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 2005-12-05

00:01:14 Nick ashridah is now known as Lost-ash (n=ashridah@
00:03:06FebsIs it worth my while to write documentation for len0x's recent patch?
00:03:21FebsOr do you think that you (or others) will change or revert it?
00:03:36amiconnI don't know. If it was just me, I would revert it immediately
00:03:52Bagderwhat patch?
00:03:57Bagderthe clipping thing?
00:04:43BagderI still wait for him to rever the wrapping integers
00:05:09amiconnWrapping integers were introduced by TiMiD
00:05:16Bagderah, my mistake
00:05:31BagderI have no real opinion on the clipping thing, it has never bothered me
00:06:06amiconnIt's the option clutter that really annoys me, plus so much happening behind the scenes
00:06:20_FireFly_wrapping integers ?? what's that
00:06:21 Quit lostlogicx (Read error: 110 (Connection timed out))
00:06:23BagderI'm against option clutter as well
00:06:27*DrumRBoy320 says 'sorry for making it an issue'
00:06:35 Quit lostlogic (Read error: 110 (Connection timed out))
00:07:11amiconnI'm working on the volume % -> dB transition, with removed limting by the prescaler
00:07:12Bagder_FireFly_: when you move beyond max integer it wraps to 0
00:07:18Bagderand vice versa
00:07:33Bagderfor bass/treble/volume etc
00:08:04FebsYikes, that's a bad idea. I can see why you want it reverted.
00:08:15FebsIf I'm reducing the volume to zero, I don't want it to jump back to 100!
00:08:16amiconnThe prescaler thing was really trivial, just moving two lines of code...
00:08:27BagderFebs: yeps
00:08:29mirakFebs: I agree that's not good
00:08:37amiconnme too
00:08:53BBubamiconn: cant you just revert the other approach and add yours? :)
00:08:54_FireFly_Bagder: ah thanks
00:08:57amiconn(but by far less annoying than the clipping option thing)
00:09:14linuxstbBagder: I've found some problems with the new tools building system. If you create a build directory for the Archos Player or iPod (for example), type "make", then "make clean", then "make", then not all the required tools are compiled.
00:09:26linuxstbFor the Player, it's convbdf, for the iPod, it's scramble.
00:09:29amiconnBBub: Sure I could
00:09:41amiconnThe question is whether I should
00:09:47Bagderlinuxstb: player doesn't neec convbdf
00:09:58linuxstbBagder: It complains that it's not there.
00:10:10Bagderfor the sim?
00:10:33Bagderit just struck me it doesn't make any tools for the sim
00:10:34linuxstbI'm just checking again now...
00:10:45linuxstbNo, these are "normal" builds.
00:11:06BagderI'll check it out
00:11:09amiconnBagder: While you are at the Makefiles, could you perhaps have a look at integrating the self-extractor?
00:11:20linuxstbYes, player builds complain when compiling sysfont.o
00:11:29*Febs decides not to work on documentation for the prevent clipping option.
00:11:48Bagderlinuxstb: ok, the player shouldn't need that
00:11:54FebsThere are some other things that need attention anyway, like documentation for crossfeed and crossfade.
00:11:56mirakit would be nice to have two bootloaders, one that default to rockbox, the other to the original firmware
00:12:04mirakjust to have the choice
00:12:10Bagderso make your own
00:12:11*amiconn wants crossfood ;)
00:12:25*Febs laughs out loud.
00:12:25linuxstbBagder: I thought Players had some custom characters - do they not use convbdf?
00:12:36mirakBagder: I am scared you know
00:12:54BagderI know
00:13:25Bagderwe don't need a bootloader to load the original firmware, imho
00:13:52amiconnmirak: Why would you want that?
00:13:53mirakfor H300 that's not enough mature
00:14:01Bagderso don't install it
00:14:05Bagderor help us fix it
00:14:08mirakBagder: comeon
00:14:12*amiconn was about to say the same
00:14:13mirakBagder: sure
00:14:38mirakdon't take that as an offence
00:14:48BagderI just find it a sorry attitude
00:14:56Bagderand you're hardly the first to suggest it
00:15:15mirakso what's the problem ?
00:15:22Bagderthere's no problem
00:15:40FebsMirak, what's the problem with simply pressing REC + ON?
00:15:45mirakok so let's say I asked nothing
00:16:14mirakFebs: there is no debate
00:16:38mirakit was just a suggestion ...
00:17:16MoosI never heard here any player die cause Rockbox ;)
00:18:09Mooslot of people are scaring for "nothing"
00:18:13mirakFebs: just that for now I see more rockbox as the second option. I am sure this will change
00:19:07Bagderit is specificly said to be for developers
00:19:29mirakI plan to develop
00:19:30XavierGrI think that the first bootloader fo the H100 was the other way around
00:19:32Bagderwhy would Rockbox h300 developers boot anything else?
00:19:35mirakthat's not the issue
00:19:42XavierGryou would have to press REC+ON to start Rockbox.
00:20:09Bagderstill, you're asking for just another load to put on someone else's shoulders
00:20:26Bagderthus, my opinion
00:20:37Moosneed to sleep here, good night all
00:20:46 Quit Moos ("Glory to Rockbox")
00:21:02mirakBagder: I would change it myself. Linus advise to not build our own bootloader
00:21:09mirakI think I will follow his advice
00:21:47BagderI know he says that
00:22:02Bagderbut then he doesn't provide any "reversed" options either
00:23:01Bagderamiconn: care to repeat to me how I builds a compressed image?
00:23:20BBubmake zip
00:23:25BBubif you meant that
00:23:27amiconnFirst, the Makefile needs to be fixed to use the correct dirs
00:23:48muesli_-who is hosting
00:24:02Bagdermuesli_-: we do
00:24:03amiconnThen, building the .ajz will need rockbox.ucl as a prerequisite
00:24:48amiconnFinally, the rule for building the .ajz/.mod from the bin needs to be replaced by a script, which should do the following:
00:25:15amiconn- call scramble with rockbox.bin, and check the return code
00:25:28Bagderlinuxstb: fixed
00:25:33amiconn- if ok, then the script has done its work
00:26:21amiconn- if the return code says the firmware is too big:
00:26:26amiconn- call 'make' in firmware/decompressor
00:26:52amiconn- call scramble again, this time using 'compressed.bin' built in the last step
00:28:09amiconnThe Makefile in firmware/decompressor should do the majority of its work in <build_dir>/firmware/decompressor, but use <build_dir/rockbox.ucl as input
00:28:23amiconnCurrently it does all its work in the source dir
00:28:50amiconnI couldn't figure out how to use the magic :(
00:30:11Bagderok, I'll start working on it
00:30:17linuxstbBagder: You forgot to add scramble to the ipodbitmaptools. But I could do that.
00:30:20amiconnIt would be really good if someone with an unflashed (unflashable?) archos could test a compressed ajbrec.ajz
00:30:27Bagderlinuxstb: sorry, please do
00:30:51amiconnI'm convinced that it will work, but it could be interesting to compare boot time with uncompressed vs. compressed
00:31:16 Quit _FireFly_ (Read error: 110 (Connection timed out))
00:31:54 Join _FireFly_ [0] (
00:32:22_FireFly_Bagder: for the wrapping integers to stop is a simple change only needed
00:32:40BagderI'm sure
00:32:53_FireFly_gui/select.c:80 change from gui_select_limit_loop(select, false); to gui_select_limit_loop(select, true);
00:34:15amiconn_FireFly_: Does this prevent the wrap for numeric settings only?
00:34:33amiconnBecause, it makes sense to allow wrap for list-like settings
00:35:01_FireFly_not tested
00:35:28_FireFly_hmm it seams that this change it for all
00:36:01XavierGrcan someone please explain to me what this declaration means? int currentSample IDATA_ATTR;
00:36:35XavierGrit decalres an integer with the name currentSample but what's with the IDATA_ATTR extension?
00:36:38 Quit JonSenior ("using sirc version 2.211+KSIRC/1.3.10")
00:36:53BagderXavierGr: it puts the variable in the internal memory
00:37:22XavierGrfor faster access?
00:37:32amiconnYes. Apart from that, the declaration isn't 100% correct (anymore)
00:37:50amiconnIt should be int currentSample IBSS_ATTR;
00:38:02amiconn..because the variable isn't initialised
00:38:09_FireFly_amiconn: to have it only for integers two changes are needed
00:38:19XavierGrit will give better performance or...?
00:39:23_FireFly_remove the line gui/select.c:80 and change the line option_select.c:42 from opt->limit_loop=false; to opt->limit_loop=true;
00:44:49 Quit ender` (Read error: 104 (Connection reset by peer))
00:45:31JdGordonhas any1 tried making their own usb-otg cable?
00:45:47 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
00:45:47_FireFly_oops the line select.c:80 must be moved to settings.c:1593 (before if(gui_syncselect_do_button(&select, button)))
00:45:49JdGordondo the 2 data wires need to be swpped? or green to green and white to white?
00:46:09XavierGryou mean sync and charge?
00:46:50 Join LinusN [0] (
00:47:34amiconnhi LinusN
00:50:07XavierGrOh Hiyo Linus
00:50:49 Join leftright [0] (
00:52:19*amiconn now has a working volume setting in dB
00:52:23 Quit cwncool ("CGI:IRC (Ping timeout)")
00:52:32amiconnThere are 2 minor problems though.
00:53:11_FireFly_amiconn: have you got the needed changes to stop the wrap around for integer-options??
00:53:22amiconn(1) For archos recorders an Ondios, very low numeric values don't fit in the status bar space for the volume icon
00:53:23_FireFly_or should i make a patch
00:53:55amiconn(MAS3587 and MAS3539 go down to -115 dB which means 4 chars)
00:54:27amiconn(2) If an old .cfg file is loaded, it sets the volume way out of range.
00:55:04JdGordon_FireFly_: y stop wrap around of int options??? it just makes for a PITA :(
00:55:19linuxstbJdGordon: Can you test a build for me on your H300?
00:55:33_FireFly_JdGordon: ??
00:55:36XavierGrwhat a PITA?
00:55:46JdGordonpain in the ass
00:56:02linuxstbJdGordon: It's here:
00:56:10XavierGrPita is pie in Greek :)
00:56:18linuxstbAs you can probably guess, it enables reading the date/time from the RTC (I hope).
00:56:31linuxstbSorry -
00:57:37JdGordoncant find my reset button pusher :(
00:57:52JdGordonthe clock is shown :D
00:58:03linuxstbDoes it show the right time?
00:58:19JdGordonno, but ill set it correct
00:58:29JdGordonit likes to flash 00:00
00:58:31amiconnBlergh, why is the player wps broken??
00:58:35JdGordondone it twice already
00:58:56JdGordonhaha wed 2035 :D
00:59:19linuxstbThe time/date setting in Rockbox doesn't work - you'll need to set the date/time in the iriver firmware.
00:59:25JdGordonah ok
00:59:28linuxstbIt's read-only at the moment in Rockbox.
00:59:38_FireFly_have to go good night everybody
01:00:23LinusNlinuxstb: zip?
01:01:02linuxstbwhat zip?
01:01:15LinusNah, that wasn't the patch?
01:01:26linuxstbNo, just a binary to see if it worked.
01:02:25JdGordondoes the 1.29K fw not have the clock?
01:02:33JdGordoni cant find the options to set the correct tome
01:02:53JdGordonno, found it
01:03:00linuxstbLinusN: I've basically just replaced rtc_read_multiple(1, &rtcbuf[1], 7) in timefuncs.c with a call to pcf50606_read_multiple(0x0a, &rtfbuf[1], 7).
01:03:52 Quit DrumRBoy320 (Read error: 104 (Connection reset by peer))
01:03:57JdGordonyupm yhr coeect time is shown
01:04:56JdGordonbut the time flickers to midnight every so often
01:06:08linuxstbThat doesn't sound good.
01:06:46JdGordonit happens when im looking through the menus... i dont know if it happens if i leave the player not not tho...
01:07:11 Quit Kohlrabi ("Leaving")
01:07:42BBubis it possible to run the cpu of the h1xx/h3xx at any speed you like?
01:08:05JdGordonlinuxstb: seems to have stopped after a reboot..
01:09:23***Saving seen data "./dancer.seen"
01:09:42linuxstbWhat do you mean by stopped?
01:10:24 Join dropandho [0] (
01:10:25JdGordonit didnt flash midnight in the last 2 min i was playing with it...
01:10:41linuxstbSo you mean it's working fine now?
01:10:45JdGordonlooks like it
01:11:06dropandhohey linus! just wanted to see how the tick pcb was commin along
01:11:28linuxstbAt least it doesn't seem to break the h300, so I'll probably commit it when I commit my related ipod changes (the ipod and h300 have similar RTCs)
01:11:46Bagderamiconn: ok, makefile seems to be working, I'll have a test version for you tomorrow to check out
01:12:48*Bagder runs off to bed
01:14:58XavierGrgnight Bagder.
01:17:38 Quit _FireFly_ (Read error: 110 (Connection timed out))
01:19:43LinusNdropandho: i don't have it yet
01:20:24LinusNlinuxstb: the rtc can't be read from a background thread without disturbing the button driver
01:20:38JdGordonthat could be why it was flashing?
01:21:09LinusNi don't know
01:25:02 Quit muesli_- (Read error: 110 (Connection timed out))
01:28:16linuxstbLinusN: What do you mean by "background thread" ? Do you mean any thread apart from main?
01:28:30LinusNi mean any thread at all
01:28:35 Quit Rick (Read error: 104 (Connection reset by peer))
01:28:49LinusNthe adc driver reads from the pcf50606 in interrupt context
01:29:38LinusNso the adc driver can interrupt an ongoing i2c transaction
01:30:21LinusNeither disable the irq or read the rtc in an interrupt as well
01:30:37XavierGrLinusN: how h100 outputs sound. I mean rockbox feeds the uda with what sort of data? WAV? Sorry for the hilarious question but I dont know anything about it.
01:30:49BBubpcm i guess
01:30:54LinusNXavierGr: yes, we feed it PCM data
01:31:53XavierGrso we could always redirect a xxx2wav plugin to transfer the wav data to the chip instead of a physical file right?
01:32:09LinusNin theory, yes
01:33:15XavierGrI am playing with the mid2wav plugin which needs the old xxx2wav files (god bless you that you kept it in the attic)
01:33:30LinusNhave fun
01:33:35XavierGrit is can playback more than realtime in 11025Hz sample rate
01:34:43JdGordondoes any1 know how to rig up a usb F->F adaptor so i can use usb-otg?
01:35:10JdGordonf-f so i can use 2 regular usb cables...
01:38:16 Join SergeantOleary [0] (
01:40:05 Quit BBub (""Nonstop Belze." -")
01:40:43amiconnLinusN: I now have a working patch with volume setting in dB, without limiting and without the option overkill added today. Volume > 0 dB means possible clipping. Wanna try it?">
01:42:35 Join aliask [0] (
01:43:12LinusNno time right now, but i'll have a look later on if i remember it
01:43:46 Quit SergeantOleary ("CGI:IRC (Ping timeout)")
01:44:08amiconnIf I don't commit before the next daily build, the strings in englih.lang need to be deprecated properly.
01:44:23amiconnI just removed them in my patch
01:44:50*amiconn goes bug hunting
01:45:23amiconnThe default player wps doesn't display anything...
01:56:27amiconnLinusN: I've updated my patch (just 2 default values changed)
01:57:40LinusNyou removed all scaling
01:58:11amiconnYou mean the options added today?
01:58:29LinusNyes, and the scaling we had before that
01:58:39amiconnYes and no
01:58:46amiconnI removed the options
01:59:45amiconnPrescaling is done to prevent clipping at lower volume, as long as it can be corrected by boosting the main volume
01:59:57amiconn...but prescaling no longer limits max volume
02:01:18amiconnIf prescaling would require to boost the main volume above the max, we used to limit the volume before. Now prescaling is reduced instead (with the possibility of clipping)
02:01:40LinusNbut no bass/treble scaling?
02:02:15LinusNlowering bass/treble to prevent clipping at high volume
02:02:19amiconnAh, no
02:02:46amiconnThat was never an option for me as long as treble/bass use dB (which they should continue to use)
02:02:50LinusNi thought that was one of the points with the new scaling options
02:03:33amiconnAll these options contradict KISS
02:03:48amiconnEven before we had something happening behind the scenes.
02:04:33amiconnWith these options, much more is happening behind the scenes
02:04:51amiconnThe settings don't represent reality anymore
02:06:09amiconnBefore, only volume was wrong, and in a somewhat consistent way. Now all of volume, bass and treble can be wrong, and even in an unpredictable way if the option to adjust the non-current value(s) is select
02:06:49LinusNi think i agree
02:07:16LinusNhow about the player, is it the same there?
02:07:53LinusNit used to scale the volume if the bass was raised, didn't it?
02:08:27amiconnYes. Before today, player and iriver were handled the same apart from the actual ranges
02:08:50amiconnWith my patch, both are handled the same again, apart from the actual ranges
02:09:07 Join YouCeyE [0] (n=YouCeyE@unaffiliated/youceye)
02:09:19LinusNwhich means that a volume > 0dB may clip
02:09:41LinusNand a volume <= 0dB will not
02:10:02YouCeyEis there fwpatcher.exe for linux?
02:10:05amiconnOne more (slight) difference is that volume on iriver goes -60 ... +24 dB with my patch, accounting for the max. bass boost of 24 dB
02:10:07LinusNYouCeyE: no
02:10:31YouCeyEhow do i go about setting up rockbox on linux for h120?
