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 2010-07-01

00:01:10 Quit pamaury (Read error: Connection reset by peer)
00:01:13 Quit Zigtown (Ping timeout: 260 seconds)
00:01:46CIA-6r27201 build result: All green
00:01:46 Quit newipodman (Quit: CGI:IRC)
00:02:02 Join stripwax [0] (~Miranda@87-194-34-169.bethere.co.uk)
00:02:14amiconnwow
00:03:26amiconnBuschel ^
00:03:48*Buschel becomes nervous :)
00:04:54amiconnSee fs task
00:05:30Buschelyay \o/
00:06:33saratogastripwax: ping
00:06:34wodz-1.7MHz more and it will run unboosted :-)
00:06:43amiconnNah
00:06:52 Quit MethoS- (*.net *.split)
00:06:52 Quit Kitar|st (*.net *.split)
00:06:52 Quit Strife89 (*.net *.split)
00:06:52 Quit solrize (*.net *.split)
00:06:52 Quit mc2739 (*.net *.split)
00:06:53 Quit Beta2K_ (*.net *.split)
00:06:53 Quit Zambezi (*.net *.split)
00:06:53 Quit _threeothree (*.net *.split)
00:06:56amiconnReal playback always needs a few MHz extra
00:07:15stripwaxpong?
00:07:43*amiconn is curious what effect the patch will have on a low-iram, broken-cache arm7tdmi
00:07:52amiconnAka PP5002 - will test shortly
00:08:13 Join Zambezi [0] (Zulu@80.67.9.2)
00:08:18 Quit bieber_ (Remote host closed the connection)
00:08:26saratogastripwax: i can't seem to get profiling to work
00:08:30 Nick Zambezi is now known as 16SAAHAE3 (Zulu@80.67.9.2)
00:08:34 Join MethoS- [0] (~clemens@134.102.106.250)
00:08:34 Join Kitar|st [0] (Kitar_st@BSN-143-104-83.dial-up.dsl.siol.net)
00:08:34 Join Strife89 [0] (~Strife89@adsl-80-136-122.mcn.bellsouth.net)
00:08:34 Join solrize [0] (~phr@adsl-71-141-88-179.dsl.snfc21.sbcglobal.net)
00:08:34 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
00:08:34 Join Beta2K_ [0] (~Beta2K@d24-36-97-38.home1.cgocable.net)
00:08:34 Join _threeothree [0] (generic@server1.unitedservers.de)
00:08:34 Join Zambezi [0] (Zulu@unaffiliated/zambezi)
00:08:49Buschelamiconn/wodz: can you give this a short try -> http://www.pastebin.org/368682
00:08:50saratogawhen I try to add the profstart() function, it cannot the function
00:08:50lestatargevaerts: copy/overwrite worked fine...am back on fuze v2 rb 26906...thanks! :-)
00:09:01 Quit mc2739 (Ping timeout: 244 seconds)
00:09:03saratogado I have to do something besides configure a profile build
00:09:13Buschelamiconn/wodz: this should work on CF as well (and gives minor speed up on ARM)
00:09:16 Quit _threeothree (Excess Flood)
00:09:19 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
00:09:26 Join threeothree [0] (generic@server1.unitedservers.de)
00:09:46 Join anewuser [0] (anewuser@unaffiliated/anewuser)
00:09:48 Quit Zambezi (Ping timeout: 244 seconds)
00:10:06 Quit wodz (Quit: Leaving)
00:11:38 Quit saratoga_ (Ping timeout: 252 seconds)
00:13:28 Quit saratoga (Ping timeout: 252 seconds)
00:13:39 Quit ender` (Quit: Do not meddle in the internals of kernels, for they are subtle and quick to panic.)
00:16:17 Quit simonrvn (Disconnected by services)
00:16:53 Join simonrvn_ [0] (simon@72.0.217.63)
00:18:00 Join saratoga [0] (9803c6dd@gateway/web/freenode/ip.152.3.198.221)
00:19:34 Join bieber_ [0] (~quassel@162-78.97-97.tampabay.res.rr.com)
00:20:43amiconnBuschel: There are PP5002 results now. Does that paste apply to svn or on top of patch v8
00:20:44stripwaxsaratoga - hey. so yes, you need to put the profile_start and profile_end (or whatever they are) calls at begin/end of the code you want to profile
00:20:44amiconn?
00:21:32Buschelwodz: if you want to support further tweaking of FS #11445 for MCF5249 you may want to exchange some of the ICONST_ATTR_FAAD_LARGE_IRAM in the hcb_*.h with ICONST_ATTR
00:21:55Buschelamiconn: to svn, already submitted as v09 (minor change)
00:22:33Buschelamiconn: good, even PP5002 is reasonable faster
00:22:33amiconnOn PP5002 this may help as well, although on that one it *may* be better to put time critical code into iram rather than data
00:22:48 Quit kugel (Remote host closed the connection)
00:22:55 Quit Luca_S (Quit: CGI:IRC)
00:25:33Buschelthe patch gives 10-23% speed up on different targets. I did not expect that much gain when I started to play around with the source...
00:29:28amiconnWhich of the hcb_ thingummies is used most?
00:32:34BuschelI did not test yet. I only measured the usage of the called huffman functions, but not the used codebooks...
00:32:42Buschelnext step :)
00:33:13amiconnNo change on MCF5250 with v9. PP5002 result is on fs
00:33:47Buschelgood, so it still works. that was the major test ;) thanks
00:34:35Buschelneed to get some sleep now. thanks for the support!
00:39:15 Quit Buschel (Ping timeout: 252 seconds)
00:43:30***Saving seen data "./dancer.seen"
00:52:44 Quit jgarvey (Quit: Leaving)
00:57:54 Quit bieber_ (Remote host closed the connection)
01:00
01:04:02 Quit petur (Quit: Zzzz)
01:07:38 Quit DerPapst (Quit: Leaving.)
