19:09preglowpress menu and select simultaneously, and keep them pressed
22:20:51jhMikeSpreglow: not really. rockbox.h line 102 which could be any instance. how do I get a stack trace?
22:21:56 Join DerPapst [0] (n=DerPapst@p5B23C66A.dip.t-dialin.net)
22:22:08preglow"bt"
22:22:28jhMikeSfrom within gdb or bt as a command?
22:22:34preglowwithin gdb
22:22:37preglowit's a command, meaning backtrace
22:22:52jhMikeSafter crash?
22:22:58preglowaye
22:24:26ForgottenMemoriestill... blank and dead
22:24:30preglowtry again
22:24:39preglowyou need to press them at the exact same time then keep them pressed for five seconds
22:24:52ForgottenMemoriestill dead...
22:25:01ForgottenMemorieand when I plug it into my computer it wont' turn on either
22:25:01preglowhas it been charged properly?
22:25:15ForgottenMemorieyep, i tried plugging it into the charger and it won't turn on
22:25:31preglowdid it have plenty battery left when it died?
22:25:39ForgottenMemorieyeah.. it was green..
22:25:48ForgottenMemorieI guess my ipod is broken.....?
22:26:06preglowi doubt it
22:26:15preglowno on has ever managed to break their ipod with rockbox
22:26:20Soapmighty strange coincidence for it to have died at this exact moment.
22:26:46preglowusually either the battery is _really_ empty, or it just seems dead and can be reset with the menu + select combo
22:26:54ForgottenMemoriewell... if it helsp I unplugged my ipod when it was ont he apple screen
22:26:57 Join Zagor [0] (n=bjst@46.35.227.87.static.tab.siw.siwnet.net)
22:27:02ForgottenMemorieand then it died
22:27:15Soaptoggle the hold switch to the on (red) and back to the off (white/silver) position, then press menu+select for 45 seconds w/o moving your fingers once. Every time you twitch your fingers you need to start over.
22:27:46 Join PaulPosition [0] (n=noneofye@modemcable228.133-82-70.mc.videotron.ca)
22:28:57 Join jpt9 [0] (n=chatzill@eevee-15.dynamic.rpi.edu)
22:29:17 Quit jpt9 (Client Quit)
22:29:18 Join linuxstb_ [0] (n=chatzill@rockbox/developer/linuxstb)
22:29:20ForgottenMemorieyep
22:29:22ForgottenMemoriestill dead
22:29:38Soapthen put in on charge for 12 hours and try agin.
22:29:52Soapand I do mean 12 hours. W/O peeking.
22:29:57ForgottenMemoriewell... i'm nto sure if it's charging since it's still blank when I plug it in
22:30:05ForgottenMemorieif I plug it into anything it stays blank
22:30:24SoapI have read everything you have said - and I promise you this is the best course of action.
22:30:32ForgottenMemorieo... ok
22:30:35ForgottenMemoriethanks
22:30:38ForgottenMemoriei'll try it
22:30:47jhMikeSpreglow: jhmikes.cleansoap.org/bt.bmp">http://jhmikes.cleansoap.org/bt.bmp
22:32:29preglowwhat in the hell
22:32:58preglowthat's a weird place for it to crash
22:33:30preglowbah, no
22:33:34preglowi forgot i've got local mods
22:34:02 Part handmadematters_
22:34:24 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
22:35:33preglowjhMikeS: i'll try checking it out myslef
22:35:36preglowmyself too, yes
22:37:16 Quit Thundercloud (Remote closed the connection)
22:37:16jhMikeSstupid thing is that the bt still doesn't tell which instance :\
22:37:29preglowyeah, i know
22:37:39preglowsomething to do with inlining, probably
22:37:59*jhMikeS will pull the inline and try
22:39:38 Join JdGordon [0] (n=jonno@c210-49-113-143.smelb1.vic.optusnet.com.au)
22:40:18jhMikeSnb_celp.c line 1180
22:40:49preglowhmm
22:40:58preglowwell, what it said
22:41:03preglowstream is corrupted
22:41:41preglowi dunno what to do about that apart from finding out why
22:42:12jhMikeSit's trying to decode mp3. that should be perfectly safe though.
22:42:36preglowthe first clip is mp3?
22:42:54preglowwhy would "yes" be mp3?
22:43:09jhMikeSno, I'm trowing an old voice at it
22:43:20jhMikeSthis did work before
22:43:47jhMikeSperhaps the use of lookahead changed that
22:44:07 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
22:44:16preglowwhy would that change it? that only changes your part in things, not the decoder's
22:44:23preglowthe decoder does nothing but tell you what the lookahead value is
22:45:51jhMikeSit might fail in a different way over being started on an mp3 sync word
22:46:55 Quit ompaul (Client Quit)
22:47:58 Join Thundercloud [0] (n=thunderc@resnet04.nat.lancs.ac.uk)
22:48:53preglowwell, sure, but what bug are you investigating now, really? i thought you were looking at the "yes" issue
22:49:43 Join nature [0] (n=nature@pool-151-203-92-183.bos.east.verizon.net)
22:49:54*amiconn has a suspicion...