02:10:33amiconnOn the player, the maximum bass boost is 15 dB, yet the volume goes from -78 to +18 dB
02:11:06LinusNYouCeyE: you have to build and run the mkboot utility
02:11:09amiconn...and not -81 .. + 15 dB
02:11:22YouCeyELinusN, thanks.. any direct webpage help link?
02:11:30LinusNYouCeyE: you want to patch 1.66?
02:11:32 Join Rick [0] (
02:11:34amiconnThe datasheet says the volume is specified as -78 .. +18 dB, so I just adhere to the datasheet here
02:11:41YouCeyE1.65 is also fine
02:11:43FebsSo if EQ is flat, the maximum volume will be 0dB? And then if bass boost is applied, the volume will display > 0 dB?
02:12:04LinusNYouCeyE: i can patch 1.66 for you
02:12:25amiconnFebs: No. The maximum volume on iriver is +24 dB with my patch, but values >0 dB *may* clip depending on the bass/treble boost
02:12:30YouCeyEonce i load rockbox bootloader, can i choose what firmware to use? similar to dual booting system either with windows/linux?
02:12:41YouCeyEgreat :) LinusN
02:12:47YouCeyEmay i know how to do that?
02:13:03 Quit thegeek (Read error: 110 (Connection timed out))
02:13:34amiconnSince the probability of clipping also depends on the actual track, there is no point in making the 0 dB point slide (and it would confuse the user)
02:13:39linuxstbYouCeyE: It's described here:
02:13:47LinusNYouCeyE: eu, us, kr or jp version of 1.66?
02:13:56YouCeyEus version please
02:14:12YouCeyEi dont know if there is any big difference except the language
02:14:59FebsSo if EQ is flat (i.e., no bass or treble boost), 24dB is equivalent to the current 100.
02:15:26FebsBut there is no possibility of clipping.
02:15:59YouCeyEthanks LinusN :)
02:16:22amiconnThere is also no possibility of clipping if you have volume at +24, bass at +24 and then play a track that doesn't contain bass
02:16:37YouCeyELinusN, can you please save the history of commands and pm me?
02:17:01LinusNYouCeyE: it's all in the IriverBoot wiki page
02:17:16LinusNin this case, i ran fwpatcher.exe with wine
02:17:34LinusNjust for the kick of it :-)
02:17:36FebsI understand that, but it's contrary to the accept convention for what 0dB represents. I think that this is the same point that preglow made the other day ...
02:18:40amiconnYes, it's debatable, but I think it makes a good indication for the possible clipping limit
02:19:20YouCeyEcool LinusN
02:20:19amiconnI could also define 0 dB to be the real maximum (simple), but then we don't have this nice indication...
02:20:36YouCeyEwhere can i dl tools for scrambling and descrambling.. is it in rockox dev kit?
02:20:51YouCeyEi have mkboot in ubuntu already
02:21:02LinusNYouCeyE: i don't know if they're in the devkit
02:21:07leftrighthmm, now that you have explained your "0" limit, it makes sense
02:21:11LinusNthey're in cvs
02:21:23LinusNi think it makes sense too
02:23:22webguest53amiconn: when i set the volume to 0 db and then set bass boost to +18, will the volume be 0db or change to 18 db?
02:23:43amiconnNo, it will stay at whatever value you set it
02:24:17amiconnIncreasing the bass boost will just do what the user expects - boost the bass
02:24:24leftrightgood, there's no trickery in the background
02:25:04amiconnThe only 'trickery' is the (compensated) prescaling, but that's inaudible
02:25:09webguest53so 24 db (without bass boost) equals todays 100% ?
02:25:31*Febs has a sense of deja vu.
02:25:45 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
02:26:19 Quit Rick ("I don't need to be here.")
02:26:46*amiconn wonders what's going on with wps loading on the player...
02:29:44LinusNi gotta sleep
02:29:55LinusNamiconn: i like your volume approach
02:30:38LinusNi liked the scaling options too, but your approach is more my style
02:30:42LinusNnite all
02:30:45 Part LinusN
02:35:09 Part leftright
02:36:03 Quit aliask ("Chatzilla 0.9.69 [Firefox 1.5/2005111116]")
02:46:16 Quit JdGordon ("Miranda IM! Smaller, Faster, Easier.")
02:57:16YouCeyEIf you have the bootloader installed and your jukebox still boots into the iriver firmware, check to ensure that the rockbox.iriver file and .rockbox directory are both located in the ROOT of your jukebox.
02:57:26YouCeyEwhat is this rockbox.iriver file?
02:57:40linuxstbIt is the main Rockbox binary.
02:57:44YouCeyEi just installed bootloader and copied .rockbox to /media/ihp120
02:57:50YouCeyEstill noluck
02:58:01YouCeyEboots into iriver firmware
02:58:08linuxstbYou need to download a file for the h120 and unzip it to /media/ihp120/
02:58:48linuxstbThe very latest version is always here:
02:59:06linuxstb(automatically rebuilt after every CVS commit)
02:59:45YouCeyEso basically i missed rockbox build files..
02:59:54YouCeyEi thought i didnt need to install them
03:00:49YouCeyEdo i need to copy .rockbox dir to /media/ihp120 or to /media/ihp120/root?
03:01:09linuxstbunzip -d /media/ihp120/
03:01:22linuxstbThe file contains both the .rockbox directory and rockbox.iriver
03:02:00YouCeyEmy bad..
03:02:07YouCeyEi unzipped in local dir
03:02:18YouCeyEand didnt see rockbox.iriver separately..
03:02:25linuxstbYes, that can be confusing.
03:04:19linuxstbSorry, the unzip command should be "unzip -d /media/ihp120"
03:04:36YouCeyEi got it working now..
03:04:46YouCeyEjust got confused with rockbox.iriver file
03:04:53YouCeyEhow can i change the theme?
03:05:43linuxstbThe "themes" are limited to the WPS (while playing screen). Press the A-B button to bring up the menu and then choose Browse Themes.
03:06:04linuxstbThere are lots of other options to configure if you want to play though - just explore the menus.
03:06:37linuxstbYou probably want to change the font - General Settings -> Display -> Browse Fonts
03:08:24dropandhooff topic- any of you guys know the arguments for LAME encoding?
03:09:09 Join XavierGr_ [0] (
03:09:12YouCeyEgot it :)
03:09:19YouCeyEthanks all of you guys..
03:09:23YouCeyEu r amazing
03:09:25linuxstbdropandho: lame -?
03:09:27***Saving seen data "./dancer.seen"
03:09:59dropandhoopensource mp3 encoder
03:10:58linuxstbdropandho: I know :) I am suggesting you type "lame -?" to get a list of options
03:11:18*amiconn blames _FireFly_ for breaking the player wps (committed by TiMiD)
03:11:34dropandhohehe- oops! thanks
03:12:32linuxstbamiconn: I'm guessing that was broken a while ago - I'm surprised no users have complained. Or are there no Player users left in the world?
03:12:50amiconnI'm not sure
03:13:27amiconnThere are probably not many users which test dailies
03:13:40linuxstbAh yes - there are users:
03:13:48amiconnMost of them are happy with a release
03:14:02amiconnSome even run oldish releases
03:14:09amiconn(like 2.2)
03:15:01amiconnlinuxstb: Ahahaha, I even read that forum article, but didn't get that he means the wps...
03:15:14 Part Sando
03:15:55amiconnI noticed it myself today, while testing my volume patch...
03:15:55XavierGr_okay need a little math help here. I want to find what percentage of x is y.
03:16:21XavierGr_Now how was it....
03:18:16XavierGr_I think I found it.
03:18:44webguest53y/x*100 ?
03:20:31XavierGr_yes that's what I tried. God it was that easy and I couldn't remember it. My brain is buggy again.
03:24:22webguest53amiconn: is there a way to convince you that having an option to automatically lower the bass- and trebleboost on high volume is a good thing an should stay in rockbox?
03:24:46 Quit XavierGr (Read error: 110 (Connection timed out))
03:24:49 Join DJDD_ [0] (
03:25:36 Join JdGordon [0] (
03:26:30 Nick XavierGr_ is now known as XavierGr (
03:26:50 Join Midgey34 [0] (
03:30:56 Join elinenbe [0] (
03:33:07 Quit Sucka ("a bird in the bush is worth two in your house")
03:48:12 Join psycho_maniac [0] (
03:48:42 Quit DreamTactix291 (Read error: 110 (Connection timed out))
03:53:35 Quit JdGordon ("Miranda IM! Smaller, Faster, Easier.")
04:04:13 Join Sando [0] (
04:07:37 Quit Lost-ash (Remote closed the connection)
04:07:39 Join Lost-ash [0] (n=ashridah@
04:14:36 Quit CheeseBurgerMan (Read error: 104 (Connection reset by peer))
04:16:03 Join Jungti1234 [0] (n=jungti12@
04:16:12Jungti1234hi all
04:17:23 Join CheeseBurgerMan [0] (n=BurgerBo@
04:17:30Jungti1234hi CBM
04:23:32Jungti1234Doesn't rockbox support 'Apev2 tag'?
04:27:07 Join golf7 [0] (
04:31:59 Quit dropandho ()
04:33:34 Join [CBM] [0] (n=BurgerBo@
04:50:24 Quit CheeseBurgerMan (Read error: 110 (Connection timed out))
04:50:30 Nick [CBM] is now known as CheeseBurgerMan (n=BurgerBo@
04:54:40 Quit psycho_maniac ("CGI:IRC")
04:55:07 Nick Lost-ash is now known as ashridah (n=ashridah@
05:01:46 Quit Nibbler (Read error: 104 (Connection reset by peer))
05:02:14 Join Nibbler [0] (
05:03:23 Quit Rob2222 (Read error: 104 (Connection reset by peer))
05:03:36 Quit Nibbler (Read error: 104 (Connection reset by peer))
05:03:37 Join Rob2222 [0] (
05:03:59 Join Nibbler [0] (
05:05:45 Quit Nibbler (Connection reset by peer)
05:06:12 Join Nibbler [0] (
05:06:16 Quit Nibbler (Read error: 104 (Connection reset by peer))
05:07:55 Join Rick [0] (
05:09:31***Saving seen data "./dancer.seen"
05:15:02 Join colombian [0] (
05:16:22 Quit Midgey34 ("Download Gaim:")
05:21:47 Nick ashridah is now known as Lost-ash (n=ashridah@
05:23:35 Join Nibbler [0] (
05:25:49 Join talaus [0] (
05:28:15 Join DreamTactix291 [0] (
05:28:17 Quit paugh ("Leaving")
05:30:02 Join Paul_The_Nerd [0] (
05:33:25 Join lostlogic [0] (
05:37:56 Join lostlogicx [0] (
05:38:37 Part colombian
05:53:38 Join lostlogic_ [0] (
05:54:28 Quit lostlogic (Read error: 110 (Connection timed out))
05:57:45 Nick lostlogic_ is now known as lostlogic (
05:58:58lostlogicThoughts on using Java for the profiling output processor?
06:00:12 Quit Paul_The_Nerd ("Chatzilla [Firefox 1.5/undefined]")
06:10:12Jungti1234Applied simulator myself.
06:33:24 Join Soul-E [0] (
06:33:30Soul-Ewhassup whassup
06:40:19Soul-Ewhats that
06:40:45Jungti1234basic PMC-100
06:41:38 Quit DJDD_ ("Trillian (")
06:58:05 Quit webguest53 ("CGI:IRC")
07:03:01 Quit StrathAFK (Read error: 110 (Connection timed out))
07:04:35 Join Strath [0] (
07:09:35***Saving seen data "./dancer.seen"
07:11:28 Part Sando
07:24:59 Quit CheeseBurgerMan (Read error: 110 (Connection timed out))
07:26:47 Join lostlogic_ [0] (
07:27:36 Join lostlogi1x [0] (
07:27:36***Alert Mode level 1
07:27:36DBUGEnqueued KICK lostlogicx
07:27:36DBUGEnqueued KICK lostlogic
07:27:36***Alert Mode level 2
07:27:36DBUGEnqueued KICK lostlogic_
07:27:36DBUGEnqueued KICK lostlogi1x
07:27:36***Alert Mode level 3
07:28:45Soul-Eso any news on a US firmware bootloader
07:34:52 Quit dpassen1 ()
07:37:37***Alert Mode OFF
07:42:06 Quit lostlogic (Connection timed out)
07:43:56 Quit lostlogicx (Read error: 110 (Connection timed out))
07:49:34 Quit DreamTactix291 (Read error: 110 (Connection timed out))
08:00:12 Join Gibbed [0] (
08:00:17 Quit Rick (Nick collision from services.)
08:00:26 Nick Gibbed is now known as Rick (
08:09:12 Join Sando [0] (
08:10:48 Join DJDD_ [0] (
08:24:20 Join ender` [0] (i=ychat@
08:31:17preglowwould anyone with a h3x0 care to enable speed adjustment and try it out?
08:31:31preglowno reason it shouldn't work
08:33:59Jungti1234Took a screenshot by simulator :
08:37:15 Join Membrillo [0] (
08:37:23Febspreglow, where is speed adjustment located?
08:37:42preglowseveral places, gimme a sec
08:41:35preglowyou need to add some H300 defines to a set of files, i remember screens.c/h had some
08:41:39preglowand gui/gwps.c
08:42:52preglowyou can probably just change the H100 defines to H300 defines to check if it's working, and i'll do the main fix later when i can
08:43:46FebsAh. Not being a programmer, that's a bit above me. But I'd be happy to test if you want to e-mail me the files I'd need to test it.
08:44:04preglowahh, right, then it's got to wait
08:44:14preglowhaven't got access to a dev environment right now
08:45:07FebsWhen you do, feel free to e-mail whatever needs to be tested. You can use the e-mail address from my wikiname page.
08:45:32preglowi think i'll just commit it when i can, there's no reason it should not work
08:48:08CtcpIgnored 4 channel CTCP requests in 0 seconds at the last flood
08:48:08*amiconn changed his mind concerning the 0 dB volume point on iriver
08:49:22preglowamiconn: hmm?
08:49:24MembrilloUmmm, ive updated the ipodVOL WPS to work with H300 (images resized etc). Its here... Anyone want to test\include it in rb, its here
08:50:19 Nick _Vladoman is now known as Vladoman (
08:50:52amiconnpreglow: I have a patch that again removes the optionitis introduced yesterday, and replaces the volume setting with a dB scale, without capping
08:51:22preglowand what does it do to make everyone happy without an extra option?
08:51:56amiconnI started with my idea to set the 0 dB volume point 24 dB below maximum, to cater for the possible distortion with up to 24 dB of bass boost
08:51:59Bagderthese days we can't change anything and still keep everyone happy
08:52:30amiconn...but now I changed my mind and will set 0 dB to be the real maximum, like the UDA1380 datasheet defines it
08:53:11amiconnIn fact we did almost the same for a long time on the recorders and Ondios
08:54:28amiconnWhile the MAS3587 datasheet calls the real maximum +12 dB, even setting it to (12 - MAX(bass, treble)) doesn't guarantee a distortion-free signal
08:54:51preglowi think defining 0db to be anything but the maximum is ludicrous
08:55:11preglowbut ok
08:55:21amiconnWell, both the MAS3587/3539 and MAS3507+DAC3550 datasheets do so
08:55:54Bagderamiconn: try this =>
08:56:06amiconnMAS3587/3539: max. is +12 dB, DAC3550: max is +18 dB
08:56:09Bagdershould build a compressed image
08:56:12amiconnThe UDA1380 datasheet says the maximum is 0 dB
08:56:14preglowyeah, but they might not be constructed like the uda, it might seem like the amplitude value affects the d/a in another way
08:56:46amiconnSo now I think it's best just to use the scale that the DAC defines. This ensures line level at 0 dB volume setting
08:56:54Bagderamiconn: it doesn't use a proper check so this builds a compressed image even though it fits uncompressed
08:58:41amiconnpreglow: The latest version of my patch will give you -84..0 dB on iriver
08:59:05preglowamiconn: and how does it handle eq?
08:59:17amiconn(-78..+18 dB on player, -115..+12 dB on all other SH1 archoses)
08:59:18preglowi assume no one around here has got a 56k assembler lying around?
09:01:07amiconnpreglow: In order to avoid distortion, it prescales the signal when treble and/or bass are boosted, but it compensates by boosting the main volume instead. If the main volume can't be boosted further, it reduces the prescaling instead (possibly causing distortion)
09:03:02 Join einhirn [0] (
09:03:06 Join _FireFly_ [0] (
09:03:08preglowso it's basically exactly what we do now, but with different way of handling a high main volume?
09:03:16Jungti1234hi Membrillo
09:06:07Bagder"The most severe of these are the vulnerabilities found in curl and the PCRE library used by Safari," said Thomas Kristensen, chief technology officer for security site Secunia
09:06:07amiconnpreglow: Yes, it's almost what we did until yesterday, with 2 changes: (1) no volume cap, (2) a dB scale
09:06:11Bagderoops ;-)
09:06:28_FireFly_Bagder ;)
09:06:42Bagderfunny when ones mistakes wound up in big companies security alerts ;-)
09:07:22Jungti1234Membrillo: How about this?
09:08:51preglowbtw, is it only me, or is queueing broken?