01:10:15 Quit ranma (Ping timeout: 260 seconds)
01:10:52 Quit efyx (Remote host closed the connection)
01:17:19 Join ranma [0] (ranma@mx.tdiedrich.de)
01:18:41 Quit flydutch (Quit: /* empty */)
01:21:39 Quit Dhraakellian (Read error: Operation timed out)
01:21:55 Quit kachna (Ping timeout: 265 seconds)
01:29:03 Quit ranma (Read error: Operation timed out)
01:33:41 Quit anewuser (Ping timeout: 264 seconds)
01:33:52 Join anewuser [0] (anewuser@unaffiliated/anewuser)
01:35:59 Join Dhraakellian [0] (~ntryon@cpe-67-240-248-41.rochester.res.rr.com)
01:42:22 Join ranma [0] (ranma@mx.tdiedrich.de)
01:48:28 Quit Stephen__ (Quit: Leaving)
01:51:52stripwaxsaratoga - (did that make it work for you?)
01:52:31stripwaxoh you also need to make sure the thing your building is compiled with the right flags in the makefile
01:52:45stripwax-DPROFILE_OPTS i think does that
01:52:56stripwax^your^you're
02:00
02:14:41 Quit stripwax (Quit: http://miranda-im.org)
02:21:48 Quit feisar_ (Read error: Connection reset by peer)
02:22:38 Join feisar [0] (jljhook@irkki.fi)
02:22:59 Quit balintx (Ping timeout: 276 seconds)
02:23:06 Nick feisar is now known as Guest5171 (jljhook@irkki.fi)
02:23:11 Quit MethoS- (Read error: Connection reset by peer)
02:26:03CIA-6New commit by jethead71 (r27202): Enable game sounds in PacBox. Sound is OFF by default but can be enabled from the menu. Enable a function for SWCODEC in the middle of the plugin API, ...
02:28:06CIA-6r27202 build result: 6 errors, 0 warnings (jethead71 committed)
02:32:48 Join balintx [0] (~balintx@fibhost-67-58-201.fibernet.hu)
02:34:12 Join Zigtown [0] (~Zigtown@CPE00259ce0fdb2-CM0014f8cc807a.cpe.net.cable.rogers.com)
02:43:34***Saving seen data "./dancer.seen"
02:46:04 Quit Dhraakellian (Read error: Connection reset by peer)
02:46:16CIA-6New commit by jethead71 (r27203): Fix r27202 red. MCF5249 plugin IRAM is already very occupied.
02:48:04CIA-6r27203 build result: All green
02:50:39JdGordonbieber: the editor is looking prettty sweet :)
02:51:26JdGordonhehe I crashed it :)
02:51:36JdGordonenabling "volume changing" crashes with segfault
02:53:02 Quit Zigtown (Ping timeout: 276 seconds)
02:57:37 Quit balintx (Ping timeout: 260 seconds)
03:00
03:02:18 Join Dhraakellian [0] (~ntryon@cpe-67-240-248-41.rochester.res.rr.com)
03:05:09 Join balintx [0] (~balintx@fibhost-67-58-201.fibernet.hu)
03:09:53saratogastripwax: (maybe the net split blocked my messages) I added those functions and the correct define, but gcc complains that they cannot be found
03:11:23 Quit hebz0rl (Quit: Ex-Chat)
03:13:48bieberJdGordon: Thanks for the heads up, I'll take a look at it tonight
03:14:05JdGordonalso, is the AA image being resized correctly?
03:15:09bieberYeah
03:15:21bieberIt pays attention to the alignment flags and everything
03:16:39bieberHow did you get it to segfault? Volume changing is working fine for me
03:17:18JdGordonit might be the theme... AF_flow for the d2
03:17:26JdGordonenabkling the volume chaning check box crashes
03:17:55JdGordonxpos and ypos for AA doesnt seem to update auomatically?
03:18:20bieberOkay, I'll check that theme
03:18:36bieberSometimes the preview doesn't update right away when the text changes, I'm not really sure :/
03:18:53bieberFor some reason the cursorChanged event just doesn't get thrown sometimes
03:19:26JdGordonisnt there a textchanged event?
03:19:37JdGordonor draw at the end of the parse?
03:20:15JdGordonnegative values for viewport positions dont work also
03:22:15JdGordon-8 for y should put it 8 pixels from the bottom
03:22:25JdGordon-8 for x should put it 8 pixels from the right
03:22:54JdGordon(or left if %ax is in front of it and it is using a RTL lang)
03:26:58bieberOh yeah, I forgot about that
03:27:29 Quit BlakeJohnson86 (Ping timeout: 276 seconds)
03:27:31bieberI'll go ahead and take care of negative positioning now, do images and album art also do that, or just viewports?
03:27:46JdGordononly viewports
03:27:57JdGordoneverything else is viewport relative so no need
03:28:00bieberI'll take a look at the event chain, too. Rendering is _supposed_ to be attached to textChanged, but we'll see
03:28:24JdGordonyou can use negative width/heights also which pretty much does the same
03:28:40 Join BlakeJohnson86 [0] (~bjohnson@2002:1876:a27b:0:227:13ff:fe65:1262)
03:28:43JdGordon-10 width means 10 pixels from the right edge also
03:28:53bieberOh, wait, AA is just ignoring it's x/y positions
03:29:03bieberAnother item on the to-fix list
03:36:28 Join simonrvn [0] (~simon@210.52-ppp.3menatwork.com)
03:38:34 Quit simonrvn_ (Ping timeout: 252 seconds)
03:38:44 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
03:39:58bieberOkay, looks like it's not the mv tag, it's the image display that's being triggered when %mv returns true
03:42:36bieberOh wait, duh
03:42:51bieberAny conditional with parameters could crash it like that
03:42:54*bieber facepalms
03:45:26CIA-6New commit by bieber (r27204): Theme Editor: Fixed segfaulting conditional bug
03:47:20CIA-6r27204 build result: All green
03:53:51bieberJdGordon: Does the status bar count as a line of text, or will text just appear under it if you start it on the first line?