22:50:58amiconnLooks like the clip positions are off by 4 bytes, because the header grew, and someone forgot to adjust one offset
22:51:14amiconnAnd obviously libmad didn't care, but speex does
22:51:20preglowwhat, and all the other ones still work???
22:51:40 Join safetydan [0] (n=safetyda@rockbox/developer/safetydan)
22:51:44amiconnThey're all off by 4, but they're stored back-to-back
22:52:11jhMikeSmp3 will find sync words automatically
22:52:12amiconnOr no
22:52:21amiconnThey're not offset
22:52:42amiconnBut the last table entry overwrites the first 4 bytes of the 'yes' clip
22:52:54preglow\o/
22:53:25*amiconn goes to verify this theory
22:53:34*jhMikeS wonders why the sim bites it trying to print a debug message thought...that's problem #2
22:54:09amiconnCompare voicefont.c line 141 and line 189 ...
22:54:29jhMikeSthat's why I use template structures :)
22:54:41 Quit OlivierBorowski ("Konversation terminated!")
22:54:47amiconnEven a simple #define would avoid that
22:55:07pixelmathe increased stack seems to help on M5 too (got a freeze after enabling voice at first but since a restart it doesn't crash anymore)
22:55:11 Join Llorean [0] (n=llorean@cpe-70-113-103-34.austin.res.rr.com)
22:55:41jhMikeSpixelma: what's the usage show?
22:56:05 Quit bertrik ("need sleep")
22:56:18 Quit Lear ("ChatZilla 0.9.78.1 [Firefox 2.0.0.9/2007102514]")
22:56:22pixelmalooking
22:56:39pixelma96%
22:56:49pixelmafor voice
22:57:03amiconnSo it's slightly different per target...
22:57:08amiconnH300 gets 95%
22:57:42jhMikeSthat's after an increase?
22:57:45*preglow wonders why
22:57:57amiconnjhMikeS: See frontpage...
22:58:00pixelmathat's with the latest svn build
22:58:31eigmawheeeee, the DSP sends text messages to my screen now
22:58:36jhMikeSinteresting
22:59:05amiconn'yes' clip works with the fix...
22:59:29 Nick parafin is now known as parafin|away (i=parafin@paraf.in)
23:00
23:00:00amiconnObviously speex can't resync to a stream
23:01:13 Join porno [0] (n=e_ibarra@189.156.61.145)
23:01:16 Join billenium [0] (n=billeniu@c-69-249-243-110.hsd1.pa.comcast.net)
23:01:22preglowamiconn: voice codecs very seldomly waste bits on sync words
23:01:42preglowbecause they have so few bits to start with
23:02:24jhMikeSobviously the sim problem isn't stack problem
23:02:38pornohey does anybody knows how to increase the size of the upper bar in an ipod 5g 30gb?
23:03:02jhMikeSit should handle any garbage gracefully...if it doesn't that needs fixing
23:03:09naturei was just given a 3rd gen ipod coming from a windows box with an old (1.5 years ago) version of Rockbox. I'm on os x 10.4.something. I put in the latest .rockbox folder (the mac could only mount the ipod when it was booted into the original firmware, but it let me copy to and fro with no trouble.). my current troubles are: the mac doesn't recognize the ipod booted into rockbox and , more importantly, now (with full and health
23:04:00Lloreanporno: You can't at the moment, without programming
23:04:06Lloreannature: cut off at "full and health"
23:04:31naturethanks
23:04:31Bagderpreglow: so, do I need to change anything for the automated voice builds ?
23:04:34*jhMikeS really thinks voicefont.c and talk.c should share a template header
23:04:35nature(with full and healthy battery and even plugged in ) everytime it boots into rockbox, it shuts down immediately with "battery empty RECHARGE shutting down" (and then restarts and shuts down and restarts). Any ideas. Should I use one of rbutil's installation schemes? which one?
23:04:51preglowBagder: i don't think so
23:05:02Bagdergreat
23:05:10preglowBagder: if there's no question which codec to use: no
23:05:15preglowand i don't believe there is
23:05:21JdGordonamiconn: have you got a patch with the backlight using arbitrary values instead of the indexes?
23:05:29Lloreannature: First off, Rockbox doesn't have a USB mode, so it's expected that Rockbox isn't recognized and you have to be in the original firmware.