09:09:00preglowwhen i queue an entire dir now, the disk spins up after every track
09:09:07preglowintroducing gaps between ogg files
09:09:36***Saving seen data "./dancer.seen"
09:10:19Slasheripreglow: Hmm, i will try that
09:10:35preglowbeware, though, i haven't tried the latest build yet
09:10:40preglowbut this isn't very old
09:13:20Slasheripreglow: yes, buffering works fine but for some reason disk spins up
09:13:35Slasherii don't know yet why
09:13:56Slasheriprobably playlist updating or something like that
09:16:53preglowthat's probably why i get the ogg gaps as well
09:17:01preglowi get reproduce it when queueing, not else
09:17:40preglowplaylist shouldn't be updated when queueing
09:17:45 Join MrShlee [0] (
09:17:46Slasheriyes, it delays the track skipping
09:17:58Slasheriso the problem is probably with playlist_next blocking
09:18:36MrShleeDoes anyone know actually how many more cycles it takes to decode ogg then it does to decode mp3?
09:19:05SlasheriMrShlee: you can look at the cpu boost ratio
09:19:27amiconnpreglow (and others): Wanna try my volume patch:">
09:19:38MrShleeim looking at batt life more then anything.
09:19:47MrShleeif the cpu is idle/normal/boost
09:20:17MrShleeand cpu life is fairly static for each period.. Id like to monitor the difference.
09:20:44Slasheriamiconn: Hmm, it has volume in dB-scale?
09:21:55Slasheriand compared to a line level signal?
09:22:01Slasherithat looks really good
09:23:24Bagdermarkun: around?
09:23:37preglowamiconn: not yet, but will later
09:23:52amiconnSlasheri: yes
09:24:04YouCeyEhow to copy wps files
09:24:10YouCeyEcustom ones
09:26:24*Bagder has a configure fix pending that removes the sed /c use
09:27:41Bagdershould make rockbox build on freebsd again
09:27:46Bagderand similar
09:29:20amiconnpreglow, Slasheri: The patch currently has one drawback (that was also there before but usually didn't hit) - if a .cfg file contains an out-of-range volume: setting, funny effects may happen
09:29:38 Join DreamTactix291 [0] (
09:29:39amiconnFurthermore it fixes a bug with the scaling on player and iriver
09:29:54amiconn(that was also there before)
09:30:37_FireFly_amiconn why does the cfg-loader doesn't check if the settings are valid ?
09:31:38MembrilloJungti1234: whatever rocks your boat :) I like the battery time left on mine, but yours is more simple, and that might be what you want
09:31:40amiconnThat's another story...
09:32:38amiconnThis will have to change, definitely, if we want to switch to a .cfg file for standard settings
09:33:28Jungti1234:) Yes. It was going to do analogously with iPod.
09:39:28Slasheri.cfg file is a good idea as long as the loading is fast
09:40:04amiconnYes, it has several advantages over the config sector
09:41:42 Join Zagor [0] (
09:42:34 Join Vlad0man [0] (
09:44:23Bagderbtw, I found a nice online book the other day:
09:45:46Bagderto the point and detailed for people not up to speed on how to run open source projects
09:46:12Zagorpost a link on the neuros list :-)
09:46:31BagderI doubt that'd help
09:46:49Bagderthey are doomed to repeat their mistakes
09:47:00ZagorI fear so too.
09:47:30Bagderand now there list has turned into a support list for the existing products
09:53:04Bagdernow my boss tells me I should work on my work ;-P
09:53:20*Bagder is his own boss atm
09:56:20Bagdernow, the downside of working from home is the lack of a coffee machine!
09:56:35*Bagder walks off, cursing
09:58:48 Quit Vladoman (Read error: 110 (Connection timed out))
10:03:04*Bagder reappears, smiling
10:03:22Jungti1234hey guys
10:03:25Jungti1234Wake up!
10:03:53Jungti1234Sleepyheads :P
10:04:16preglowi've been awake for four hours now :/
10:04:38MrShleebtw.. Is anyone elses H3XX running VERY HOT?
10:04:43MrShleemine is fair boiling :S
10:04:50MrShleewell was..
10:09:52Bagderwell, it seems rockbox runs out of battery a lot faster than the original on the h3x0
10:10:05Bagderand there is a reason for that
10:10:20Bagderperhaps that reason helps making the unit slightly warmer
10:16:34preglowweird for a unit that's so close to h1x0 in design
10:16:37preglowi've never run that hot
10:17:00Bagderthe h3x0 seems to be powering things a lot more flexible and software controlled
10:17:19Bagderso there is a chance rockbox is doing something slightly wrong atm
10:17:53Bagder(all this based on my interpretations of info from Linus)
10:19:06 Nick Lynx_awy is now known as Lynx_ (
10:21:57 Join KN|stiff [0] (i=phhome@
10:24:02MrShleeit sometimes takes 3-4 tries to boot into original firmware. is that the back screen problem?
10:24:26BagderMrShlee: what bootloader version?
10:24:48Bagderthere is a v3
10:24:58Bagderthat might work better, it has been reported so by some
10:26:26Lynx_my archos recorder has been stolen last week :(. How much would I need to pay for a H320?
10:27:12MrShleeGood question
10:27:47Bagder~300 USD
10:28:48preglowarchoses get stolen? :P
10:29:09BagderI got my previous stolen too
10:29:21Lynx_preglow: it was in my backpack, along with a cellphone, a digicam, an usb harddisk and other stuff :((
10:30:55Lynx_yes, ouch
10:39:05Lynx_hmm, still has h340, for 349 eur
10:39:41preglowsomeone with a h300 please test whether playback speed adjustments work:
10:40:03XavierGramiconn: What if the user sets the bass to maximum and volume to maximum as well? Won't this reproduce clipping?
10:40:18preglowit will possibly clip, yes
10:40:34preglowwith bass on +24db, it will most certainly clip
10:40:59XavierGrIf it clips we will have users running here that Rockbox sounds like "crap" (remember the other thread?) and it clips while iriver fw is nice.
10:41:14XavierGrBecause of their ignorance they are going to blame rockbox
10:41:31preglowand because of their ignorance, i am going to ignore them
10:41:39preglowsomeone put up a faq or seomthing
10:41:51MrShleeXavierGr.. you CAN blame Rockbox, its in alpha and made line by line..
10:42:05MrShleethe rockbox might have problems but they WILL get fixed
10:42:10MrShleethats the difference.
10:42:11XavierGrYes but it will be a bad hit for no reason.
10:42:27preglowXavierGr: i believe in giving the users who know what they're doing the tools to do so
10:42:33XavierGrMrShlee: I think you missed the point
10:42:37Bagderthe clueless masses will remain clueless to the end of time
10:42:41preglowXavierGr: i wont limit rockbox because some people don't care to find out how stuff works
10:43:20XavierGrstill an option to cap the bass seems nice for the ignorant masses. I don't care because I never use eq.
10:43:29preglowit's not like rockbox was started to win a popularity contest anyway
10:43:31MrShleeI personally think sound quality is better. the volume gain is less and it sometimes skips
10:43:37MrShleeyou lose some you win some.
10:45:44XavierGrpreglow: Some had other opinion on this as I can remember. Some said that they want Rockbox more user friendly and idiot proof. Again I just say that an option is better for the newbies, either way I know that Rockbox will be better.
10:46:07markunBagder: I'm here now
10:46:19Bagdermarkun: I committed a fix for the sed case in configure
10:46:28markunAh, thanks!
10:46:43preglowXavierGr: an option isn't better for the newbies, an option is never better for the newbies
10:46:54preglowan idiotproof default is better for newbies
10:47:15preglowand i don't want rockbox to end up idiotproof
10:47:21preglowit'll probably suck for people who know they're doing, then
10:47:29XavierGrwell you right you said it.
10:47:41XavierGr^you are right
10:48:38preglowanywho, if we end up with a default that allows clipping, we put up a faq entry on it, and let people educate themselves
10:48:46preglowpeople who don't care to read it, will get ignored
10:48:57*preglow needs asm56000.exe :/
10:49:18 Join Polo_o [0] (
10:50:47linuxstbThe main thing I don't like about the audio scaling is that it is done in secret. Rockbox should always be transparent.
10:51:45Jungti1234Ϻ  ξ ڴ!
10:51:49Jungti1234hi markun
10:52:06markun+n :)
10:52:11preglowlinuxstb: agreed
10:52:31XavierGrmarkun: how the porting on the Gigabeat is going?
10:52:36preglowlinuxstb: i'll try to some ipod coding later, btw
10:52:38linuxstbpreglow: Will your ipod button driver need to communicate via i2c from an interrupt?
10:52:45preglowlinuxstb: very probably
10:53:08linuxstbSo we'll have the same problem as the H300 when reading the RTC.
10:53:22preglowrtc is connected via i2c?
10:53:36preglowi thought it was just a couple of plain registers mapped somewhere in memory
10:53:46linuxstbYes - it's on the pcf50605 PMU
10:54:05linuxstbThe H300 has a pcf50606 (they are almost identical)
10:54:15preglowthen hooray
10:54:44preglowbut yeah, we access the rtc like an ordinary register in ipodrockbox right now
10:54:52preglowor is there some i2c secrecy hiding behind it?
10:54:57linuxstbThat's a different rtc...
10:55:03preglowoh, there are several
10:55:25linuxstbIt's not really an rtc - the one we read now is equivalent to Rockbox's current_tick variable.
10:55:35linuxstbi.e. it's reset to zero when you reboot.
10:55:47preglowcalling it an rtc is a bit... weird then
10:56:17markunXavierGr: not much progress. I was trying to make a cross-compiler to make ARM Linux binaries from FreeBSD, but I want to make sure I use the same glibc, gcc and linux versions as they did.
10:56:17linuxstbYes - but that's how IPL refers to that register.
11:00:49preglowbut anywho, i'll see if i can't get some general interrupt handler code going afterwards
11:00:57preglowthe code i have now deals only with the tick timer
11:03:36XavierGrmarkun did you saw jlo's post about crosfeed adjustments? Seems that he knows about that stuff.
11:06:16preglowyes he does
11:06:17markunI read it now, but one thing can't be right. By changing -0.2 to +0.3 it is not longer a hight pass filter, but a low pass filter.
11:06:31preglowthat filter needs to go anyway :-)
11:06:35markunyes :)
11:07:20XavierGrso is it possible to have crossfeed without volume loss?
11:07:27Bagdermarkun: try crosstool, neat little thing for cross compiler build for linux
11:07:40Bagderfor anything really
11:08:07markunIt would maybe be better to just leave out the high-pass filter, instead of changing it to a low-pass one
11:08:12linuxstbpreglow: Linus's suggestions for reading the RTC were either to disable interrupts during the read, or to read the RTC from an interrupt.
11:09:07preglowmarkun: he does suggest leaving out hrtf compensation
11:09:38***Saving seen data "./dancer.seen"
11:13:37preglowmarkun: his highpass has the same cutoff as the lowpass, it seems
11:13:44preglowthat makes the code even faster, so i hope that's the case :)
11:16:27 Join yngwi [0] (
11:21:26amiconnpreglow: One point that convinced me to allow clipping is that we do exactly this on the recorder for years. No-one really complained, there was just the occasional 'why does it do this'? question
11:22:04amiconnThen we explained that 0 dB equals 92 %, and going higher might cause clipping
11:22:27amiconn(even without bass boost that is)
11:22:48amiconnWe should just make an appropriate FAQ entry
11:22:49preglowamiconn: if they understood that, then this is going to be straightforward
11:22:51Jungti1234help me
11:23:06Jungti1234Why does error happen?
11:24:57MrShleecross compiler?
11:25:06linuxstbmarkun: Looking at the gigabeat firmware upgrade, the binaries there were compiled with gcc 2.95.3 20010315
11:29:55 Quit yngwi ("Chatzilla [Firefox 1.5/undefined]")
11:30:55XavierGrwhy would someone wants to playback with a sample rate over 44.1Khz?
11:31:05XavierGra signal must be sampled at over twice the bandwidth of the desired signal. For instance, to reproduce signals in all the audible spectrum, which includes frequencies of up to 20 kHz, it is necesary to use DACs that operate at over 40 kHz.
11:34:42preglowwhich they do
11:34:46linuxstbXavierGr: I'm not sure what you're asking. I have lots of 48KHz files which mainly come from digital TV/Radio broadcasts.
11:35:04preglowtodays dacs operate at well above a megahertz
11:35:19preglowwith a somewhat reduced bit depth, heh
11:35:41XavierGrMy question is why use more than 44.1 if the audio spectrum is between 20Hz-20Khz?
11:36:20preglowyou're trying to explain it scientifically, so you're bound to get nowhere
11:36:27XavierGrwhat's the point? You will get bigger filesize and no better quality
11:36:43preglowthe people who claim they need 96khz aren't exactly trying to be objective about it
11:37:41XavierGrHa! So there are people that declare they can say the difference between 96Khz music and 44.1 khz?
11:38:29preglowand i believe a number have in blind tests as well
11:38:37preglowbut again, there are a lot of factors
11:38:40XavierGrany blind tests?
11:38:53preglowand it's not _impossible_ to hear frequencies above 20khz either
11:38:59preglowi can myself, but i don't hear them as such
11:39:05Zagorscientifically, it's not so strange actually. consider a 20kHz sine and square wave. at 40k sample rate, they will look the same.
11:39:05preglowit's more a strange sensation in the ear
11:39:05 Join BBub [0] (
11:39:06Jungti1234Does no one know reason?
11:39:24XavierGrwell how long beyond that? maybe 22Khz, more?
11:40:03preglowZagor: provided you've done your antialiasing properly ;)
11:40:05linuxstbJungti1234: Have you made any changes to your copy of the source code? Have you applied any patches?
11:40:21preglowundo it
11:40:35Jungti1234I patched it.
11:40:45linuxstbIt looks like a problem with that patch.
11:41:06Jungti1234By the way, a person who make the patch is fine.
11:41:13preglowsure, but his patch isnt
11:41:25XavierGrJungti1234: If you just want to compile and don't care about that plugin (rockblox) type make -k to see if you get more errors
11:41:30Zagorpreglow: yeah, well they will both be a sawtooth wave (and sound like it). obviously, with a couple more samples per wave the difference can be recorded.
11:41:44preglowZagor: hell no, they'll both be sines
11:42:33preglowat that frequency it's more or less down to luck what they'll look like anyway, but they should both be sines
11:42:34merbananXavierGr: with higher sf you dont need so sharp cutoff filter at sf/2
11:42:37preglowbut now: lunch
11:42:44Zagorpreglow: sines require post-processing, doesn't it?
11:42:53XavierGrwell, what's the sample rate in human ear after all? It must be a cap there too, right? ;p
11:43:42BBubthe cap is actually at about 20 khz
11:43:50BBuband its lower when you're older
11:44:27XavierGrBBub samplerate is different from audio spectrum
11:44:38BBubi know
11:44:48BBubyou gotta go at least twice as high
11:45:25XavierGrThat's what I read, but some claim that sample rate must be higher than double.
11:45:51XavierGrAnd here I ask. Which is the internal samplerate of human ear.
11:46:09BBubthen it would be about 40 kHz
11:46:21ZagorXavierGr: the human ear is analog
11:46:37BBubbut a 60 year old man might still "sample" at 36 kHz
11:46:57 Quit KN|stiff (Read error: 104 (Connection reset by peer))
11:46:59XavierGrZagor: But it will have a samplerate even if analog.
11:46:59cannardi was just thinking that, its no like the brain can only processes in cycles
11:47:04Zagor40kHz is enough to produce sound at 20kHz. but not enough to reproduce *any* sound at 20kHz
11:47:06XavierGrIt is like the eye.
11:47:14BBubXavierGr: the human ear doesnt sample
11:47:20XavierGrHuman eye can't see more frames than a set number.
11:47:37BBubit has little hairs for every frequency
11:47:41ZagorXavierGr: it's not that simple
11:47:43Jungti1234hey XavierGr
11:47:52BBuband from that it measures the intensity
11:47:58Jungti1234Also, error happens.
11:48:50XavierGrWell the brain processes the hearing so it must devide it into samples, no?
11:49:04Zagorthe brain is not a digital computer
11:49:23Zagorif we heard samples, mp3 compression would not work
11:49:35XavierGrBut that's the way with th ears. AFAIK
11:50:09XavierGrJungti: What you want to do?
11:50:17BBubwell, your eyes dont even see a whole picture
11:50:20Jungti1234Error happens why.
11:50:38BBubyou can only see a small spot and your brain completes the scenery
11:50:50XavierGrJungti: compile the rockblox plugin with the patch or just compile the code?
11:51:02cannardthats how REM sleep works
11:51:09cannardi did it once while concious
11:51:13merbananyour brain is actually predicting the picture
11:51:27Jungti1234XavierGr: Japanese applied patch that make.
11:51:43XavierGrSo that means that I mustn't ask how many megapixels eye resolution is eh? ;p
11:52:06XavierGrJungti: that you made? Or that you applied?
11:52:15merbananXavierGr: well there is one such limit also
11:52:20BBubthe eye has a maximum resolution
11:52:35merbananbased on the properties of the eye
11:52:47XavierGrso eyes have a samplerate and resolution but not ears?
11:53:14Jungti1234I applied it.
11:53:31merbananonly resolution can be derived from the eyes fysical aspect
11:53:55XavierGrJungti: I cannot help then. If it is a patch it means that it is out of date.
11:54:00preglowZagor: how requires post-processing?
11:54:28XavierGrYou can go to the line that the error occured (in the file) and see fo yourself if you can fix it.
11:54:32preglowears have bandwidth, not sample rate
11:54:54pregloweyes have actually got a spatial sampling rate
11:55:00 Join KN|stiff [0] (i=phhome@
11:55:01XavierGrDo we agree then that eyes have samplerate?
11:55:48preglowbut the ear does not detect samples
11:55:53preglowhearing is frequency domain based
11:56:03preglowit works like a filter bank, or a spectrum analyser, if you will
11:56:08XavierGrJungti: If you know C. Then refer to the line of the error in the file. And see if you can fix it.