04:00
04:04:28JdGordonit doesnt count as a line
04:04:35JdGordonI think...
04:07:42CIA-6New commit by bieber (r27205): Theme Editor: Implemented negative x/y/w/h values in viewport declarations
04:07:53bieberOkay, I'll just leave it as it is for now
04:08:52bieberAlso, should an empty line cause a newline to be rendered? If so, what about lines that only have tags on them that don't display anything?
04:09:32CIA-6r27205 build result: All green
04:13:00 Join Tarvis [0] (~Tarvis@106.203.100.97.cfl.res.rr.com)
04:19:40 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
04:24:45 Quit pixelma (Disconnected by services)
04:24:45 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:24:46 Quit amiconn (Disconnected by services)
04:24:48 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:24:59 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:25:08 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:27:19 Join Zigtown [0] (~Zigtown@CPE00259ce0fdb2-CM0014f8cc807a.cpe.net.cable.rogers.com)
04:27:49JdGordonbieber: I tihnk they do, and yeah you get blank lines
04:28:47bieberI recall CustomWPS saying that a line with only one tag by itself and a newline won't render an empty line
04:29:59CIA-6New commit by bieber (r27206): Theme Editor: Made blank lines render newlines in the output
04:31:55CIA-6r27206 build result: All green
04:32:11 Quit Zigtown (Ping timeout: 265 seconds)
04:43:37***Saving seen data "./dancer.seen"
04:46:43 Quit TheSeven (Ping timeout: 260 seconds)
04:46:52 Join Barahir_ [0] (~jonathan@frnk-590fc06d.pool.mediaWays.net)
04:49:46CIA-6New commit by bieber (r27207): Theme Editor: Fixed album art positioning bug, removed unneeded debug code
04:50:38 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:50:40 Quit Barahir (Ping timeout: 264 seconds)
04:51:29CIA-6r27207 build result: All green
04:55:51 Quit Dhraakellian (Read error: Connection reset by peer)
05:00
05:01:23 Join Dhraakellian [0] (~ntryon@cpe-67-240-248-41.rochester.res.rr.com)
05:06:01 Quit Dhraakellian (Ping timeout: 265 seconds)
05:07:31 Join Dhraakellian [0] (~ntryon@cpe-67-240-248-41.rochester.res.rr.com)
05:12:54 Join RockNewb [0] (www-data@giant.haxx.se)
05:14:00RockNewbhey guys
05:14:42RockNewbis there a new forum for Sansa v2 now that the old thread is locked?
05:15:57RockNewbi cant get no staisfaction
05:16:42RockNewbIS ANYONE THERE
05:16:55RockNewbor, HOW DO I TELL IF ANYONE IS THERE
05:17:01RockNewbhelp
05:17:14Mode"#rockbox +o JdGordon" by ChanServ (ChanServ@services.)
05:17:20JdGordonask a question or get kicked
05:17:28JdGordonand have some patience
05:17:37RockNewbhank you
05:20:34RockNewbBasically, the sansa v2 thread is locked, but i looks like the port is finished, but is it supported by Rockbox Utility?
05:23:14Mode"#rockbox -o JdGordon" by ChanServ (ChanServ@services.)
05:23:40JdGordondoesnt look like it
05:24:12RockNewbthank you so much.
05:38:12 Join S_a_i_n_t [0] (S_a_i_n_t@203.184.0.128)
05:38:20 Quit Horscht (Quit: Verlassend)
05:44:42 Quit Tarvis (Quit: /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit)
05:45:03binaryhermitsome of the sansa v2s are supported by rbutil
05:45:16binaryhermiteverything but the fuzev2 iirc
05:45:46binaryhermitand I'm not sure whether the latest version supports the fuzev2
05:48:56 Join Strife89DS [0] (~nds@adsl-80-136-122.mcn.bellsouth.net)
05:49:10 Quit Strife89 (Remote host closed the connection)
05:50:38 Quit Strife89DS (Client Quit)
05:51:02 Join Strife89DS [0] (~nds@adsl-80-136-122.mcn.bellsouth.net)
05:51:42RockNewbthanks binary hermit. Does anyone know if the sansa c200 v2 is supported by rbutil?
05:57:39CIA-6New commit by jethead71 (r27208): PacBox: Premultiply sound prom data on load rather than during emulation. Use 16-bit data for 'raw' output instead of int.
05:59:20CIA-6r27208 build result: All green
06:00
06:04:33mc2739RockNewb: according to this wiki page: http://www.rockbox.org/wiki/TargetStatus#New_Platforms_Currently_Under_De Fuzev2 is supported by the Rockbox Utility, but c200v2 is not
06:06:02binaryhermitc200v2 is also listed as "unusable"
06:14:06 Quit Strife89DS (Max SendQ exceeded)
06:18:47 Join esperegu [0] (~quassel@145.116.15.244)
06:21:38 Join Strife89DS [0] (~nds@adsl-80-136-122.mcn.bellsouth.net)
06:28:37 Join lmjubuntu [0] (~lmjubuntu@58.195.42.183)
06:29:45 Join Boldfilter [0] (~Boldfilte@adsl-82-152-180.jax.bellsouth.net)
06:30:00 Part Boldfilter
06:31:55ranmabinaryhermit: C200v2 works pretty well, you just have to do the install manually for now.