23:06:03Lloreannature: This is being worked on, but no estimates when it'll change. As to your problem, I don't know. But if the build is 1.5 years old (or was, rather), I'd suggest installing the newest bootloader.
23:06:17natureFirst, thats fine (does it make a difference that im plugged in by firewire?)
23:06:28pornothnx
23:06:29jhMikeStalk.c uses struct voicefile to access the header
23:06:53naturesounds good, ill get back to you.
23:06:59Lloreannature: Well, firewire support may be further off than USB. ;)
23:07:24amiconnBut firewire reboot should be 100% reliable
23:07:47Lloreanamiconn: His previous build was _very_ old, and his current build doesn't apparently boot far enough to get to that, though
23:09:25 Quit Robin0800 (Read error: 110 (Connection timed out))
23:09:37***Saving seen data "./dancer.seen"
23:10:11jhMikeSpreglow: any other nifty gdb commands I should try out?
23:10:28preglownone i can think of, i really suck in gdb
23:10:35preglowi mostly just use bt and print
23:10:56Bagder'n'ext and 's'tep and 'c'ontinue
23:11:16*jhMikeS wonders if the DEBUGF is somehow the wrong one
23:11:45jmspeexamiconn: If you want re-sync ability with Speex, that's what Ogg is for
23:11:50preglowBagder: you think the new voice stuff merit a news item? might ward off some people who don't pay attention
23:12:01Bagderyes it certainly does!
23:12:20preglowjmspeex: oy, will the decode_lost stuff ever be used at all in file based decoding?
23:12:36jmspeexpreglow: no
23:12:43naturesorry guys, now I'm getting the bad folder icon and im in disk mode. the installation process said that it wrote a log, where can i dg that up?
23:14:29jhMikeSpreglow: if I take that particular DEBUGF out, I just get frame erros and silence feeding it mp3 (which is what should happen). weird.
23:14:51amiconnIn fact the 'yes' clip was broken before... but one missing mp3 frame wasn't noticeable
23:15:11preglowjhMikeS: very
23:15:21*amiconn needs to perform a full update cycle on all his targets... will take a while
23:17:09 Quit Frazz ("Leaving")
23:17:53 Join Mouser_X [0] (n=mouser_x@LAYL001.digis.net)
23:19:01jhMikeShehe
23:19:36Bagderfront page updated
23:20:37preglowthanks
23:21:00jhMikeS#if defined(DEBUG) || defined(SIMULATOR)
23:21:00jhMikeS#undef DEBUGF
23:21:00jhMikeS#define DEBUGF ci->debugf <== !!
23:21:53JdGordonamiconn: you see my q before?
23:22:11amiconnJdGordon: yes, no
23:22:30amiconnThat means I don't have such a version yet
23:23:05 Join MethoS- [0] (n=clemens@pD955DD06.dip.t-dialin.net)
23:23:57JdGordonk
23:24:09 Quit Seed (Nick collision from services.)
23:24:10amiconnI first wanted to be sure that the settings system can be adapted without problems
23:24:11JdGordonill add a fake setting for testing then
23:24:16 Join Seed [0] (i=ben@bzq-84-108-237-178.cablep.bezeqint.net)
23:24:43amiconnI will prepare that after updating my targets...
23:25:30JdGordonshould I add a min/max values? or just use the first and last array value for them?
23:25:50jhMikeSpreglow: yup. wrong DEBUGF
23:26:03amiconnJust first & last array value should be ok, I think
23:26:31amiconnI can't imagine a setting where the settings system should not offer the full range...
23:27:41 Join Robin0800 [0] (n=Robin080@cpc1-brig8-0-0-cust97.brig.cable.ntl.com)
23:29:14*jhMikeS wonders why a ci pointer is declared in the core...or extern struct *ci declared elsewhere emulates the struct ci
23:29:27JdGordonamiconn: and should I bother adding an option to force the value to one in the list? or always add the users value in?
23:30:27amiconnI like the dynamic addition idea. This way the user can even reset his not-normally-in-the-list value as long as he stays in the list
23:31:00natureSo after updating the firmware from the very oldversion, I can't boot into either rockbox or the original code, i end up in disk mode. I've tried installed a few different times. now my options are to try and troubleshoot this or just restore everything via apple's scheme. given that im just maybe Halfway competent, which course would you recommend for me. and if i'm going to troubleshoot, where are the logs i should be diggi
23:31:19 Quit jhMikeS (Nick collision from services.)
23:31:25 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
23:32:03natureand if i'm going to troubleshoot, where are the logs i should be digging through?
23:32:06*amiconn likes the new voice file sizes :)
23:33:32jhMikeSpreglow: why are DEBUGF and LOGF redefined?