11:56:17Jungti1234'static struct rgb color = { 0x80, 0x80, 0x80 };'
11:56:37amiconnBagder: I tried your compressed.patch. It seems to work so far. I can't test the firmware atm, forgot my USB cable...
11:56:38Jungti1234sorry I don't program well.
11:57:15XavierGrJungti: How on earth a Japanese patch broke rockblox?
11:57:32amiconnSome remarks:
11:57:37amiconn(1) The build dir is apps/decompressor although the source dir is firmware/decompressor. Any reason for this?
11:57:38Jungti1234I don't know. ^^;
11:58:02XavierGranyway have to reboot.
11:58:12 Part XavierGr
11:58:21amiconn(2) It says 'LD rockbox.elf' where it should say 'LD compressed.elf'
11:58:56amiconn(3) When compiling uclimage.c, it shows the full path instead of just the filename
12:00:09amiconn(4) "Image too big, makes a compressed version!" should probably read "...making.."
12:05:38 Quit Jungti1234 ("bye")
12:06:28 Join Jungti1234 [0] (n=jungti12@
12:09:55linuxstbJungti1234: I don't think "struct rgb" exists in Rockbox any more - it was removed a few weeks ago.
12:11:35linuxstbHow old is the rockblox patch you are using?
12:12:49Jungti1234I downloaded here :
12:14:46 Join XavierGr [0] (
12:14:48Bagderamiconn: 1) because the link stuff is made in apps and the makefiles are made to only travserse down its on path 2) really? its the same .elf isn't it? 3) I'll fix 4) I'll fix
12:15:56Bagderah, ok I see the 2) point now
12:16:46 Join thegeek [0] (
12:17:34BagderI'll see what I can do about the 1) as well
12:18:04 Join Zak1392 [0] (
12:18:33Zak1392anybody know what crossfeed is?
12:18:58Zak1392what is it?
12:19:47cannardit mixes a tiny bit of left/right channel to the right/left channel, to simulate listening with speakers
12:19:50cannardinstead of headphones
12:20:13cannardsome people dont like that wierd headphone feel when there heavy panning in songs
12:20:51Membrillowhen installing a new version of rockbox, is it ok to just overwrite the old version? or should you delete the old and put the new one in?
12:21:04Lost-ashcannard: when done right, it can also move audio from inside your head to infront of you.
12:21:18Lost-ashof course, there doesn't seem to be a minute delay in rockbox's crossfeed afaict
12:21:27cannardharder to do with ety's i would say :p
12:21:30Jungti1234it is ok to just overwrite the old version
12:21:45markunLost-ash: there is
12:21:49markun13 samples
12:21:57Lost-ashMembrillo: i always just unpack over the top. rockbox deliberately forgets the settings when someone bumps the version
12:22:04cannardthanks ash, i was hoping someone would chime in with a better explanation than my piss poor attempt
12:22:45 Join hshah [0] (
12:23:39hshahTiMiD: you around?
12:24:28Lost-ashmarkun: hmm. it should be about 60, not 13
12:24:34Lost-ashwait, samples. nevermind
12:24:41*Lost-ash isn't doing the right math ;)
12:26:26 Quit elinenbe (" HydraIRC -> <-")
12:30:17 Quit Membrillo ()
12:31:15linuxstbJungti1234: That Rockblox patch is the same one as this:
12:31:20linuxstbIt's 3 months old...
12:32:28Jungti1234It was old.
12:34:57 Join webguest14 [0] (
12:36:30Zak1392what do the crossfade settings mean?
12:37:19cannardno need for rockbox crossfeeding, just build it into your headphone amp
12:39:33preglowfor those who are actually willing to lug around a headphone amp
12:39:45MrShleeany major River changes in the last week.. worth upgrading the build?
12:39:56Jungti1234hey Zak1392
12:40:33Jungti1234Does this help to you?
12:42:15Jungti1234good night
12:42:26 Quit Jungti1234 ("bye")
12:43:52Zak1392sort of
12:45:56 Quit MrShlee (Read error: 104 (Connection reset by peer))
12:45:58 Quit hshah ("CGI:IRC (Ping timeout)")
12:47:18 Join webguest06 [0] (
12:53:02Zak1392any children of bodom fans here?
12:53:10Bagderamiconn:, all 4 remarks fixed and the conditional set to work "for real"
12:53:21SlasheriHmm, it seems that tagcache becomes a reality :) But at first it will be available only in wps and in future for browsing too
12:53:43SlasheriZak1392: Hmm, not quite fan that much but i like it ;)
12:54:40Zak1392i'm listening to their live cd Tokyo Warhearts on my h340 and it sounds amazing gapless :)
12:54:51Slasherihehe, that's nice to hear :)
12:55:26Zak1392i wish Avenged Sevenfold would release a live cd soon...
12:55:46Zak1392they're another one of my faces :)
12:56:53 Quit webguest06 ("CGI:IRC (Ping timeout)")
12:58:18linuxstbSlasheri: How does it work? Does it generate a tag dabase from scratch?
12:58:19 Join webguest06 [0] (
12:58:31SlasheriZak1392: i like the most industrial/gothic genre but some metal bands like that children of bodom are not bad either :)C
12:58:39XavierGrSlasheri: What is tagcache?
12:58:42Slasherilinuxstb: yep, no pc is needed for that
12:59:00Slasheriit uses the same idea as dircache: do as much on background as you can
12:59:03linuxstbDoes it use the code in get_metadata() to read the tags?
12:59:23Slasherii already have a working cache initialization routines
12:59:23linuxstbCool. So it obviously supports every file format that Rockbox does.
12:59:32Slasheriit takes about 5-10min. to build the cache
12:59:37linuxstb(unlike songdb)
12:59:38Slasheri(10 min. with music playing on background)
13:00:03XavierGra database biult on the go?
13:00:19Slasheriand in future when browsing support is added, it is possible to browse the cache file on disk before it's loaded into ram completly
13:00:26linuxstbWhy does it take so long to build?
13:00:46SlasheriXavierGr: yep, and it's then loaded into ram
13:00:58Slasherilinuxstb: reading metadata from >5000 songs takes some time.. :)
13:01:03ZagorSlasheri: why not expand the existing database rather than build a new?
13:01:08Slasheriand it seeks disk a lot
13:01:13XavierGrthe whole tag database to ram?
13:01:30XavierGrhow big?
13:01:30SlasheriZagor: i have tried thinking that too, but it would be hard to do
13:01:38XavierGrin size?
13:02:03SlasheriZagor: tagdb is not designed to be loaded into ram, and i couldn't even get it working so i could have tried it..
13:02:18SlasheriXavierGr: i have not an idea yet, but it seems to be smaller than the dircache
13:02:34ZagorSlasheri: the solution is not to reinvent the wheel again...
13:02:38Bagderit depend how much tags you have of course
13:02:44SlasheriZagor: yes, that's true..
13:02:47BagderI fail to see how it can be smaller than the dir cache
13:02:52Bagderif all files are tagged
13:02:56Bagderand you have mostly music
13:03:07XavierGrSlasheri: Your multicrafts always amaze me!
13:03:31SlasheriBagder: currently the tagcache file (with _full_ filenames) was ~500 KiB and dircache 450 KiB. But, when tagcache is on ram, it doesn't have the filenames at all (only pointers to the dircache)
13:03:39Slasheriso it must be smaller (at least with my files)
13:03:48Bagderbut the tags are longer than the file name
13:03:54Bagderat least in my files
13:04:06Bagderof course it depends on what tags you keep
13:04:16Slasheriyep, that's weird indeed.
13:04:33Zagor50KB tags for 5000 files is 10 bytes per file. not many tags in your music... :-)
13:04:44SlasheriBut i will know better when i get the ram loading work
13:04:53XavierGrIt can easily climb to 1MB for a full player.
13:05:07Slasherihehe, true. Might be some bugs :)
13:05:36BagderI have full tagging on my 6000+ files
13:05:45Bagderit should be way over 30 bytes per file
13:05:52Slasheriin fact that doesn't make sense at all.. with full filenames, the tag file should have been much bigger
13:05:57Slasheriso there must be a bug
13:06:00Bagderthat's 180K
13:06:05XavierGrBagder then take the patch and test it.
13:06:22BagderI don't use those things
13:06:41BagderI probably use a smaller portion of Rockbox than most users ;-)
13:06:47 Join ^BeN^ [0] (
13:06:49*amiconn does like the idea, but not the implementation
13:07:01amiconnMore code diversity...
13:07:26Bagderin fact, I've been ponder if throwing out the current db related stuff would be good or bad
13:07:50BagderI think it blocks development as it is right now
13:08:00Bagderits there, so no one writes anything new
13:08:10Bagderbut its so broken and hard to use so hardly anyone uses it
13:08:17linuxstbI agree. As it seems to be mostly abandoned and not well understood, I don't see a problem in throwing it out.
13:08:39amiconnImho it's not broken
13:08:49amiconnIt works and is usable
13:08:52Zagorit was very simple and easy to understand when I wrote it...
13:08:56Bagderit has numerous bugs
13:09:11amiconnIt's certainly not the best solution, agreed on that
13:09:13BagderZagor: yes, but someone altered that...
13:09:42***Saving seen data "./dancer.seen"
13:09:47amiconnBagder: The only real bug I know about is the minute-long playlist building when selecting somethin in the song list
13:09:56*preglow never used the db
13:10:11amiconnOther than that, there are some not-yet-completed features
13:10:14Bagderwell, I've not even tried to research the reported bugs, perhaps they're all moot
13:10:25Zagormy whole purpose in writing the thing was to set the tone and direction for the simple "rockbox style" design.
13:10:32Slasheriwhen i tried enabling the tagdb, my player crashed immediately without any message in logf..
13:10:38Zagorso those interested in it could continue to build on it
13:10:39XavierGrfucking allergies. I've been snezzing for half an hour!!!
13:11:11Zagormaybe I was naive
13:11:36Bagderlet's say that HCl and I had different views on how to do things
13:11:42Zagor...and now I've been inactive for too long to be able to complain :-)
13:11:42Bagderand I didn't care to resist all the way
13:11:55amiconnZagor: I agree that the KISS principle is a bit disregarded lately, which I don't like as well
13:12:16BagderI've been meaning to re-implement the db one day
13:12:26ZagorBagder: just go back a bit in cvs ;)
13:12:34preglowcome on, let's just adopt itunesdb!!11
13:13:11Slasherii think that implementing api to the db (for retrieving data) would be a good thing, now most of the could would need to be rewritten for the ram version..
13:13:11amiconnThis is one reason why I made the volume patch that removes the limiting options again, which are imho overkill
13:13:25Bagderamiconn: I'm with you on that
13:13:34XavierGrspeaking of it, do the daily biuld reach back to the firsts builds when rockbox started?
13:13:42ZagorSlasheri: you're missing the point. having two versions is total waste.
13:13:49BagderXavierGr: certainly not
13:14:00BagderXavierGr: we only keep some 30 days or so
13:14:23XavierGrI thought so..
13:15:03Bagderbut everything is in CVS so you can easily reproduce any at will
13:15:19Bagderfromd day 1
13:15:25Bagderat least day 1 in CVS
13:15:33Zagorif we only write new code and never stop to think of purpose and coherence, we'll end up with a multi-MLOC chaos before long
13:17:25*amiconn fully agrees
13:19:20ZagorMillion Lines Of Code
13:21:30Zagorhey, I didn't mean to kill the channel... :-)
13:22:02preglowtoo late, we're depressed now
13:22:20*Bagder transfers lots of data
13:22:32ZagorBagder: how many connections are you up to now?
13:22:32Bagderuseless data
13:22:45BagderI'm still measuring the current implementation
13:22:55BagderI'm a bit amazed of its speed ;-)
13:23:02BagderSummary from 500 simultanoues transfers:
13:23:07Bagder390 out of 500 connections provided data
13:23:14Bagder29863 calls to curl_multi_perform(), average 148 alive. Average time: 700us
13:23:35BagderAverage number of readable connections per select() return: 120
13:23:52Bagder700us is not a lot of time
13:24:09linuxstbI've never used the tagdb, but what can that do that can't be done by automatically building lots of playlists and putting them in a directory structure such as Artist/Album/Genre/Year etc ?
13:24:14Slasheriin fact, i think the new tagdb could work on archos too without dircache (and without loading itself to the ram). But i can't promise yet without testing the performance..
13:26:34Zagorlinuxstb: search
13:26:38amiconnBagder: I hope you will fix the inconvenience in curl I encountered a few weeks ago
13:26:40 Join criis_ [0] (
13:26:47Slasheridircache is only needed for loading the db into ram, so it can also operate directly from disk as tagdb currently does
13:27:31amiconnall: What do you think would be the best way to get a decision on the volume scale/options/etc patch?
13:27:40linuxstbZagor: What kind of searches can you do?
13:27:42amiconnA poll? dev ml?
13:28:10Bagderamiconn: I say dev ml and see if anything sensible comes up ;-)
13:28:44XavierGrml stands for?
13:28:55Febsmailing list (?)
13:28:56Zagorlinuxstb: none in my code, but the data model is designed to do combination searches. i.e. "abba" and 1974 etc
13:29:08XavierGrah yes
13:31:51criis_Hey. Does anyone know if running rockbox on a european H340 is a good idea. I've heard that it runs somewhat well on the US one, but they are different.
13:32:29Zak1392it should be fine
13:35:31criis_Right. I'll give it a shot then.
13:36:37Zak1392mines an intl. version
13:38:35criis_Right. I thought I read about problems with the intl. one but I must remember incorrectly then.
13:47:45Nixsosthere were some booting problems, but they should be fixed with bootloader V3
13:48:48 Quit KN|stiff (Read error: 110 (Connection timed out))
13:49:36criis_OK, cool.
13:49:50 Join webguest95 [0] (
13:50:15criis_Which reminds me. There was a german guy on the mailing list a couple of days ago who had installed it on his 320.
13:50:28criis_That kind of answered my question really.
13:51:15 Join webguest38 [0] (
13:55:53criis_Argh, since when have the changed the korean firmware to some
13:55:57criis_lousy exe-file?
14:01:14_FireFly_criis_ go to the global iriver side there you bekome the normal zip-version
14:02:32BBubi hope he doesnt become a zip-file :P
14:04:30_FireFly_on the korean side the self-extracting-exe-file is in a zip-file ;)
14:08:47criis_Thanks for that link.
14:08:59criis_I admire a man who can navigate iriver's site. :)
14:11:00BBub_FireFly_: what i meant is that become means "to turn into" ;)
14:11:30 Quit Zak1392 (
14:11:30 Quit lostlogi1x (
14:11:30 Quit Maxime (
14:11:30 Quit merbanan (
14:11:30 Quit goa (
14:11:30 Quit wubbla (
14:11:30 Quit Lynx_ (
14:11:33 Quit XavierGr (Read error: 113 (No route to host))
14:12:01NJoinZak1392 [0] (
14:12:01NJoinlostlogi1x [0] (
14:12:01NJoinMaxime [0] (
14:12:01NJoinmerbanan [0] (
14:12:01NJoingoa [0] (
14:12:01NJoinwubbla [0] (
14:12:01NJoinLynx_ [0] (
14:12:51_FireFly_Zak1392 what ??
14:13:35Zak1392what do you mean what?
14:16:02_FireFly_Zak1392 why do you said woah... ??
14:16:08*linuxstb finally has a remote for his h140 :)
14:16:57linuxstbI was expecting the LCD to be larger - Sudoku isn't going to be easy to play...
14:17:57_FireFly_yeah you need scrolling
14:19:14linuxstbWhat's the "ticking problem" ?
14:21:13preglowlinuxstb: not everyone's got it
14:21:21preglowlinuxstb: but it's the data bus leaking over to the audio
14:21:23criis_linuxstb, I imagine it's the old Iriver remote problem where some hardware fault made the remote give a slight (but annoying) tick.
14:22:00criis_Strike that. Apparently it's a software fault (thanks preglow).
14:22:24linuxstbI assume it would be obvious if I had it?
14:23:17 Join XavierGr [0] (
14:23:51 Quit Febs (Read error: 110 (Connection timed out))
14:24:21criis_linuxstb, yup, unless you're deaf. :)
14:24:40criis_It made me return my 140 and get a 340 instead.
14:25:36preglowno, it is a hardware fault, they say
14:26:09criis_preglow, ah, must have misunderstood you.
14:26:39linuxstbDoes logf still work on the remote?
14:26:43preglowlinuxstb: yea
14:27:11 Quit webguest14 ("CGI:IRC (EOF)")
14:29:04amiconnmarkun: The unicode patch causes conflicts against current cvs...
14:29:19*amiconn wanted to try unicode together with compressed.patch
14:30:45 Join Sandking [0] (
14:31:16_FireFly_linuxstb: yepp,
14:36:05Bagderamiconn: that would indeed be the grand test
14:38:14 Join Kohlrabi [0] (
14:47:11markunamiconn: Which did you try, the one on my website? I can make you a new one
14:50:14 Quit Zak1392 ()
14:54:06amiconnmarkun: I tried the last one I downloaded... the conflicts are caused by morse input and other things
14:54:06 Quit webguest06 ("CGI:IRC (EOF)")
14:54:29 Join Febs [0] (n=medifebb@
14:55:45preglowSlasheri: you got any idea about the queueing bug?
14:56:23Slasheripreglow: i haven't been able to test that without remote lcd, i will try when i get home :)
14:56:31markunI will fix it. I'm looking at conflicting code in playback.c, but can't see what the code is supposed to do.