06:32:35binaryhermitwell, it is listed as unusable
06:33:08binaryhermitnot having a c200v2, I assumed that "unusable" meant major issues
06:33:25RockNewbyeah, I was wondering why the forum thread was closed before the various statii were synced
06:33:45RockNewbI just did a manual install and it seems to work
06:38:47 Quit RockNewb (Quit: CGI:IRC (Ping timeout))
06:42:10 Part lmjubuntu
06:43:39***Saving seen data "./dancer.seen"
06:44:05 Quit BHSPitMonkey (Quit: Ex-Chat)
06:44:17ranmabinaryhermit: You can help with the rbutil support by testing this C200v2 bootloader: http://forums.rockbox.org/index.php?topic=25048
06:44:45binaryhermithow? I don't have a c200v2
06:45:45ranmaAh, that should rather have been directed to RockNewb, but he already quit
06:46:02 Quit t0rc (Quit: Leaving)
06:46:57 Join lmjubuntu [0] (~lmjubuntu@58.195.42.183)
06:53:54 Quit soap (Ping timeout: 265 seconds)
06:55:43 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
07:00
07:11:31 Part lmjubuntu
07:13:20 Nick scorche` is now known as scorche (~scorche@rockbox/administrator/scorche)
07:14:48 Join Strife89 [0] (~Strife89@adsl-80-136-122.mcn.bellsouth.net)
07:16:15 Quit Strife89DS (Max SendQ exceeded)
07:27:47 Join soap [0] (~soap@rockbox/staff/soap)
07:43:23 Join LinusN [0] (linus@rockbox/developer/LinusN)
07:46:55 Join Buschel [0] (~~andree@p54A3EE2B.dip.t-dialin.net)
07:56:54 Quit binaryhermit (Read error: Connection reset by peer)
07:57:10 Join binaryhermit [0] (~binaryher@99.141.196.197)
08:00
08:01:51 Join Zigtown [0] (~Zigtown@CPE00259ce0fdb2-CM0014f8cc807a.cpe.net.cable.rogers.com)
08:06:47 Quit BHSPitMonkey (Quit: Ex-Chat)
08:16:12 Join funman [0] (~fun@rockbox/developer/funman)
08:21:30 Quit S_a_i_n_t ()
08:22:19 Join ender` [0] (krneki@foo.eternallybored.org)
08:26:54 Quit rvvs89 (Read error: Connection reset by peer)
08:27:11 Join S_a_i_n_t [0] (S_a_i_n_t@203.184.3.101)
08:32:33funmansaratoga: ping
08:37:14 Quit Buschel (Ping timeout: 264 seconds)
08:43:43***Saving seen data "./dancer.seen"
08:45:41 Quit anewuser (Quit: http://xrl.us/detroitVI Sixth edition of 25 samples + 48 hours + detroit party (july 31) (maybe streamed online!))
08:58:41bieberJdGordon: Do we no longer have the tag that dumps the playlist in the current viewport?
08:58:44 Quit Strife89 (Ping timeout: 260 seconds)
08:58:52JdGordon%Vp
08:59:08bieberIt's not in the manual any more :O
09:00
09:01:20JdGordonwas it ever?
09:01:22 Join fyre^OS [0] (~nnscript@cpe-69-203-151-59.si.res.rr.com)
09:01:40 Quit fyrestorm (Read error: Connection reset by peer)
09:01:43 Join Rob2223 [0] (~Miranda@p4FDCAFA8.dip.t-dialin.net)
09:01:43 Join yawny [0] (user36@pr0.us)
09:02:14bieberI'm not sure. I thought it had all of the WPS tags documented
09:02:16 Quit AlexP (*.net *.split)
09:02:44 Quit elcan (Ping timeout: 260 seconds)
09:03:00 Join jfc^3 [0] (~john@dpc6682208002.direcpc.com)
09:03:03 Quit sneak (Remote host closed the connection)
09:03:54 Quit Rob2222 (Ping timeout: 260 seconds)
09:04:12 Join ender [0] (krneki@foo.eternallybored.org)
09:04:31 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
09:04:36 Join AlexP [0] (~alex@rockbox/staff/AlexP)
09:05:05 Join sneak [0] (debian-tor@gateway/tor-sasl/sneak)
09:06:45 Quit jfc (Ping timeout: 265 seconds)
09:07:05 Quit ender` (Ping timeout: 260 seconds)
09:07:05 Quit FOAD (Ping timeout: 260 seconds)
09:07:10 Join FOAD [0] (~dok@83.160.60.104)
09:07:27 Join flydutch [0] (~flydutch@host129-160-dynamic.15-87-r.retail.telecomitalia.it)
09:07:43 Quit advcomp2019 (Ping timeout: 265 seconds)
09:07:44JdGordonbieber: is that the last tag you need to implement?
09:07:46jhMikeSfunman: the correct DS for as3525v2 is 1.13 or something else?
09:08:01bieberUmm, I think it's the last major feature that's altogether missing
09:08:12bieberI still have a lot of work to do with all the special cases in the config dialog, though
09:08:18JdGordonnuts :p Vp is a PITA
09:08:47funmanjhMikeS: i have 1.13 which i believe is the latest, but i believe other versions are correct too
09:08:49bieberIt is, but I figure I can cheat a little, since I don't have to bother with real data :P
09:09:06JdGordonactually, if we can parse/render it properly we should be able to change it to only need one code block
09:09:25JdGordon(something for later)
09:09:26funmanjhMikeS: btw there's no as3525*v2* datasheet, we use the one for as3525
09:10:01JdGordon%Vp(0,%?ia<%ia|%fn>)
09:10:04JdGordoneek! :p
09:10:26bieberAre conditionals allowed in the Vp tag?
09:10:44JdGordonnot in svn
09:10:47JdGordonyes in your parser
09:10:48jhMikeSfunman: maybe I'm getting confused by all this v1/v2 stuff. is there actually a v2 or is this because of v2 players?
09:11:08bieberOhh, because the arguments are just parsed as code
09:11:11S_a_i_n_tUm...conditionals in a %Vp line?!?
09:11:17JdGordonthey are only not in svn because that would really mean reimplemeneting the whole parser for the tag
09:11:25S_a_i_n_thow the heck does *that* work lol?