23:33:49 Quit Thundercloud (Remote closed the connection)
23:34:37pixelmaspeaking of file sizes... Bagder do you have an idea what's causing the one round delayed numbers in the delta table for random targets?
23:35:00Bagderuh, no... I wasn't aware of any such
23:35:10 Quit MethoS- ("Konversation terminated!")
23:36:02Bagderany example from the current table?
23:36:25pixelmasee for example the M5 and the m:robe after today's big voice commit
23:38:05Bagderthe table shows the size from the "log" at least, so the log should be where the error is...
23:38:09 Quit mdevz (Read error: 104 (Connection reset by peer))
23:38:25 Join mdevz [0] (i=mdevz@cpe-065-184-173-244.ec.res.rr.com)
23:39:55 Quit porno (Read error: 110 (Connection timed out))
23:42:39Bagderand the log shows the m5 with a different rev number...!
23:42:48amiconnpreglow: Hmm, the bug in voicefont.c dates back to 5 Aug 2007 (langv2 cleanup)...
23:45:05Bagderand here it is... "Warning: couldn't get zip from rbclient@wardell.ucd.ie/build-iaudiom5"
23:46:04jhMikeShrm...an e200 DEBUG build won't build at all...errors in gui/wps_debug.c :\
23:46:50amiconnDEBUG target builds only work for archos (and even then not for all - the Ondios are excluded), and iFP, afaik
23:47:09JdGordonamiconn: last tihng, for the config values, should I add the option to have some strings mapped to the first values in the list? so it would say "always off" instead of -1 and on for 0 and then just numbers for other values?
23:47:18preglowamiconn: that's a relatively young bug in rockbox terms :)
23:48:23amiconnJdGordon: Hmm. Would it be simple enough to allow optional strings for all values in the list
23:48:26amiconn?
23:48:48Bagderok, the problems seem to be connectioned to barry's host
23:48:53amiconnThe other option would be a formatting function for the setting, as you suggested
23:48:54Bagderconnected even
23:48:57JdGordonthe way i was going to do it was just map the comma seperated string list to the first list values
23:49:18ForgottenMemorieMy Ipod is Dead, blank screen, If I plug it anywhere it's still blank, tried reseting, nothing happens, still blank. Any other Ideas?
23:49:23JdGordonbacklight: on looks better than backlight: -1
23:49:43amiconnyeah, sure
23:50:21 Quit Jon-Kha (Remote closed the connection)
23:50:56amiconnA formatting function with a switch() could do that too - but if the settings system handles it, it might be more efficient if more than one setting uses it
23:50:59 Quit Toki (Nick collision from services.)
23:51:10amiconnAnd there will be more than one...
23:51:13 Join Toki_ [0] (n=hsdbvlkb@gateimb.imb.lebedev.ru)
23:51:39JdGordonwell it will have a text formatter for the screen, but thats not the same as the text for the .cfg
23:51:41*preglow tries out his shrunk speex
23:52:23 Quit ForgottenMemorie ()
23:52:35amiconnAh, you mean in the .cfg?
23:52:55JdGordonyeah
23:53:19 Part linuxstb_
23:53:30amiconnHmm. I wouldn't care too much if there are just the numbers, but if it's simple enough, why not...
23:53:33*jhMikeS thinks printf should support 64-bit ints on LOGF builds
23:53:53amiconneh?
23:54:20Bagder64 bit *ints*?
23:54:22jhMikeStrying to LOGF spx_int64_t barfs a bunch of warnings
23:54:53Bagderah int64_t...
23:55:23PaulJamchanging the values that get written in the config file is maybe not such a good idea. i mean wouldn't it mean that some settings silently change when i update from an older build which still used the old values?
23:55:28amiconnJust log "0x%08x%08x", (unsigned)(val>>32), (unsigned)(val)
23:55:30jmspeexpreglow: FYI, I managed to shrink testenc down to 60 kB on PC!
23:55:43jmspeexjhMikeS: Speex no longer uses 64-bit ints
23:55:58preglowjmspeex: hah, not bad
23:56:08preglowjmspeex: that is still including the decoder, i take it?
23:56:14jmspeexthat's amd64 BTW
23:56:22jmspeexthis is nb encode+decode
23:56:24preglowwhich means bigger binary c0ompared to x86
23:56:29 Join Gnu47_ [0] (n=Gnu47@private.ntwk.thita.net)
23:56:58*preglow just tried decoding nb clips on target, rate was off by two
23:57:10 Join radinp [0] (n=pradin@psyduck-08.dynamic.rpi.edu)
23:57:20 Quit Gnu47 (Nick collision from services.)
23:57:24 Nick Gnu47_ is now known as Gnu47 (n=Gnu47@private.ntwk.thita.net)

Previous day | Next day