14:57:01 Nick lostlogic_ is now known as lostlogic (
14:57:01DBUGEnqueued KICK lostlogic
14:58:00 Quit Lynx_ (
14:58:00 Quit wubbla (
14:58:00 Quit goa (
14:58:00 Quit merbanan (
14:58:00 Quit Maxime (
14:58:00 Quit lostlogi1x (
14:58:21lostlogicpreglow: profiling works, btw... files play at about 1/4 speed while profiling, but it does seem to do it's job... hopefully today I can write the parser to convert PCs to symbols and make pretty the output.
15:00:00preglowlostlogic: excellent!
15:00:35lostlogicpreglow:"> and Makefile.patch are up-to-date if you want to play.
15:01:02NJoinlostlogi1x [0] (
15:01:02NJoinMaxime [0] (
15:01:02NJoinmerbanan [0] (
15:01:02NJoinwubbla [0] (
15:01:02NJoinLynx_ [0] (
15:01:47linuxstblostlogi1x: Do you have an example of the output of your profiler?
15:01:56lostlogiclinuxstb: yeah, lemme put some up.
15:02:12 Quit pabs ("Reconnecting")
15:02:12 Part criis_
15:02:13 Join pabs [0] (
15:03:07 Quit wubbla (Remote closed the connection)
15:03:12 Join amiconn_ [0] (
15:03:39 Join wubbla [0] (
15:04:37lostlogiclinuxstb:"> a brief profiling run of vorbis decoding.
15:04:38Bagderamiconn: were you gonna elaborate on the "curl problem" you mentioned?
15:05:50 Quit _FireFly_ ("Download IceChat at")
15:05:57lostlogicI think the output may be squirly as I look at it, something like duplicates getting output. I guess bugs are to be expected :(
15:06:25 Join _FireFly_ [0] (
15:09:43***Saving seen data "./dancer.seen"
15:10:46 Quit Lynx_ (
15:10:46 Quit merbanan (
15:10:46 Quit Maxime (
15:10:46 Quit lostlogi1x (
15:10:47 Join lostlogicx [0] (
15:10:58NJoinmerbanan [0] (
15:10:58NJoinMaxime [0] (
15:11:50NJoinLynx_ [0] (
15:16:34amiconn_Bagder: One or two months ago I used curl, for the first time, to test something
15:18:02amiconn_It took me a while to figure out what was going wrong. Imho curl should output the http status when doing a http transfer and something went wrong
15:18:41Bagderwell, in curl's view, a non-200 HTTP code is not wrong
15:18:56amiconn_The code was 501
15:19:09Bagderyou asked for it, you got it
15:19:12amiconn_...and I would have expected at least some message
15:19:26 Quit amiconn (Read error: 110 (Connection timed out))
15:19:27 Nick amiconn_ is now known as amiconn (
15:19:33amiconn(warning, whatever)
15:21:04BagderI understand your point, but curl is more low-level than so
15:21:26Bagderit does have an option if you want errors on http errors
15:22:56amiconnI don't want errors on http errors, but not displaying any hint for a non-200 http status seems a bit strange to me
15:23:06amiconn(unless silent operation is requested)
15:23:26Bagderwell, curl gets the HTTP you asked for, and you got HTTP back successfully
15:23:55amiconnNo, I didn't.
15:24:17Bagderif you didn't get HTTP back, then were was the status code?
15:24:38amiconn501 means 'Not implemented', and I don't count this as a successful transfer
15:24:48Bagderthe HTTP transfer was successful
15:24:54Bagderthe server-side request was not
15:25:38amiconnNitpicking? ;) Yes, the *request* wasn't successful
15:25:45Bagderit is far from nitpick to me
15:25:49Bagderit is the essence of curl
15:26:01Bagderit does HTTP
15:26:09Bagdergetting a 501 back is not a HTTP failure
15:26:27Bagderit is perfectly fine HTTP
15:26:45amiconnOf course not, but afaiu curl is for transferring data to and from URLs
15:26:58preglowyou did transfer data, heh
15:27:11amiconnYes, but not successful, that's my point
15:27:13Bagdersure, if you use it as a kind of browser on the command line, you might feel a lack of such extras
15:27:16preglowit was successful
15:27:25preglowfrom curls point of view
15:27:58amiconnpreglow: If I request a resource, and only get back a status telling me it can't fulfill the request, that's definitely not successful afaics
15:28:10Bagderit depends on what you wanted
15:28:17Bagderlots of people use curl to stress test
15:28:23Bagderthey _want_ the 501 back
15:28:30 Join massiveH [0] (
15:29:38amiconnBagder: If so, I can imagine curl should still tell what status it got back
15:29:50Bagderyou can make it
15:30:02BagderI mean, with options
15:31:35 Part massiveH
15:33:16amiconnI couldn't find such an option... and since the is the -f option to silence HTTP error output, I somehow expected that it would show errors by default
15:33:49amiconn(also -s for silent operation)
15:33:53Bagderyes it does, it just doesn't consider a HTTP 501 response an error ;-)
15:34:35amiconnSo what option should I use to see 50x status codes?
15:35:06Bagderamiconn: I'd say either you'd use -i to get all the headers, or you'd use -w to get the specific code
15:35:43amiconnBtw, afaiu I understand http, 4xx codes mean 'client error', 5xx codes mean 'server error'. Both are errors...
15:36:10Bagderbut you don't seem to understand curl's part in all this
15:36:13Bagderor how I see its part
15:37:07Bagderit is on purpose
15:37:29amiconnI wasn't looking for a specific code, I was investigating what was going wrong. The correct code would have been 201 as I was testing HTTP PUT
15:37:52Bagderbut then you should've checked the server's response headers
15:38:53amiconnYes of course. Somehow I expected curl to display whether something went wrong (informative, without stopping the operation)
15:39:36amiconnThe way curl worked somewhat reminded me of how internet explorer would do it...
15:40:11BagderI disagree, on both
15:42:44markunBagder: Is your compressed.patch finished?
15:43:08BagderI think it works as supposed
15:43:37markunWe could have unicode rockbox tonight :)
15:43:54markunIf amiconn has no problem with it
15:44:22linuxstbmarkun: Will developers have to do anything differently after your Unicode patch is applied?
15:44:37 Quit Sandking ()
15:45:15markunyes, I made a utf8length function which should be used instead of strlen in some cases
15:45:26markunMaybe it should be renamed to something more posix
15:45:49linuxstbSo utf8length returns number of characters, and strlen still returns the number of bytes?
15:46:34markunAnd all strings internally should be utf-8 encoded
15:47:20linuxstbSo does that mean we should use utf8 encoding for the Rockbox source files?
15:47:43linuxstbIIUC the current standard is iso-8859-1
15:48:07markunIt doesn't matter if there are no hardcoded strings
15:48:55amiconnAll .lang files needs to be utf-8
15:48:57markunBut to edit language files you will need a unicode editor anyway, so we could also have utf-8 as standard encoding
15:48:57linuxstbPlugins have hardcoded strings (at the moment). But I would expect them all to be in English and plain ASCII.
15:49:41markunamiconn: Did you try unicode with compress-2.patch?
15:49:47Bagderwcslen() seems to be the commonly used name
15:50:14amiconnmarkun: unicode4.diff has problems.
15:50:26amiconnrecorder/keyboard.c:290: error: `j' undeclared (first use in this function)
15:50:42amiconn(building for recorder v1)
15:50:43markunI thought I changed that..
15:51:34linuxstbBagder: Isn't that for wide characters, rather than utf8 ?
15:52:20Bagderum, perhaps
15:52:30BagderI'm really lost when it comes to unicode...
15:53:28 Quit _FireFly_ ("Friends help you move. Real friends help you move bodies.")
15:54:28preglowit is
15:54:42preglowalmost nothing uses utf8 internally, so you might as well just call it itf8length
15:54:48preglowutf8length, yes
15:55:35linuxstbThere is also the wcswidth() function (for wide chars) that tells you how many terminal columns that string will occupy - which is different to the number of characters...
15:56:30preglowthings can get really hairy if you really want to dive into unicode
15:57:07linuxstbI can't find any POSIX functions for operating on utf8 though.
15:57:55preglowyou wont find it
15:58:04 Join Paul_The_Nerd [0] (
16:00:13linuxstbmarkun: Do you use wide characters as well as utf8?
16:00:56markundouble width chars?
16:02:30linuxstbI mean a string consisting of the decoded utf8.
16:02:54preglowhe only uses utf8, i believe
16:03:07preglowwhich isn't such a big case for us, since we really don't do much string processing
16:03:43chendo__does rockbox do unicode filenames yet?
16:03:44markunwell, we use a decoded string for the bidi code before printing it
16:03:55markunchendo__: almost :)
16:04:14chendo__mmm... i think my filenames are SJIS anyway :(
16:04:44amiconnOn-disk filenames are always UCS-16 with FAT32 afaik
16:04:56amiconnVFAT to be precise
16:04:59chendo__what about ID3 tags?
16:05:41markunchendo__: They can be SJIS if you like
16:05:56chendo__but can rockbox read them properly?
16:06:00chendo__i get garbled stuff
16:06:08markunwe are working on it
16:06:55 Nick chendo__ is now known as chendo (
16:07:16chendohow's USB mode for iriver coming?
16:16:00amiconnBagder: Hmm, now I am confused.
16:17:18amiconnunicode patch together with compressed-2.patch seems to have worked because I get an ajbrec.ajz of the expceted size, however, I can't find the build dir of the decompressor nor any output from 'make' ...
16:19:00Bagderthe decompressor is in firmware/ now
16:19:20amiconnNot here.
16:19:40amiconnI have firmware/common and firmware/drivers as usual
16:19:42 Join _FireFly_ [0] (
16:20:20chendoif i want to develop a plugin for rockbox, where's the best place to start? i've read the helloworld.c and the contributing pages
16:20:52Bagderamiconn: then remove the rockbox.ucl file and do make V=1
16:21:02Bagdershould reveal the details
16:21:22 Join actionshrimp [0] (
16:22:22 Quit actionshrimp (Client Quit)
16:23:57amiconnBagder: The first call to scramble tries to scramble rockbox.ucl - !!
16:24:32Bagderchange order of the dependencies
16:24:35amiconnOf course this leads to approximately the right size...
16:24:45BagderI mean make it $(BUILDDIR)/$(BINARY) : $(OBJDIR)/rockbox.bin $(FLASHFILE)
16:24:50amiconnIt needs to scamble rockbox.bin instead
16:25:20 Join actionshrimp [0] (
16:26:56Bagderand the second scramble uses args too
16:27:00Bagderwrong args
16:27:15amiconnHmm, but now the unicode patch doesn't exceed 200 KB (!!)
16:27:24CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
16:27:24*amiconn wonders why
16:28:17amiconnmarkun: Is arabjoin still disabled for archos?
16:28:26markunamiconn: I forgot to remove the #define NO_ARABJOIN
16:29:11markunI'll remove NO_ARABJOIN altogether, ok?
16:29:11Bagderfixes both scramble invokes
16:29:51chendohow the hell are we gonna get video on rockbox... i'm looking at the jpeg plugin and it's huge
16:30:39Bagderchendo: like apps/video.c perhaps? ;-)
16:31:21Bagderbut if you think jpeg is huge, then I don't know what you expected
16:31:53chendowell, divx/xvid would be a pain, no?
16:32:05chendounless we port ffdshow somehow
16:32:11Bagderwhy would they be more pain than any other codec?
16:32:48BagderI mean, sure there is lots of work
16:32:53thegeeksurely any mpeg4-based codec would be waay too cpu-intensive?
16:33:09Bagderiaudio x5 plays xvid at 13 fps
16:33:10thegeekI assume this is meant for hxxx ?
16:33:16thegeekit does?
16:33:29Bagdersmaller resolution though, but still
16:34:03chendoiriver offical firmware plays xvid...
16:34:11chendoat 220x140 and 10fps though
16:34:36chendoi'd be happy if rockbox plays that :P
16:34:53Bagderthen start working! ;-)
16:35:09chendoi suck at C :(
16:35:29 Quit actionshrimp ("a bird in the bush is worth two in your house")
16:35:37chendomy C++ is only slightly better, i prefer python
16:36:17chendootherwise i was gonna try making an ID3 editor
16:36:36chendoshould be within my capabilities
16:36:50chendois there a full documentation on rockbox's functions though?
16:37:11markunamiconn: did you get my last link?
16:37:21BagderI'm off, back again tonight
16:38:23amiconnmarkun: Yes.
16:38:35markunShould be over the limit again
16:38:39amiconnHydraIRC does have a really useful URL catcher
16:38:57amiconnNo problems in finding URLs posted hours ago
16:39:23chendomy bad. found the API docs
16:40:23markunchendo: what kind of plugin would you like to work on?
16:40:50chendohad a quick glance at the feature request list... ID3 editor seems to be a good one
16:40:54chendounless there is anything easier
16:41:11 Join actionshrimp [0] (
16:42:40_FireFly_amiconn: now i hve a patch to disable the wrap-around of integer-options :)
16:42:41chendowhat does "NOT CURRENTLY SUPPORTED" in the API docs mean?
16:42:52_FireFly_but the is currently down :(
16:45:18chendoclose to 2am, i'm off.
16:45:35FebsI've made some updates to the Main Menu page of the WikiManual:
16:45:51Febsincluding addition of sections for Prevent Clipping, Crossfeed, and Crossfade.
16:46:31FebsSlasheri, if you could either fill in the options chart for crossfade or give me a 2 sentence reminder of what the various functions do, I'd appreciate it.
16:47:33FebsI also changed the "move to next folder" option to "auto change directory" to be consistent with what the option is now actually called ... although I think that "auto change directory" is much less intuitive than "move to next folder."
16:48:14FebsThe "prevent clipping" section contains a disclaimer that the option is still under discussion and may change.
16:49:47amiconnOkay, compressed unicode rockbox seems to be built correctly. I'll test the binary tonight
16:52:15SlasheriFebs: hi, the options are: Enable crossfade: always - enables crossfader for all track changes and new track selection, shuffle - enables crossfading only when shuffle mode is activated, off - no crossfading. Fade in/out delay: How many seconds to wait before starting fading the stream (these options will change to a new name in future). Fade in/out duration: How many seconds the fading should take for the selected stream. If duration is 0, there is no ...
16:52:21Slasheri... fading at all (still delay settings have the effect). Fade out style: crossfade - normal crossfading, mix - fade out duration will only mix the stream to be faded out without fading it
16:52:27 Join leftright [0] (
16:53:58leftrightFebs, there's feature missing from the Replay Gain type option on the wiki, which is "Track gain if shuffling"
16:54:11amiconnSlasheri: With the 'fade offset' idea, you would both save one delay setting and the style setting
16:54:24FebsSlasheri, got it, thanks. Will update the wiki in a few minutes.
16:54:57Slasheriamiconn: one style setting, how?
16:55:03Febsleftright, I haven't yet used replaygain, so I need to work with it before I can update the wiki. Feel free to add it yourself in the interim.
16:55:19FebsAs Linus would say, it *is* a wiki, after all. :)
16:55:32leftrightI would if I knew how, Febs, I tend to break things
16:55:36Slasheriamiconn: i will add that offset, but there is much more fixes to the crossfade that probably should be done before that
16:57:28FebsSlasheri, is there any difference between Fade out style = mix and fade-out duration=0?
16:57:41amiconnSlasheri: Iiuc the style means whether both streams are faded at the same time, or if the second stream is faded in while the first remains constant, then the first stream is faded out while the second one is already playing at full level
16:58:00amiconnAll these things are possible with just 3 delay settings
16:58:15amiconnFade out delay, fade in delay, and crossfade offset
16:58:31 Join Sucka [0] (
16:58:36amiconn(of course the crossfade offset can be both negative and positive)
16:59:26amiconnCrossfade would be completely deactivated if all 3 times are 0
16:59:54SlasheriFebs: There is, in fact i have to check the code.. :)
17:00:06*leftright goes to look for a wiki tutorial
17:00:40Febsleftright, if you click "edit" on any wiki page, it will provide you with links to wiki formatting references.
17:02:37Slasheriamiconn: Eh, interesting.. It seems you are correct, there is no need for the mixmode at all :)
17:02:59Slasherisetting fade out duration to 0, should do the same thing
17:03:15Slasherii don't know what was the idea adding it..
17:03:26Slasheriamiconn: oh, in fact there is..
17:03:42Slasherifade out duration sets the crossfade buffer length
17:03:53Slasheriso that style setting is necessary
17:04:09amiconnImho it's not
17:04:26Slasheriso setting fade out duration to 0, will not mix the songs
17:04:51Slasherimixing both streams without fading is not possible without that setting
17:05:13amiconnThe crossfade buffer length should be fade_in_delay + crossfade_offset + fade_out_delay if crossfade_offset is positive,
17:05:51Slasheriamiconn: with crossfade offset, there is no need for fade in/out delay..
17:06:14Slasheriit's just |crossfade_offset| + fade_out_delay
17:06:17amiconnand MAX(MAX(fade_in_delay, fade_out_delay), crossfade_offset) if crossfade_offset is negative
17:06:59Slasherioh, i mean |crossfade_offset| + fade_out_duration
17:07:10amiconnHmm, it seems I should draw a little image of how I imagine it could (should) work
17:07:26SlasheriHmm, probably that would help :)
17:07:36amiconnErm, replace fade_*_delay im my descriptions with fade_*_duration
17:07:42Slasheriah :)
17:08:13Slasheriamiconn: fade_in_duration is a separate thing and it doesn't need buffer space
17:08:21Slasheriit's processed on-the-fly
17:08:23amiconnIt doesn't?