09:11:32funmanjhMikeS: i call it as3525v2 because it looks like as3525 but there are small differences
09:11:41JdGordonS_a_i_n_t: %ia is replaced for each track as normal
09:11:50S_a_i_n_tHmmmm.
09:12:10jhMikeSfunman: lol, what the heck is it...really?
09:12:12JdGordonwe might make the 2nd code block optional instead of forcing it to one line
09:12:13bieberMy plan is to just write a really simple mini-renderer that will ignore anything unsupported in the arguments
09:12:17funmanjhMikeS: dunno
09:12:46jhMikeS?? it's unmarked?
09:12:48funmani suppose an as3525 extended either by request of SanDisk, or by non-availability of the old components (a research design?)
09:12:54JdGordonbieber: well each param is just a LINE or SUBLINES so cant you just feed that back into the same renderer?
09:13:30funmanit's marked SanDisk
09:13:58 Join [Saint] [0] (~St.]@203.184.3.101)
09:13:59funmanthere are pictures of clip+/fuzev2 on the wiki, i didnt check if it was the same marking than clipv1/fuzev1
09:14:02bieberI could, but it's a little more complicated than everything else, since the tag values are shifting as it iterates through the entries
09:14:04jhMikeSI see...must be the cool thing to do now
09:15:04*TheSeven is reminded of the s5l8700 vs. s5l8701
09:16:04*[Saint] can't seem to get autojoin channel working in irssi
09:19:35funmantoni (Antonius Hellman) was given svn access but never committed? :/
09:19:38 Join einhirn [0] (~Miranda@p548516F2.dip0.t-ipconnect.de)
09:20:05jhMikeSI think he did a couple times but I'm not sure
09:20:41funmannope, all his patches were committed from Flyspray by other people
09:21:21funmani think mono mp3 encoding was broken in r9498/fs#5033 if not before
09:21:23ranmafunman: Since http://forums.rockbox.org/index.php?topic=25048.0 was tested on both revision 1 (mine) and revision 0 (seani), I'd like to upload it to download.rockbox.org, however I've never done that, how do I?
09:22:00 Join Zagor [0] (bjst@rockbox/developer/Zagor)
09:22:22funmanranma: you must go through a Swede (Bagder or Zagor), but i wonder if we shouldn't wait for USB to be stable and make new bootloaders for fuzev1/clipv1/e200v2/c200v2 at the same time?
09:23:06funmanwas this bootloader made before the SD changes and does it have a version number (and not r2xxxx) ?
09:23:39jhMikeSfunman: mono mp3 is broken?
09:23:51pixelmaAlexP: manual - in the WPS tags appendix the hold status conditional description is currently only included for targets with hold switches, This once was true but now you can also use it on soft hold targets and get the status. Can you think of a more "generic" description so we can save the \opt? (currently the headline is "Hold switch" which would have to be phrased differently for soft hold)
09:23:57funmanin the plugin yes
09:24:17funmani have a workaround (fake right channel) but i'm not sure if it's enough
09:24:43ranmaIt was made before the SD changes and is versioned "1.0", hmm, I guess that should have been 1.1 anyway, since the one at http://download.rockbox.org/bootloader/sandisk-sansa/c200v2/ already is "1.0"
09:24:57funmanthere is a bootloader already ? Oo
09:25:15funmanbut it doesn't work on variant 1 that's it?
09:25:39ranmaWell, it'll probably work, but have a black display during the bootloader stage.
09:26:03 Join Strife89 [0] (~Strife89@adsl-67-57-51.mcn.bellsouth.net)
09:26:04 Join Steamcade [0] (WinNT@adsl-86-1-119.tys.bellsouth.net)
09:26:07jhMikeSfunman: mp3_enc.codec is based off it, and it works (tried it just now).
09:26:11 Join kachna [0] (~kachna@r4ax178.net.upc.cz)
09:26:38SteamcadeStupid question. i'm working on updating my various setups. ipod Video, is it 5G or 5.5G?
09:26:40funmanranma: I think we said bootloaders must use major revisions (2.0 etc)
09:26:43SteamcadeiPod Video 30GB
09:26:49funmanjhMikeS: but rockbox always feeds stereo data
09:26:58ranmaAh, ok.
09:26:59funmanLRLRLR, mono wave are LLLLLL
09:27:58jhMikeSit does mono mixdown or copy
09:28:24funmanwindow_subband1 assumes channels are packed
09:29:52funmanif i replace LLLLLL by L?L?L?L? i have something correct but the quality isn't up to stereo so there must be something else
09:30:24jhMikeSrecording just does LRLRLRLR -> MMMMMMMM, maybe it's supposed to have that?
09:30:38funmandoes it?
09:30:44jhMikeSyes
09:31:04funmancan you point me where?
09:31:06Steamcadeanyone? iPod Video 30gb. is it 5G or 5.5G. googling it give sme coinflicting reports.
09:31:12jhMikeScheck starting around line 2087
09:34:35ZagorSteamcade: I'm not sure what the definition of 5.5G is. the 2048 byte sectors?
09:34:50Zagorwe just call it ipodvideo and ipodvideo64mb
09:34:56Steamcadehrm
09:35:09SteamcadeI'm trying to install Hybrid
09:35:17Zagordoes rbutil use the 5G and 5.5G names?
09:35:46SteamcadeI'll have to check. hang on
09:36:28SteamcadeI just had to restore to factory mode... rockbox and Zeroslakr are easy to update, I hate having to redo everything whenever apple updates
09:36:53funmanjhMikeS: where does this planar format come from ? don't driver output lrlrlr and not lllrrr ?
09:37:55 Quit TheSeven (Ping timeout: 276 seconds)
09:38:16 Join TheSeven|Mobile [0] (~theseven@rockbox/developer/TheSeven)
09:38:46jhMikeSwhere there a planar format? it takes the raw PCM output from the buffer
09:38:46TheSeven|Mobilehm, rockbox seems to do a terrible job at resampling
09:39:24funmanjhMikeS: ah nevermind the comment confused me
09:39:38 Quit [Saint] (Quit: leaving)
09:39:53jhMikeSTheSeven|Mobile: define "terrible job"? yes, it's linear interpolation. something better where it can be used would be nice.
09:40:19funmanjhMikeS: did you compare mp3 quality for mono vs stereo ? does it match closely ?
09:40:35SteamcadeGot it
09:40:47 Join [Saint] [0] (~St.]@203.184.3.101)
09:40:48TheSeven|Mobileresampling 48k to 44k1 => everything > 12kHz gets cut off
09:40:55SteamcadeI was attempting to isntall 4G firmware, like an idiot
09:41:02jhMikeSfunman: I think so. I wasn't the only one working on this. I think Davide-NYC scutinized it closely while it was being implemented.
09:41:12 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:42:15*jhMikeS 's played back DVD captures made with S/PDIF and they didn't seem muffled
09:42:51TheSeven|Mobilewell, actually it should switch the codec to 48kHz and not do any resampling, but even for resampled audio the quality was crap (nano2g)
09:44:18pixelmathe resampler being crap is known
09:44:39 Join Rob2222 [0] (~Miranda@p4FDC9475.dip.t-dialin.net)
09:44:46jhMikeSI just made a radio recording at 48kHz and it's certainly not cutting off at 12kHz
09:45:29TheSeven|Mobilewell, it's not a complete cutoff, but strong attenuation judging from the fft plugin
09:46:44jhMikeSit's cutting at about 15kHz (68% of 22050), about where FM shound
09:47:35TheSeven|Mobiletry playing back a 48khz mp3 file on a device where the codec doesn't support that natively
09:47:45*TheSeven|Mobile need to reboot his phone
09:47:53 Quit TheSeven|Mobile (Read error: Connection reset by peer)
09:48:11 Join efyx [0] (~efyx@lap34-1-82-225-185-146.fbx.proxad.net)
09:48:30jhMikeS48000 is supported by mp3
09:48:38 Quit Rob2223 (Ping timeout: 264 seconds)
09:48:52*pixelma thought that on Rockbox swcodec targets always resamples to 44.1 kHz
09:49:28*binaryhermit believes pixelma is correct
09:49:29funmanpixelma: nope, perhaps it does for dsp effects though
09:50:36jhMikeSpixelma: not for recording, just for playback
09:51:41jhMikeSmp3 supports 16, 22, 24, 32, 44, 48...depending on the desired bitrate, which determines the available format
09:52:14 Join TheSeven|Mobile [0] (~theseven@rockbox/developer/TheSeven)
09:52:25funmanplayback resamples input to 44.1 ?
09:52:28pixelmajhMikeS: I believe he meant the hardware
09:52:29 Part TheSeven|Mobile
09:52:57 Join TheSeven|Mobile [0] (~theseven@rockbox/developer/TheSeven)
09:53:02TheSeven|Mobilere
09:53:26jhMikeSfunman: yes it does
09:53:43funmanwhy that?
09:54:04 Quit einhirn (Read error: Connection reset by peer)
09:54:33 Quit rado1 (Quit: Leaving.)
09:55:27jhMikeSswitching would be a big nightmare, really. it might be better to be able to just set it to something other than 44.1...and excessive rates are sort of pointless.
09:56:02pixelmaI guess it would be a real nightmare for things like crossfade
09:56:46 Quit [Saint] (Quit: leaving)
09:56:57jhMikeSand voice...it would break up speech across switches. it isn't glitch-free, so pauses in audio playback would be necessary
09:56:57funmanwhat's the point of playback drivers supporting different frequencies then?
09:57:39TheSeven|Mobileprobably some plugins needing it
09:57:51jhMikeSplugins can use it. recording uses it. it wouldn't be so bad if it weren't expected to change in the middle of playback.
09:58:07TheSeven|Mobile(someone recently complained about 24khz not working on nano2g for some reason)
09:58:37Zagorsurely we can upsample the voice if needed?
09:58:59jhMikeSeffects like crossfeed and tone filters would need to be checked out and reworked
09:59:02funmanalso voice isn't using 44.1kHz?
09:59:14jhMikeSwe do upsample the voice ... to NATIVE_FREQUENCY
09:59:28jhMikeSI think the spx is 16k
09:59:35Zagorright, I mean dynamically to whatever is currently playing
09:59:35S_a_i_n_tTheSeven|Mobile: That's regarding Doom and Rockbox...and yes, the sound is AWFUL
09:59:39 Join detaos_ [0] (~quassel@ip72-218-104-242.hr.hr.cox.net)
10:00
10:00:17Zagorcrossfeeding can either be disabled, or trigger resampling
10:00:21jhMikeSZagor: sure, but then if it switches in the middle of speaking, it'll break momentarily.
10:00:31Zagortrue
10:00:41 Quit detaos (Ping timeout: 276 seconds)
10:01:15 Join noname [0] (~jonno@203.184.3.101)
10:01:41 Join swilde [0] (~wilde@aktaia.intevation.org)
10:01:43 Nick noname is now known as Guest64767 (~jonno@203.184.3.101)
10:02:00ZagorI would speculate those who want 48khz will not mind terribly a possible glitch in speech at track transitions
10:02:04jhMikeSin the same vein, no crossfade between tracks of different rates could be smooth
10:02:13 Quit Guest64767 (Client Quit)
10:02:14Zagorno, only with resampling
10:02:23jhMikeSit'd be easier to just set the output to run at 48 all the time, really
10:02:31TheSeven|Mobilehm the switching shouldn't cause more than a slight pop/click if done the right way
10:02:36funmanjhMikeS: what do you think of merging codecs/mp3_enc.c and plugins/mp3_encoder.c ? for example including a common .c file
10:02:47jhMikeSone thing about 44.1 is that it's rather universal, but 48 is not
10:03:18Zagoryes, if we're limiting ourselves to one frequency 44.1 makes the most sense
10:03:33jhMikeSfunman: sounds like a plan. the mp3_enc.c is the more optimized one of course.