17:08:35Slasherionly fade out buffer is required
17:08:55*amiconn wonders why/how fade in can be handled on the fly, but fade out cannot
17:09:12leftrightdoes anyone know what gain type is reverted to if "track gain if shuffled is deselected"
17:09:24Slasheriamiconn: fade out buffer is processed completely at once (and fade out applied to it)
17:09:30 Quit actionshrimp (Read error: 110 (Connection timed out))
17:09:44Slasherithen when new data is inserted into pcm buffer, it will be mixed/faded with the fade out buffer that is being played
17:09:45***Saving seen data "./dancer.seen"
17:09:57Slasheriand when fade out buffer is completely played, fading in might still continue
17:10:21amiconnHmm, then it works rather different from how I imagined it would
17:10:43Slasheriyes, it saves some buffer space and is not too complicated :)
17:10:51leftrightok, so I've make a change to the wiki, how do I exit and save changes ?
17:10:55amiconnI don't think it saves buffer space
17:11:26Slasheriamiconn: it there is fade_in_delay (the offset) or fade_in_duration is longer, then all that space is saved
17:11:48Slasheri(in fact that is the after crossfade fading)
17:12:26Slasherithe there might be: fade out - crossfade - fade in parts at once
17:12:46linuxstbleftright: Click on Preview, and then if you are happy, click on save (and tick the "Release edit lock" checkbox)
17:15:22leftrightI assum that the gain mode last used is the mode that is reverted to if shuffle is deslected ?
17:17:38*leftright looks around the room
17:17:47Febsleftright, just looking at the replaygain menus, it seems to me that the only way to deselect "track gain if shuffled" is to select either "album gain" or "track gain."
17:18:13leftrightor deselect shuffle
17:18:25FebsAh, I see your point.
17:18:49FebsIn that case, it seems that "track gain if shuffled" implies that it is album gain otherwise.
17:18:58webguest95amiconn: i like the idea of having only 3 crosfade settings, but i have a question: wouldn't a positive value of crossfade_offset mean that there is an additional gap inserted?
17:19:10leftrightthats what I'm assuming as well Febs
17:19:45amiconnwebguest95: No.
17:20:06leftrightanyway my first wiki addition ever is done,
17:20:14amiconnA gap would only be inserted if crossfade_offset > fade_out_duration + fade_in_duration
17:21:31webguest95amiconn: ah, i thougtht that crossfade_offset would be the overlap of the tracks
17:25:09Slasheriwebguest95: it is
17:25:57Slasheriamiconn: in fact code can't handle that situation, so no gap will be inserted. The fade in delay will be just shortened
17:27:07 Quit webguest38 ("CGI:IRC")
17:31:16amiconnI'll provide a little image later
17:32:19 Quit Paul_The_Nerd ("Chatzilla [Firefox 1.5/undefined]")
17:36:18 Part Polo_o
17:37:51 Join Polo_o [0] (
17:44:10 Join stupid_ro [0] (
17:48:05 Join rooomish [0] (
17:48:08webguest95amiconn: would the new crossfadesettings be like this?
17:48:16 Quit rooomish (Client Quit)
17:49:06amiconnyes that's what I've been thinking of
17:49:51 Join Philip [0] (
17:50:40amiconnThe offset could also allowed to be negative, this would mean both tracls would be playing at full volume during this time
17:52:44amiconnSetting all 3 times to the same value would resemble a 'clasic' crossfade, offset == 0 would resemble mix mode
17:53:13webguest95in my opinion it would be easier if the user would be able to set the overlap (only positive values) instead of the crossfade_offset
17:53:36amiconnHmm, perhaps
17:53:46amiconnThe final result would be the same
17:54:27webguest95i think setting the overlap is more intuitive
17:54:44webguest95(fir me at least)
17:54:48amiconnThen overlap == fade_in == fade_out would be the classic crossfade, and overlap == fade_in + fade_out would be mix mode
17:54:59 Join bobwise [0] (
17:55:11 Quit bobwise (Client Quit)
17:55:42amiconn...and the overlap would be directly connected to the crossfade buffer length
17:55:47 Join bobwise [0] (
17:56:39amiconnDoesn't sound bad, imho
18:02:33 Quit cannard (Read error: 110 (Connection timed out))
18:02:33 Nick stupid_ro is now known as cannard (
18:06:25 Quit bobwise ("CGI:IRC (EOF)")
18:18:40 Quit Kohlrabi (Read error: 104 (Connection reset by peer))
18:18:58 Join hshah [0] (
18:23:57 Part leftright
18:24:45 Join Moos [0] (
18:27:39 Part Nibbler ("Leaving")
18:32:04 Join Drumrboy320 [0] (
18:33:57 Join webguest35 [0] (
18:36:17hshahTiMiD: you around mate?
18:37:21 Quit Drumrboy320 ("CGI:IRC (EOF)")
18:39:15 Quit webguest35 ("CGI:IRC (EOF)")
18:51:33 Join edx [0] (
18:54:11preglowFebs: you here?
18:57:40 Part Polo_o
18:58:14 Quit edx ()
18:58:24 Nick Lynx_ is now known as Lynx_awy (
18:58:39preglowFebs: to test playback speed thing
19:02:56_FireFly_amiconn: the broken wps for the player where also gone if the new wps-file-loader would be commited ;)
19:07:48Febspreglow, loading it now.
19:09:48***Saving seen data "./dancer.seen"
19:11:38linuxstbpreglow: What's new with the playback speed?
19:11:47preglowlinuxstb: i tried enabling it for h300
19:11:54preglowlinuxstb: couldn't exactly remember what defines were needed, heh
19:12:18 Join Nibbler [0] (
19:13:13amiconnpreglow: Why no just check your playback speed commit for H1x0?
19:13:39preglowi think i did, but i'm not a master of cvs usage
19:13:43preglowso might have missed something
19:13:56preglowdid i crash febs' pc or what? :)
19:14:00amiconnThe daily changes pages are very handy for doing this
19:14:11preglowamiconn: well, there you go, i didn't know they existed
19:14:33amiconnDaily builds page -> Source -> Older
19:14:38preglowi have most definitely forgot the playback.c/h ones
19:14:39amiconnGoes back up to 30 days
19:15:23FebsSorry, someone walked into my office! Testing it now ...
19:16:20amiconnmarkun: Do you have some test files with 'exotic' unicode names, or an utf-8 text file, complete with a screenshot how it should look like?
19:16:29Soul-Ewhat kind of job do you have Febs
19:17:03Febspreglow, what is the key combination? PLAY and +/-?
19:17:10preglowFebs: aye
19:17:15preglowFebs: play + up/down
19:18:53FebsDoesn't appear to have any effect.
19:19:02preglowFebs: you don't get a screen or anything?
19:19:25Soul-EWhat kind of lawyer? Criminal? So you can defend me?
19:19:27preglowwhen it works, you should get a screen
19:19:28FebsI'm trying this from the WPS?
19:19:41FebsSoul-E, no civil litigation. I stay as far away from criminal work as possible.
19:20:42Febspreglow, just double-checking, the rockbox.iriver file was the only file that I needed to load?
19:21:20FebsSoul-E, I'm also a working musician.
19:21:24preglowFebs: yeah
19:21:59Philipdoes anyone know the cappacity of the battery in the H3xx is in mAh??
19:22:25preglowFebs: i've found a mistake
19:22:38FebsPhilip, 1300maH
19:24:19preglowFebs: same file again
19:24:22preglowFebs: make sure it's not cached
19:25:35 Join Paul_The_Nerd [0] (
19:26:00FebsTrying to download it now. Your server is timing out on me.
19:26:02Paul_The_NerdDoes anyone else here randomly wake up in the middle of the night with solutions to problems they haven't been asked to solve?
19:26:14Philippaul: yes
19:26:24preglowit answers ping here
19:26:39FebsCould be a routing issue of some sort.
19:26:56Paul_The_NerdI just remembered waking up in the middle of the night with this idea of how to restructure a friend's webpage to make it standards compliant with the newest XHTML without breaking functionality.
19:27:23FebsOK, downloading now.
19:27:45 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:29:58 Join Kohlrabi [0] (
19:30:14 Quit Philip ("bye everybody")
19:31:31Febspreglow, same result. PLAY + up/down has no effect.
19:32:32preglowcan't one press play and other buttons at the same time on h300 or what?
19:32:58 Nick Vlad0man is now known as Vladoman (
19:33:29preglowFebs: excuse me for wasting your time
19:33:34preglowFebs: stand by
19:34:30amiconnpreglow: Pressing PLAY+ other on H3x0 should work
19:34:33preglowi know
19:34:38preglowi'm just being a fucking moron
19:35:00preglowan all too familiar scenario of late
19:35:13FebsFWIW, pressing PLAY + up/down in the file browser properly pages up/down.
19:35:19 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
19:35:40preglowFebs: same thing again, if it doesn't work now, i'll do something drastic
19:36:07FebsStand by.
19:39:40preglowif i ever need a lawyer, i hope i catch him playing with a h300 when i enter
19:39:49preglowlistening to music at an insane pitch
19:43:15preglowhooray for me!
19:43:30preglowbloody getting a set of defines working, now there's a day's work
19:48:01*Febs is listening to a *really* fast version of March of the Toys.
19:48:49markunamiconn: I will make you some tests.
19:48:58preglowjust out for smoke, then i'll commit and try to do some ipod work
19:49:03markunOne thing you can try is some the translations.
19:49:56markunYou will have to go to apps/lang and use to make the utf-8 language files
19:50:09markunDo you have a unicode font?
19:51:32 Join Lear [0] (
19:52:56Soul-Ewhat kind of musician Febs
19:53:26amiconnmarkun: No
19:55:04markunFor all files prefix
19:55:14FebsSoul-E, I run a 17-piece jazz big band. I also work for a local professional theatre, so I a lot of different styles.
19:55:21markunfonts are: and unifont.bdf
19:55:38Febs(I play guitar and bass.)
19:55:40markununicode.txt should look like unicode.jpg
19:56:13markunarabic-bbc.txt should look like arabic-bbc.png (only works with unifont)
19:57:15 Nick Febs is now known as Febs_out_to_lunc (n=medifebb@
19:58:21markunfranken.bdf is a combined font of 12x13ja.bdf and 6x13.bdf
19:59:05markunworks with the korean and japanese language files (which are not included in the patch)
19:59:20markunwikipedia is a good source to get unicode strings btw
20:04:49 Join IRCMonkey_ [0] (
20:05:33 Nick IRCMonkey_ is now known as PcHeRo (
20:09:01LearGrumble... make for a sim doesn't do make tools by itself...
20:09:13LearWait, maybe I forgot to configure... :)
20:09:50PcHeRoanybody know how long it will take for the video plugin for h100?
20:10:34markunamiconn: any luck with the unicode build?
20:10:45*Paul_The_Nerd points to the channel topic.
20:11:46 Join leftright [0] (
20:12:46leftrightLear: for the wiki, if "Track gain if shuffled" is deselected does replay gain revert to album mode ?
20:13:27Leardeselected? If you mean selected, then yes, fall back to album gain (if available).
20:14:19leftrightsorry I meant if shuffle mode is deselected, does replay gain revrt to album /
20:15:30linuxstbPcHeRo: See the topic.
20:15:57PcHeRook thx
20:16:24 Join hoffm23 [0] (
20:16:31preglowi hope people are aware the video plugin wont play divx...
20:16:38linuxstbIt won't?
20:16:50markundamn.. :)
20:16:53LearYes, that's correct.
20:23:25 Join NicoFR [0] (
20:23:51 Quit NicoFR (Client Quit)
20:24:39 Join MiLK [0] (
20:26:17leftrightLear: if in "album" gain and no info is available does it use track gain value as a default ?, and... if in "Track" mode does it use album gain value as default ?
20:26:24 Join DangerousDan [0] (
20:28:13LearAlbum falls back to track, track falls back to nothing (typical case is that you generate either track gain only, or album and track gain).
20:30:43 Quit webguest95 ("CGI:IRC")
20:35:00preglowwoop, gotta go
20:39:21amiconnWiki spam alarm!
20:39:33 Join webguest89 [0] (
20:40:46Bagderamiconn: TexNoel?
20:41:45amiconnCluttered several user pages with hidden text
20:42:18amiconnWould it be possible to forbid hidden text in the wiki?
20:43:57 Join jlo [0] (
20:44:32 Quit webguest89 (Client Quit)
20:45:12Bagderwe reject his IP now
20:45:41BagderI doubt it'll work for long but could give us some room for now
20:46:25jlohello guys
20:48:05jloto preglow, I just answered to you about crossfeed on the forum
20:48:42leftrightI hope that wiki spam alrm wasn't due to my editing the Replay Gain section, I was trying to add spaces to some text
20:48:55Bagderleftright: indeed not
20:49:15Bagderleftright: unless you added 25 hidden URLs in several pages
20:49:37 Quit leftright ("CGI:IRC (EOF)")
20:49:41markunjlo: If I sayd I made a shelving low-pass then I was wrong (but I think I didn't say that)
20:49:51 Join leftright [0] (
20:52:46jlopreglow : no you said lowpass, but maybe I don't understand how it exactly works : for me you add the whole signal (allpass) with a negative highpass, that gives lowpass no?
20:53:19markunare you talking to me or preglow?
20:54:25jlosorry, I just spoke with preglow but you are welcome !!!
20:55:47jloin fact Markun I was talking to you !
20:58:49 Part hoffm23
20:58:59markunhe negative value made it a very inefficient high-pass.
20:59:58amiconnmarkun: I tried 'make zip' with unicode5.patch
21:00:08amiconnThe patched convbdf segfaults...
21:00:22markunjlo: try to compile rockbox. It's not very difficult to change the values of the crossfeed.
21:00:33markunamiconn: That is very strange. A lof of people have been using it..
21:02:16amiconnIt always happens with the same one font: 00-Starmap_r400-8.bdf
21:02:28amiconnAll others are converted
21:02:57markundid you test it without this font then? And also with the unicode fonts?
21:03:21jlomarkun : yes that what I should do but I wanted to change the place of the delay (to delay whole crosfeed signal not only high frequencies) and that's not so easy (I just looked at the dsp.c, registers are rare)
21:03:34markunHm, 00-Starmap_r400-8.fnt has a size of 0 here..
21:03:56amiconnyes, here too
21:04:31amiconnThe zip gets built, so it doesn't stop me from testing, but of course convbdf shouldn't segfault
21:04:35markunjlo: change the C code, not the assembly code. btw, it's the low freqs that get delayed.
21:05:00 Quit mirak (Remote closed the connection)
21:05:52markunjlo: what exactly are you trying to do? Delay the whole sound by 13 samples? That wouldn't be a natural crossfeed.
21:06:49markunmaybe you should just try out your idea and make a patch if you think it sounds better
21:06:50jlomarkun : yes I was wrong it's low freq, I wonder why you haven't delayed all freqs
21:07:45amiconnI would also think delaying all frequencies by a constant time would be natural
21:08:12DreamTactix291would 13 samples even be detectable by human ears?
21:08:12 Join solexx___ [0] (
21:08:24jlomarkun : no the 13 samples that Preglow mentionned is just to compensate dsp calculation in one branch
21:08:39amiconnThe speed of sound is practically constant across the audible frequency range
21:08:51markunjlo: I wrote the code, so I'm pretty sure why I added the 13 sample delay :)
21:09:51***Saving seen data "./dancer.seen"
21:09:57markunThe theory is that your head blocks (some of) the high frequencies from the speaker on the other side of your head.
21:10:21jlo13 samples is impossible to hear, but it just gives some "peigne" filtering (in french !) when added to a non delayed signal
21:11:03markunIt is just an approximation, but I did some listening tests and preglow too and we settled for these values.
21:11:28amiconnThen I would think the cross-feed of these higher frequencies need to be attenuated (more) than for the lower frequencies, but the delay should remain constant
21:11:56markunI made the values ajustable in my own player and the sound was very different when I changed the delay. It changed the direction where the sound came from..
21:12:03jloyes the head gives a delay of about 350 ms from signal coming from a speaker thats opposite and some high frequency attenuation
21:12:22markunamiconn: That's exactly what I am doing I think
21:12:47markunjlo: are you sure it's that big?
21:12:58amiconnjlo: 350 ms?? Should this read 0.35 ms
21:13:02jlomarkun : no you delayed only low frequencies
21:13:11markunI read it was 0.3ms
21:13:17markun300 mu-s
21:13:43jlomarkun : yes 350 microseconds (0.35 milliseconds)
21:14:05markunwhomany samples in a 44100Hz signal would that be?
21:14:18dwihnoA question for you techies... Does S-ATA disks require some additional drivers to get optimal performance on Windows?
21:14:19 Join yngwi [0] (
21:14:59_FireFly_dwihno: not the disk but the host afaik
21:15:06markunFor 0.35ms I get 15 samples.
21:16:24markunjlo: maybe we should stop talking about the theory and you should just implement it how you think it is right.
21:16:38jlomarkun 441OO*0.00035 = 15 or 16 samples
21:16:56dwihno_FireFly_: so there's a possibility my disks just are running in some kind of compatible mode?
21:17:32 Quit PcHeRo ("ChatZilla [Firefox 1.0.6/20050717]")
21:17:36markunjlo: And 44100*0.0003 = 13 samples, so that's why I did the delay
21:17:55jlomarkun : you are right, but I would like to have the same delay on all frequencies
21:17:59 Join Cassandra [0] (
21:18:31_FireFly_the disks do nothing only the host-chip, to which the disks are connected could be run in a slower mode because maybe the standard windows-driver didn't support the controller complete
21:18:58markunjlo: just remove the low-pass filter then
21:19:28 Join tucoz [0] (
21:19:35dwihno_FireFly_: are there any ways to check what mode the disks are currently operating under?