10:03:35 Join noname_ [0] (~jonno@203.184.3.101)
10:04:12JdGordonbieber: ping?
10:04:18bieberHey
10:04:24 Join Yun [0] (www-data@giant.haxx.se)
10:04:46JdGordonyou know how a few tags eat the \n in svn (so a new line isnt drawn) how do you handle tha currently?
10:04:48 Quit noname_ (Client Quit)
10:05:17bieberAt the moment, the way it's working is that I only draw newlines after blank lines, or lines that have some text rendered in them
10:05:31bieberIf the line has nothing but tag side-effects, no \n
10:06:35JdGordondoesnt that mean that "%?mh<hello>" will draw a \n if it is on and not if it is off?
10:06:59bieberHmm, I'm not sure, but probably so
10:07:00funmanjhMikeS: r24271 (by n1s) : isn't the first sample LEFT, and not RIGHT ?
10:07:55JdGordonit looks like there is about a dozen tags which need this behaviour... should I put it into the tag table somehow?
10:08:06bieberThat would be helpful
10:08:14JdGordon\n at the end of the param list causes the \n to be eaten?
10:08:17 Quit Yun (Client Quit)
10:08:23bieberThen I can just disable the newline if any of them are present on the line?
10:08:25funmanah i spot a commit by Toni less than one year ago (i was grepping for 'toni' not 'Toni')
10:08:40bieberOh, you want to eat it in the parser
10:08:44JdGordonyeah
10:08:46bieberThat would actually be easier, okay
10:08:47JdGordonthat is how svn works
10:09:43JdGordonmind you, I dont like tags that do that
10:09:53JdGordon%we %wi %wd all do it for no particular reason
10:09:55jhMikeSfunman: yeah, should be Ll|Rr, or lL|rR depending on endianness
10:11:23jhMikeSfuman: hmmm...I see...it looks backwards
10:12:20funmanbtw mp3_enc.c looks similar to mp3_encoder.c according to git log (no optimization not present in the other file)
10:13:58jhMikeSmp3_encoder.rock is afaiu taken and just turned to codec format
10:14:09funmanhum.. svn logins are case insensitive? it looks like people can use any case and the server will use just that in the commit log
10:15:28jhMikeSof course, maybe they're just based off the same reference (I'm not 100% certain)
10:18:10JdGordonbieber: I dont understand the parser enough... have you got a few min to hack this in?
10:18:23bieberSure thing
10:18:24JdGordonif \n is the last param then just eat everything up to and including the \b
10:18:26JdGordon\n
10:18:40JdGordonI'll pastebin the diff for tag_table
10:19:18JdGordonhttp://pastebin.com/AR32fmuj
10:21:59funmanjhMikeS: how about using a single file (in apps/plugins or apps/codecs) and #ifdef PLUGIN / #ifdef CODEC ? versus 2 small files in apps/plugins & codecs, and a 3rd file wiht all the common code
10:23:21jhMikeSlibmp3enc?
10:25:17jhMikeSdoes mp3_encoder get made with hwcodec too?
10:26:16funmannope (no reason why except the slowness though)
10:28:22jhMikeSI guess the main stuff could go in the codec lib sections (probably the best area for common code like that). we build libmad and libspx for different purposes so I'm sure a similar sort of solution could work.
10:28:35 Quit binaryhermit (Ping timeout: 245 seconds)
10:28:41jhMikeS*libspeex
10:28:48pixelmawouldn't that mean using software encoding on hwcodec? That doesn't sound like a logical thing to do
10:29:17*jhMikeS was just told hwcodec didn't use the mp3_encoder plugin
10:30:03jhMikeSof course if it's an encoder library useable in general, it sort of makes sense
10:30:11pixelmayes, my statement was because of funman's "no reason why it isn't included)
10:30:43jhMikeSand his other statement...."slowness" :D
10:30:54funmancan the player use the MAS for mp3 encoding ?
10:31:18pixelmaI guess if you want mp3 encoding on hwcodec you would use the hw features,
10:31:30pixelmawhich means a different plugin though
10:31:57funmanjhMikeS: i just verified that r24271 indeed swaps channels, but the log mentions "fix swapping" so i want to check with n1s first
10:31:58jhMikeSwell, it records it, but I'm not sure if it can take pcm and do it
10:32:29jhMikeSfunman: or test record from one channel
10:32:40funmanwell i did
10:32:55 Quit Zigtown (Ping timeout: 240 seconds)
10:33:43jhMikeSah, yes, that's what you meant. then what the....??
10:34:05funmanperhaps the target on which he tested swaps the channels upfront?
10:34:39jhMikeScould be...I'll do a check myself here with one channel gain all the way down
10:35:03funmanjust take e200v1 and disable channel copy in pcm-pp.c
10:35:40 Join DerPapst [0] (~Alexander@dslb-088-069-138-199.pools.arcor-ip.net)
10:36:30 Quit TheSeven|Mobile (Read error: Connection reset by peer)
10:36:40 Quit 16SAAHAE3 (Changing host)
10:36:40 Join 16SAAHAE3 [0] (Zulu@unaffiliated/zambezi)
10:36:47 Nick 16SAAHAE3 is now known as Zambezi (Zulu@unaffiliated/zambezi)
10:36:55jhMikeSI checked gigs with R turned way down, and L plays back loud like it should .... hmmm
10:37:27jhMikeSpeakmeter shows correctly too
10:39:02 Join [Saint] [0] (~St.]Lasci@203.184.3.101)
10:40:02jhMikeSe200v1 didn't swap it either
10:40:02CIA-6New commit by funman (r27209): encoders are not built in sim builds ...