21:19:59markunI don't know what kind of delay the 1 pole low-pass filter adds. Don't know much about filters.
21:20:16jlomarkun : no there must be a delay but also a filter (high freq are more attenuated by the head)
21:20:38 Quit solexx (Read error: 110 (Connection timed out))
21:20:51_FireFly_dwihno: i don't know but generally it isn't a bad idea to use the driver from the manufacturer of the controller/mainboard
21:21:38jlomarkun : the delay of the DSP filter calculation, I don't know but I'm quite sure that it takes some samples
21:22:38Leardwihno: I know I needed to install a driver from within windows for full stability at least (the one during install wasn't enough).
21:23:07markunjlo: We both know how it should be implemented, only I don't know enough about filters to do it right. So please try it with the right filters and then we can listen to the results.
21:23:14_FireFly_and at least not all new controller are natively supported from windows
21:24:08dwihnoI'll check for a driver then
21:24:20amiconnmarkun: Hmm, it seems the uncode text files aren't displayed properly
21:24:27amiconn(I use unifont)
21:24:55amiconnDoes the viewer interpet the text as iso8859-1?
21:25:02markunamiconn: that's true. The changes to the viewer are not finished.
21:25:22markunIf you revert to the unpatched viewer it will interpret the text as utf-8
21:26:00markunWe need to add codepage selection in a menu to the viewer. Most of the code is in it already.
21:26:15jlomarkun : Preglow posted on the forum that maybe he 'll try to change the implementation After I can modify the values compile and try. I'll keep you informed
21:26:39markunok, thanks
21:27:46amiconnmarkun: Hmm, now I installed the unpatched viewer. It works unless the viewer applies line breaks
21:28:09markunyes, that was fixed in the patched viewer :)
21:28:15amiconn(switched to wide view in order to avoid this)
21:28:23markunShall I change the patched viewer to default to UTF-8?
21:28:26markunfor now
21:28:29amiconncomparing to your screenshot...
21:29:01markunamiconn: did you try some language files? greek, russian..
21:30:30amiconnNo, not yet
21:30:48amiconnarabic-bbc.txt looks quite different from your screen shot
21:31:07markunHm, that's not good. Can you make a screenshot?
21:31:22markunMaybe it's not joined..
21:31:39dwihnoLear: I'll follow your lead and install some chipset drivers, let's see what happens then.,
21:31:43markundid you try unicode6.diff.gz?
21:32:20amiconnThis is unicode6.diff, minus the viewer
21:32:59amiconnThe begining looks like Firefox displays arabic-bbc.txt
21:33:36amiconnScrolling right in the viewer seems to have weird effects
21:33:59tucozmarkun, I found something (that you might have seen) on crossfeed.
21:34:18tucozGo to articles, and click on "natural crossfeed"
21:35:57*Cassandra considers submitting some Rockbox feature requests.
21:36:06CassandraI think we should have Quicktime support.
21:36:11CassandraAnd VRML.
21:36:12CassandraAnd AI.
21:36:18Bagderand touch screen
21:36:32CassandraGood point.
21:36:32Bagderoh, that one was already submitted ;-)
21:36:44*Bagder recalls the days of tracktheripper
21:37:04CassandraI do love the Morse support though. That's wonderfully silly.
21:37:09amiconnBagder: The latest compressed.patch seems to work as intended
21:37:25markunamiconn: yes, scrolling right doesn't work reliably in the old viewer.
21:37:27amiconnI'm currently running a compressed .ajz of the unicode patch
21:37:39Bagderamiconn: I leave it to you to commit or adjust as you see fit, to match your work
21:37:59Paul_The_NerdCassandra: How 'bout a true text to speech engine with multilanguage support? Hehehe.
21:38:02amiconnI'd say commit it as-is
21:38:08markunamiconn: another thing is that the text is RTL, but looks strange because it's left aligned
21:38:13BagderI'll do it
21:38:29CassandraWhat I really can't understand is why we don't have an emacs plugin yet.
21:38:53CassandraOh yeah, and Linux compatibility.
21:39:20amiconnCassandra: Ehm, because we don't have 8 megabytes of RAM (not on all targets at least)
21:39:30CassandraI want to be able to log in to my Jukebox over the headphone cord.
21:39:35amiconn(regarding emacs)
21:39:57amiconnOh, and of course we don't have virtual memory support ;)
21:39:58CassandraBy whistling 1200baud carrier.
21:40:09ender`well, iRivers have at least 16MB ram, right? :)
21:40:18CassandraSurely that's not too much to ask?
21:40:20tucozmarkun: it looks like the author of the article uses different delay's for different frequences. (in some way)
21:40:51amiconnmarkun: Firefox also shows the text left aligned.
21:40:52markuntucoz: I give up, let jlo do it :)
21:41:11amiconnBut you're right, the viewer will have to support bidi properly
21:41:19tucozmarkun: fair enough. I like it though. Makes the 60's listenable again.
21:41:43jlotucoz : this article is interesting but misses something, the comb filter effect is valid for sinus measurements but first we have a delay on music and secondly the crossfeed is attenuated so the comb effect is much less
21:41:55*Cassandra doesn't really mind weird stereo recordings.
21:42:19SlasheriCassandra: hehe, 1200 baud afsk would be probably easy to use but i don't see currently very much of use for that =)
21:42:27CassandraAlthough there's a bit in "Mr. and Mrs. Mickey Mouse" by the Bonzo Dog Band that always makes me think my headphones are broken.
21:42:38CassandraIt's in mono on the right channel.
21:43:02tucozjlo: It seems like you know this. Hope you get something out of your tests :)
21:43:17*tucoz knows next to nothing on filters
21:43:22CassandraSlasheri, so you can run a web server via IP over avine carrier, of course.
21:43:45SlasheriCassandra :P or attach iriver directly to some packet radio systems ;)
21:44:02CassandraErm, avian
21:44:11Slasheriwithout the need for computer as terminal :) and then you could type with morse.. haha :D
21:44:53*Cassandra is working on a super secret Rockbox related project.
21:45:02CassandraIf I told you about it, I'd have to kill you.
21:45:20Slasherinice :p
21:45:28CassandraI will ask how the hell you're supposed to copy strings in C++ though, which may provide a little clue.
21:45:33amiconnSlasheri: You could connect 2 irivers (line in - line out and vice versa) and play a two-player game plugin...
21:45:39hshahhas TiMiD commited my WPS?
21:46:15Slasheriamiconn: yep.. then we should probably have some communication protocol driver
21:46:23tucozCassandra, string s = "aaa"; string b = a ?
21:46:28Slasheriamiconn: but do you think that would be really useful?
21:46:30Cassandraamiconn, you scare me. That would actually be no less stupid/cool than the Morse support.
21:46:50amiconnmarkun: Do you have an idea what's wrong (or not) with arabic-bbc? In fact my recorder shows the beginning exactly as Firefox, apart from font differences and a missing dot at the beginning of the second line
21:46:55Cassandratucoz, you'd think, yes.
21:47:02tucozhehe, that went wrong. string s = "aaa"; string b = s;
21:47:10Paul_The_NerdI hadn't thought of two-player games as a use for that. All I'd thought was some crazy scheme to integrate a rockbox equipped player as a peripheral for a PC in some way.
21:47:26amiconnCassandra: Imho morse support can be useful, not only for those who already know it
21:47:27dpassen1Cassandra: using the STL string library? or char* for string?
21:47:37markunamiconn: yes, the problem is that the viewer expects every byte to be a char and can just start in the middle of a UTF-8 encoded char
21:47:51tucozthat works if you use stl at least.
21:48:00Paul_The_NerdWith line in/out, could you potentially set up some sort of token ring networking method, building off of that? Just, as crazy ideas go?
21:48:08amiconnI just tried to input a few letters, and I'm fairly sure this is way faster than the standard vkeyboard with only a little practice
21:48:44Cassandraamiconn, That's why I like the Morse patch. It's a lovely piece of lateral thinking. It really makes sense as an input mechanism for MP3 players.
21:48:59 Quit MiLK ()
21:49:12Cassandratucoz, it's causing my program to dump core.
21:49:43Cassandra(technically the two strings are type wxString, but you get the idea.)
21:49:51CassandraOooh, another hint there.
21:49:53tucozreally? weird. I do that all the time.
21:51:17CassandraHmm. I wonder if my global class is getting initialised properly.
21:51:30markunamiconn: yes, that's how it looks like in my firefox as well
21:52:00Paul_The_NerdFor what it's worth, it looks like that here too.
21:52:05CassandraOh, hold on. Is = calling the constructor?
21:52:22CassandraThat could get messy if the string's already initialised.
21:52:49Cassandra(You know, I thought C++ was supposed to make dynamic allocation easier, not more of a pain in the arse.)
21:52:50_FireFly_Cassandra: it shouldn't
21:52:58amiconnCassandra: Beware, C++ may cause headaches ;)
21:53:17LearCassandra: only when you declare it.
21:53:17_FireFly_wxString only overloads the '='-operator
21:53:41Cassandraamiconn, Oh yes. I've been a C purist until now, but if you want to write wxWindows, it's pretty much gotta be in C++.
21:53:48tucozCassandra, the stl is fairly good documented at and
21:54:17LearwxString isn't STL...
21:54:25 Join matsl [0] (
21:55:06tucozLear, I know. :)
21:56:11CassandraHmm. If I use a dummy wxString from the same routine, it doesn't coredump.
21:56:12dwihnoLear: seems like the intel inf update already has installed the suitable driver.
21:56:23dwihnoLear: Then it's just me wanting more performance ;)
21:56:37CassandraLooks like the fact I'm trying to copy into a global class is causing the problem.
21:57:12_FireFly_Cassandra: do you mean a global object of a class don't you ;)
21:57:20LearCassandra: still, a simple assign should work. Maybe you're doing something that messes up the wxString storage?
21:57:27Cassandra_FireFly_, an instance, yes.
21:57:50LearThe fact that it is gobal probably increases the chances for that...
21:59:30CassandraHmmm. Maybe my instance isn't getting initialised.
22:01:08tucoz...and this has to do with rockbox? My guess is a cross-platform patcher? or maybe simulator...
22:01:44_FireFly_tucoz: we have already a sim ;)
22:01:47 Quit Kohlrabi (Nick collision from services.)
22:01:49amiconnmarkun: I tried some language files now. German works, russian too. Greek doesn't show some menu items at all
22:01:52 Join Kohlriba [0] (
22:02:07tucoz_FireFly_, yes, but that is not the same for win and linux
22:02:24_FireFly_afaik only the gui is different ;)
22:02:43amiconn...and we need a smaller full-unicode font :/
22:02:43tucozYes, then a crossplatform gui would be nice, and easier to maintain.
22:02:54amiconnunifont is huge on the recorder lcd
22:03:45tucozand by using something like wxwindows I believe that is doable. Same drawing code for windows and linux.
22:03:55 Join DrMoos [0] (
22:05:03 Quit Lear ("Chatzilla [Firefox 1.5/undefined]")
22:05:12 Quit Moos (Read error: 104 (Connection reset by peer))
22:05:31amiconnmarkun: Oh, the strings in the button bar (recorders) aren't centered properly if they contain utf-8 chars (e.g. "Men" for F1 in german)
22:06:02 Nick DrMoos is now known as Moos (
22:09:23amiconnI would like to bump the .lng file version with the utf-8 transition, and cleanup all deprecated strings
22:10:16CassandraDamn - if I initialise the config structure, then trying to create a new instance of an element whose pointer is stored in the structure causes a segfault.
22:10:30CassandraI hate C++. Even describing it makes my head ache.
22:10:43amiconnmarkun: Greek seems to be the only language that doesn't work properly
22:11:00SlasheriCassandra: hehe, that's nothing - wait until you see Symbian C++ ;D
22:12:02 Join muesli_- [0] (
22:12:26_FireFly_amiconn: what about the currently "missing" remote-settings ??
22:13:54CassandraI'm allowed to have pointers in my instance of my global class aren't I?
22:14:21 Quit Nixsos ()
22:17:30_FireFly_Cassandra: yepp but are you sure that the pointers holds a address to a valid object ??
22:18:12dpassen1why use a pointer rather than a reference?
22:18:36 Join Midgey34 [0] (
22:18:51 Quit XavierGr (Read error: 104 (Connection reset by peer))
22:20:26CassandraErm, well the line where I'm crashing is supposed to be initialising the pointer as in gv->pwxThingummy = new wxThingummy(parms);
22:20:32Paul_The_NerdAren't they, in the end, the exact same thing anyway? (I mean, one you have to dereference, but yeah...)
22:21:12dpassen1Paul_The_Nerd: references can't be NULLed as far as i remember, and you use pointers with dynamic memory allocation
22:21:42Paul_The_NerdAh. So then what's the advantage of a reference?
22:21:49dpassen1pwxThingummy is public?
22:21:54 Join DrMoos [0] (
22:22:00Cassandragv is.
22:22:17CassandraAh, hold on, I spot a flaw.
22:22:31Cassandragv is just a pointer.
22:22:52linuxstbCassandra: Are you sure you need to use that much C++? I've only played a little with wxwindows, but I managed to get quite far using C for everything apart from the GUI widgets themselves.
22:22:59CassandraIt's being initialised in my init routine, but presumably once the init exits the object it points to is being destroyed.
22:23:09 Join webguest49 [0] (
22:23:28CassandraCan I use GV* gv = new GV(); in a global context?
22:23:37 Quit Moos (Read error: 104 (Connection reset by peer))
22:23:43amiconnThe 'fun' of behind-the-scenes memory management...
22:24:10CassandraC++ - it sucks.
22:24:48CassandraI tried to work around that bu using a malloc(), but a malloc doesn't initialise the elements within the structure, of course.
22:25:04 Nick DrMoos is now known as Moos (
22:25:04 Quit jlo ()
22:25:11CassandraWow, at least I definitively understand what was going wrong, even if I have no clue how to fix it.
22:25:16dpassen1does calloc?
22:26:14 Quit dpassen1 ()
22:26:34CassandraAh, the answer appears to be yes, you can use new in a declaration of global scope.
22:26:48CassandraWhich is nice.
22:27:05tucozwhy would you use malloc in c++
22:27:23Cassandrabecause it worked and new didn't?
22:27:52Cassandra(because I wasn't using new in a global context, as it turned out.)
22:28:16*Cassandra is a lazy programmer. I code around damage rather than trying to fix it.
22:28:21_FireFly_Cassandra: you should avoid mixing malloc/new free/delete
22:28:26tucozI see, well I thought new and delete were instead of malloc and free
22:28:31CassandraI find it's a good way to learn what works and what doesn't.
22:29:11Cassandra_FireFly_, I see why now. However, it's nice to be aware when you can do it safely.
22:29:38Cassandra(Try to avoid using malloc on class instances is the basic idea.)
22:30:25_FireFly_i would say generelly avoid malloc/free in c++-code ;)
22:31:24matsli'd say stay away from new and delete to if possible
22:31:41_FireFly_matsl: why ??
22:32:00tucozCassandra, :)
22:32:04_FireFly_matsl: forget my question ;)
22:32:17tucozfunny title
22:32:23matsl_FireFly_: what? ;-)
22:33:12CassandraHmmm, now why isn't pwxStr->Format("stuff"); causing pwxStr to point to a string that says "stuff"?
22:35:55 Quit DreamTactix291 (Read error: 110 (Connection timed out))
22:35:59_FireFly_Cassandra: why not simply use the << operator
22:36:26 Quit Midgey34 (Read error: 104 (Connection reset by peer))
22:36:58CassandraErm, because I have no real understanding of it.
22:37:07_FireFly_and Cassandra the Format-fn returns a wxString
22:37:17CassandraI've realised why is isn't working.
22:37:21CassandraYeah, that reason.
22:37:22_FireFly_because this function is a class methode
22:39:35*Cassandra goes and looks up <<
22:39:58markunamiconn: I'll look into the centering problem
22:40:22CassandraOh, that's pretty simple.
22:41:05amiconnmarkun: Probably just another place that needs to use lcd_getstringsize() instead of strlen()
22:41:20_FireFly_Cassandra: this operator adds the given values to the string
22:41:24_FireFly_afaik to the end
22:41:36markunGreek used to work, don't know what broke it
22:41:54amiconnDo you get the problem on iriver as well?
22:42:34amiconnI'll try player...
22:44:53 Join Midgey34 [0] (
22:45:00markun-rw-r−−r−− 1 karl karl 13310 Dec 5 22:44 greek-utf8.lng
22:45:03markunit's too big
22:45:22markunThe limit is 13000 bytes
22:45:38markunShall I change it to 13500 or a bit more?
22:45:59amiconnHmm, but why doesn't rockbox complain, or refused to load it
22:46:23 Join muesli- [0] (
22:46:30_FireFly_amiconn: maybe it loads only the first 13000 bytes
22:46:36_FireFly_of the file
22:46:44markunI don't know. It doesn't show the "new language" string like the other files do
22:47:23amiconnI think 14000 bytes would be better (although the cleanup will save some space)
22:47:29*Cassandra grins.
22:47:34CassandraOh good. It's working.
22:47:43CassandraSoon the world will be mine! All mine!
22:48:11markunso evil to take over the world..
22:48:29CassandraI did it once, for a giggle.
22:48:44amiconnCassandra: Take over the world ? With C++ ???
22:49:18markunBagder: The simulater build gives me this: gmake[1]: *** No rule to make target `@gmake'. Stop.
22:49:32markunWhat should I change?
22:54:04amiconnmarkun: The player credits scroller is another place that isn't properly adjusted to utf-8
22:54:36markunamiconn: I wish I could build the simulator..
22:54:47amiconnWhen an utf-8 char is scrolled out to the left, it turns into a ? for a short time
22:55:48Bagdermarkun: odd, lemme check
22:56:44markunBagder: I use 'gmake' because the FreeBSD 'make' doesn't like the Makefile at all
22:57:01Bagderyes, we need gnu make
22:57:35Bagdermarkun: does the makefile look sensible?
22:57:48Bagderit sounds as if the sed still doesn't do right
22:57:53CassandraF***ing windows and it's f*** stupid file permissions.
22:59:15CassandraHmm, anyone know how to persuade windows that no, no-one really is using that file. It just crashed.
22:59:35Paul_The_NerdIs it one of those "The file is currently in use, so you can NEVER USE IT AGAIN" things?
23:00:30CassandraI suspect I may have to reboot to persuade Windows it really isn't being used.
23:00:42CassandraBloody stupid excuse for an OS>
23:01:23tucozBut isn't there a process using it? Soudns strange that you get that error if it isn't being used at all.
23:01:50leftrightsometimes just logging that user on and off helps
23:01:54Paul_The_NerdRebooting doesn't always work
23:02:04tucozBut it is often hard to find out which process that is by looking at the process list.
23:02:08 Quit muesli_- (Read error: 110 (Connection timed out))
23:02:11Paul_The_NerdSometimes the file gets locked in some obscure way that requires deleting it before the OS finished loading. There's even specialized tools for it.
23:02:21tucozhehe, nice :)
23:02:47Paul_The_NerdI have an entry in my context menu "delete on reboot" for just such things now, I've gotten so tired of it.
23:03:06tucoztools not from MS I suppose
23:03:31CassandraWhere are these tools.
23:03:41CassandraI know it's not in use. I was the only one using it.
23:04:00CassandraThe problem is it crashed, so Windows still thinks it's open.
23:04:08CassandraThis never happens in Unix.
23:04:15tucozCassandra, and the process is not in the list?
23:04:19Paul_The_NerdIf I *recall* this is the shell extension I used.
23:04:21tucozso that you can kill it?
23:04:27Cassandra(Except perhaps with automounters, but we don't talk about automounters.)
23:04:29amiconnPaul_The_Nerd: That never happened to me so far.
23:04:31Paul_The_Nerdtucoz: Files can get locked without a process actively using them.
23:04:36CassandraI already killed the process.
23:04:56amiconnWell, sometimes I did get locked files for no obvious reasons. Most of the time this is caused by the windows explorer
23:04:57tucozOh, I see. Didn't know that. Not that often in windows nowadays
23:04:59Paul_The_Nerdamiconn: I've had it happen about 8 or 9 times so far, in various ways.
23:05:14Paul_The_NerdAnd in almost all cases I haven't been able to unlock them.
23:05:29amiconnThe windows explorer may get confused if a file isn't really what it expects it to be.
23:06:04tucozlike a .doc named as .mp3?
23:06:05amiconnA reliable way is trying to check the properties of a 0 byte .wav file
23:06:30tucozI see. Interesting.
23:06:35amiconnThere's a relatively simple method to remove such files
23:06:41markunamiconn: I think I found the problem with greek. The check in language.c is now if(filesize != MAX_LANGUAGE_SIZE) {
23:06:44ender`hint: get a better filemanager than exploder
23:06:50markunprobably should be <
23:07:02tucozI use mc in windows
23:07:16amiconnCall the task manager, kill all (!) explorer.exe processes, then restart one explorer.exe to get the shell back
23:07:42amiconnDo *not* open an explorer in that dir, start a cmd shell instead, and delete the file from there
23:07:54_FireFly_amiconn: afaik if the last explorer process gets killed a new one is automatically started
23:08:10tucozamiconn, not that simple, and not that nice :)
23:08:32amiconn_FireFly_: Not here, but that wouldn't matter in this case. YOu will want your shell back anyway
23:08:36tucozbut if it works, nice tip.
23:09:29amiconnI can kill all explorer.exe tasks, and end up with an empty desktop (as long as no other programs are running) without taskbar, start button etc
23:09:40 Join LinusN [0] (
23:09:53***Saving seen data "./dancer.seen"
23:09:53LinusNmarkun: u there?
23:10:00Paul_The_Nerdamiconn: Which windows are you?
23:10:03markunLinusN: yes
23:10:07amiconnWinXP SP2
23:10:07_FireFly_yeah the desktop depends on explorer.exe
23:10:25LinusNi think the crossfeed setting should be in the Sound settings menu
23:10:37markunme too
23:10:41tucozSo what do you do with an empty desktop?
23:10:52LinusNmarkun: i suggest you move it
23:10:56CassandraI think the crossfeed setting should be implemented properly so we can use it for stereo width.
23:11:18Paul_The_NerdI know my XP has behaved both ways, though I can't recall when the behaviour changed from reloading explorer automatically to not, or vice versa (can't remember which it does now.) Probably a difference between home and pro, or something
23:11:27CassandraShame I know bugger all about sound processing, really.
23:11:33LinusNmarkun: any chance of adjusting the filter to get rid of the attenuation?
23:11:36markunCassandra: you want the delay and filters to be configurable?
23:12:06ender`Win2k/XP/2003 don't reload exploder automatically if it crashes, while 9x/ME do
23:12:11markunLinusN: jlo and preglow will take a look at the some other filters I think
23:12:12amiconnMaybe it depends on the setting whether to start folder windows in a new process (dunno how that's called exactly in english windows)
23:12:20Paul_The_Nerdender' Some form of XP does.
23:12:32ender`really? never seen one?
23:13:04Cassandramarkun, more that I'd like consistency with the Archos port.
23:13:07Paul_The_NerdYeah, my laptop was doing it in XP for a while.
23:13:16amiconntucoz: You can always use ctrl-alt-del to bring up the task manager, and start a new explorer.exe from there
23:13:29ender`task manager = Ctrl+Shift+Esc
23:13:32tucozAs someone noted here a few days ago, the build table lists 34 builds but it says 35 builds underneath it? What is the 35'th build?
23:13:46tucozamiconn, ah, of course. :)
23:13:47ender`ctrl+alt+del = windows security dialog (at least on machines that are in a domain)
23:14:03LinusNtucoz: a virus!
23:14:11amiconnYes, and there is a button for starting the task manager
23:14:15tucozthe rockbox virus
23:14:19Cassandraamicon: KIlling explorer seems to have done the trick, thanks.
23:14:27ender`btw, how do i have to set up the crossfade settings if i just want to mix 3 seconds of the songs?
23:14:46amiconnFor machines not in a domain which use the XP welcome screen, ctrl-alt-del starts the task manager directly
23:15:00ender`yeah, that's annoying
23:15:06LinusNtucoz: i don't know what the 35th build is
23:15:28CassandraUnless you have to turn the welcome screen off, say for fingerprint recognition to work, to pick a totally non-arbitrary example.
23:15:52ender`anyway, Ctrl+Shift+Esc is always task manager
23:15:53tucozLinusN, no biggie. It probably makes the build time for each target a few seconds lower than it actually is. Fake statistics is always good :)
23:16:11amiconnYes, then you'll get the security dialog the same way as for machines that are in a domain
23:16:34amiconnender`: Thanks for the tip
23:17:01ender`want another? Win+Pause = system properties :)
23:17:10amiconnI know some
23:17:20ender`Win+L, the most useful
23:18:10ender`btw, didn't you need a H120 scan with rockbox logo on it?
23:18:17tucozwow, I thought that the win button could only bring up the start-menu. Is it actually used for something else as well.
23:18:35 Join chopped_pork [0] (
23:20:11*tucoz just read what win+l does
23:21:17 Quit _FireFly_ ("Leaving")
23:21:39Paul_The_NerdDoes the right-click-thingy-button do anything useful?
23:22:25CassandraWin + D is incredibly useful. Go on - try it.
23:22:28amiconnender`: That scan should replace the current one in the device comparison chart
23:22:54amiconnCassandra: No use for me
23:23:05Paul_The_NerdCassandra: Somehow, that sounds like something I *shouldn't* do.
23:23:20ender`it's just show the desktop shortcut
23:23:22amiconnIt's just 'minimise all windows'
23:23:29ender`same as right-click the taskbar->Show desktop
23:23:43ender`(use it again to restore the windows)
23:23:44tucozCassandra, that is actually good
23:23:45amiconnNo need for this
23:24:11*amiconn has no single icon on the desktop except the trashcan
23:24:12CassandraYeah. I wasn't being sarcastic, for once.
23:24:25CassandraMe neither.
23:24:33 Part tucoz ("Leaving")
23:24:35 Quit Kohlriba ("Leaving")
23:24:37CassandraAlthough I cheat. I use Objectdock.
23:25:02amiconnI just use that quick launch bar for often used programs
23:25:23*ender` prefers the side button on his mouse + scrollwheel for switching the programs with mouse
23:25:27amiconnFor all others I just use the start menu (or start->execute)
23:25:38ender`you mean Win+R :)
23:25:41CassandraI used to do that, but it got so half my taskbar was taken up with tray icons and launch icons.
23:25:55Paul_The_NerdI created a second taskbar and organized the programs I actually use in it, then got rid of every icon on the desktop, or hid them.
23:26:06*ender` has 37 icons in his quicklaunch bars
23:26:19*amiconn only has 6
23:26:20 Join Jungti1234 [0] (n=jungti12@
23:26:25CassandraSo now it's common apps to the top. Start menu subcategories to the left, running tasks to the right, and taskbar icons top right. And all hidden unless I move the mouse towards them.
23:26:29thegeekyou're not leet until you run your own litestep theme
23:26:33ender`(2 rows + autohide + UltraMon taskbar for the 2nd monitor)
23:26:40thegeekI'm on the 13'th iteration of my theme now
23:26:45thegeekrefined over 3 years
23:26:50thegeekit's perfect
23:26:59 Quit hshah ("Leaving")
23:27:05CassandraBesides, ObjectDock is cool, and does that shrinky growy MacOS X icon type thing.
23:27:06thegeekit's actually the one thing holding me back from switching to linux;)
23:27:19*ender` disliked OS X
23:27:19Jungti1234Good Morning~
23:27:31*ender` dislikes eye candy
23:27:41thegeekeye candy is ok
23:27:47CassandraYou're going to hate the new windows then.
23:27:56Paul_The_NerdI'm actually watching ReactOS for it to finally say "Hey, I run all the programs you want to. So there."
23:27:56thegeekbut only second to performance
23:27:57CassandraEye candy is fine as long as it's functional.
23:28:07thegeekexactly Cassandra
23:28:17ender`reactos has ugly shell
23:28:50thegeekI dont think the actual gui is the main focus;)
23:28:56thegeekthey are still developing the os itself
23:29:11ender`i hope so
23:29:28ender`their filemanager looks like the Win3.1 File Manager
23:29:47CassandraBesides, presumably you can always run $WindowsAlternateShell anyway.
23:30:10Paul_The_Nerdender' I use Windows XP with the shell set to that "ugly" gray theme anyway.
23:30:23CassandraI like the grey theme.
23:30:32Paul_The_NerdI do too. ;-)
23:30:36muesli-thats nerd style ;)
23:30:38ender`i use 2003 with themes disabled and colors set to nice dark green shade
23:31:23Paul_The_NerdI'm still looking for a old tan tower for my desktop. Last time I needed a new one, the most basic looking one I could find was all black. The rest had lick... lighting inside, and windows and other crazy stuff.
23:31:23 Join bobwise [0] (
23:31:39Paul_The_NerdI kinda like my computer to look as dull as possible, for some reason.
23:32:06ender`my current desktop: (note that i never use the icons there, but i'm too lazy to delete them)
23:32:07thegeekmy beloved theme
23:32:19thegeekthat is horrible ender`
23:32:31thegeekso many icons, and _green_ ?
23:32:39Paul_The_NerdScores points for having Xenosaga wallpaper though
23:33:01ender`just finished ep1 on Saturday :)
23:33:58Paul_The_NerdI can't bring myself to finish Ep 2. I don't like it as much. I'll finish it once it gets closer to when 3 comes out
23:34:08ender`also bought ep2 PAL, but i'm playing the downloaded NTSC version, because the movies look ugly in the PAL one (and it didn't let me import the ep1 data)
23:34:39CassandraMine has cats on it.
23:34:45CassandraLike the rest of my house.
23:35:09ender`you mean like this:
23:35:50CassandraYes, except I only have 1 monitor.
23:36:08Cassandra(A Dell 24" widescreen jobbie. It's lovely.)
23:36:11Paul_The_NerdI don't have a desktop wallpaper.
23:36:42Jungti1234ah hahaha
23:36:47Jungti1234me too
23:36:59*ender` is buying another 19" (or maybe larger) LCD next year
23:38:04cannarddell 3007fpw!!!
23:38:34cannardi wanted one of those, but i might wait for the 3007, base is much nicer
23:38:40*HCl has an 26" tft :)
23:38:44cannardand its bloody'yuuuge!
23:38:56ender`probably won't be widescreen though - i prefer my monitors 5:4
23:38:58*Paul_The_Nerd doesn't like LCD monitors.
23:38:59Jungti1234I go to school
23:39:06*HCl goes afk
23:39:07 Quit Jungti1234 ("bye")
23:39:17CassandraThis is about as big as I can handle, given I have to sit 10cm from the monitor in order to be able to read it.
23:39:17Paul_The_NerdThere's always this quality to them, that doesn't look quite right to me.
23:39:37thegeekwidescreen rocks
23:39:49thegeekI'm using a 2005fpw from dell right now
23:39:52thegeekit's awsome
23:39:56thegeekMUCH better than two monitors
23:40:00CassandraI have a complex arrangement involving D clamps, and a triangular wood frame with a wall mount attachment screwed into it.
23:40:02ender`i've got a 82cm widescreen TV, that's enough
23:40:04thegeekyou dont have to deal with all the issues
23:40:22ender`i don't want a widescreen monitor because the windows would be too wide (i always use all my programs maximized)
23:40:56thegeekafter getting a widescreen I tend to multitask with windows _not_ maximized
23:41:08thegeekI used to have most windows maximized before too
23:41:09*Cassandra nods. Yup. Me too.
23:41:17thegeekbut with a widescreen you have the space to keep them windomized
23:41:35Cassandra1920x1600 is just so spacious. ;)
23:41:36 Quit BBub (""Nonstop Belze." -")
23:42:03CassandraBut I should stop trying to inspire techlust.
23:42:09thegeekI went for a 20" widescreen
23:42:18ender`i dislike high resolutions, too - there's just too many problems with Windows when you use Large Fonts (and i don't want to sit 10cm away from the monitor to be able to read the text)
23:42:21cannard 2560 x 1600 baby!
23:42:21thegeekany larger and you can't play games in perfect resolution anynore
23:43:17ender`i've got a PS2 for my games
23:43:32 Quit Febs_out_to_lunc ()
23:43:50Paul_The_NerdEww... ps2?
23:44:15Paul_The_NerdI mean, I have one, but if I had to pick *one* system it wouldn't be that one.
23:45:10ender`what other system has Final Fantasy and Xenosaga? :)
23:45:59cannardsega mega_cd?
23:46:21Paul_The_NerdBah, the PS2 final fantasies have been unimpressive.
23:46:31Paul_The_NerdI'll give you Xenosaga though. Maybe.
23:46:42Paul_The_NerdIt was good.
23:47:43*ender` wishes he could play xenogears on PS2, but he didn't have a PS1 memory card...
23:48:20Paul_The_NerdI never understood why they didn't format the first 150k as a PS1 card.
23:48:31Paul_The_NerdUse two partitions, basically.
23:48:31ender`to sell more stuff of course
23:48:36Paul_The_NerdI suppose.
23:48:46ender`well, there's always ePSXe
23:48:57Paul_The_NerdThat's my general choice.
23:49:04Paul_The_NerdEven if the framebuffer effects slow it down a lot
23:49:12ender`i never understood why the PS2 memcards are so damn slow
23:49:28ender`not on my computer :)
23:49:29Paul_The_NerdBecause they used the cheapest flash memory they could, then charged you out the ass for it anyway?
23:49:59Paul_The_NerdMine does the heat glare in the first cutscene at about 35fps. Well, my laptop does, I suppose I should say. After that the game's fullspeed.
23:50:30ender`59,9FPS at 1280x1024
23:51:29ender`i played most of xenogears without FPS limit, made the text much more bearable
23:51:57Paul_The_NerdYeah, I tend to disable it at those points.
23:52:25Paul_The_NerdPS2 has a glitch with Xenogears anyway
23:52:25 Quit bobwise ("CGI:IRC")
23:52:49Paul_The_NerdEverything is drawn at about 98% opaquity I'd say. You can VERY slightly see through everything to what's below it.
23:52:55Paul_The_NerdSo if Fei walks behind a wall, you can still see his sprite.
23:53:08ender`i need to check that out
23:53:26Paul_The_NerdIt's a *very* faint effect. I didn't notice it first until right before you recruit Rico, but once I saw it, it was noticeable
23:57:19ender`btw, is it just me, or was dubbing in xenogears horrible?
23:57:46Paul_The_NerdHorrible doesn't even BEGIN to describe it. It was Symphony of the Night bad.
23:58:18Bagderwhat is this
23:58:23Bagder#define option_select_get_title(_opt) \
23:58:23Bagder (_opt)->title
23:58:36ender`horrible is the shortest word i could come up with
23:58:37 Join Dima202 [0] (
23:58:52BagderTiMiD: STOP THAT

Previous day | Next day