10:40:51funmanso if you record mp3 with mono mode set to 'left' you have normal output ?
10:41:03funman(and channels set to mono)
10:41:21 Join einhirn [0] (~Miranda@bsod.rz.tu-clausthal.de)
10:41:33jhMikeSoops :) (think I forgot to set it :P)
10:42:11CIA-6r27209 build result: All green
10:42:16funmannever underestimate the pebkac :P
10:42:36jhMikeSnope, output is silent (R turned down, L mono mode)
10:42:49 Quit [Saint] (Client Quit)
10:43:10 Join [Saint] [0] (~St.]Lasci@203.184.3.101)
10:43:37jhMikeSswitch it to R mono mode with R turned down, I get mono L instead, yep
10:43:44***Saving seen data "./dancer.seen"
10:44:18funmann1s: stop enjoying the sun and get back here!
10:45:51CIA-6New commit by bieber (r27210): Theme Editor: Implemented JdGordon's tag newline-eating mechanism
10:46:53 Quit [Saint] (Client Quit)
10:47:16 Join [Saint] [0] (~St.]Lasci@203.184.3.101)
10:47:31CIA-6r27210 build result: All green
10:47:46pixelmanot very descriptive
10:48:19[Saint]/highlight nick -n
10:48:31S_a_i_n_toops.
10:50:52bieberJdGordon: It should be working now
10:54:59balintxhi
10:55:17funmanhm the plugin and the encoder differ too much :'(
10:55:18balintxhave an idea improving http://www.rockbox.org/wiki/LinuxSimpleGuideToCompiling
10:55:33balintx" Getting Started"
10:55:45balintxIf not, you will need to type cd ~, then type nano ~/.bashrc and press Page Down a few times. You should see a blinking cursor on an empty line. Paste this onto the line:
10:55:55balintxshould be enough to paste without nano
10:56:19balintxecho "export PATH= /usr/local/bin :$PATH " >> ~/.bashrc
10:56:31funmanbalintx: good idea, you have a wiki account already ?
10:56:40balintxno, havent
10:57:01funmanjust make one (using your real name as login) and we'll give you write permission
10:58:07 Quit [Saint] (Quit: leaving)
10:59:13 Join lmjubuntu [0] (~lmjubuntu@58.195.42.183)
11:00
11:02:14balintxfunman: where to register?
11:02:38balintxnow i see
11:06:07balintxlol
11:06:11balintxYour personal Foswiki topic BalintKoczka has been created
11:06:15balintxFoswiki :-O
11:07:29funmanyou should be able to write now (perhaps it requires logout/login if it doesn't work at first try)
11:07:58JdGordonbieber: thanks
11:08:46balintxfoswiki :D
11:08:46balintxsorry..
11:08:46balintxin my language's slang 'fos' means 'sh*t'
11:09:00jhMikeShaha
11:09:11balintxbut it's true :-O :D
11:10:19 Join [Saint] [0] (~St.]Lasci@203.184.3.101)
11:13:14JdGordonbieber: looks like you didnt get it quite like svn, but I tihnk it is good enough
11:13:38JdGordonin svn you can put garbage at the end of those lines and it will all get eateb
11:13:40JdGordoneaten
11:13:59JdGordoni.e the same as putting # on the end of the tag
11:14:06JdGordonunless I misread the diff?
11:14:07bieberOhh, okay
11:14:16bieberI'll have it seek for a newline tomorrow
11:14:23bieberright now it just looks for one immediately after the tag
11:14:39JdGordonI should be able to fix that
11:15:41balintxdone
11:18:10bieberokay, cool
11:26:49amiconnZagor: 32/64 MB is independent of 5g/5.5G
11:27:58Zagoramiconn: aha. what is 5.5 then?
11:28:34amiconn5.5G is the one that presents 2048 byte sectors over USB. It also has a more power efficient/ brighter backlight
11:29:04pixelmaoh, I withdrew a comment about 64mb before because I assumed Zagor knew...
11:29:19 Quit [Saint] (Quit: leaving)
11:29:38amiconn60GB is 5G and has 64MB RAM. 80GB is 5.5G and also has 64MB RAM. 30GB exists both as 5G and 5.5G, both having 32MB RAM
11:29:39 Join [Saint] [0] (~St.]Lasci@203.184.3.101)
11:29:49Zagorknew? :) 09:34 <Zagor> Steamcade: I'm not sure what the definition of 5.5G is.
11:29:51balintxfunman
11:30:02amiconnThat said, there are refurb devices around with 60 or 80GB, but a 32MB mainboard
11:31:03pixelmaZagor: yes, but that doesn't say anything about your knowledge 32/64MB RAM
11:31:06balintxfunman: wiki smashes up code.
11:31:31balintxecho "export PATH=/usr/local/bin:$PATH" >> ~/.bashrc
11:31:31balintxsave
11:31:31balintxreload the page
11:31:41balintxand i see: echo "export PATH= /usr/local/bin :$PATH" >> ~/.bashrc
11:32:44S_a_i_n_tbalintx: Are you talking about the $PATH in CygWin?
11:33:13funmanbalintx: i don't know foswiki syntax but you can use the preview button before saving
11:33:22S_a_i_n_tOh, right...on the wiki, nevermind
11:33:28 Join mitk [0] (~mitk@195.117.162.130)
11:34:18balintxfunman:
11:34:52 Quit [Saint] (Quit: leaving)
11:35:25balintxthere isn't any preview
11:35:31 Join [Saint] [0] (~St.]Lasci@203.184.3.101)
11:36:26Zagorbalintx: which page?
11:37:43balintx[10:55:17] <balintx> have an idea improving http://www.rockbox.org/wiki/LinuxSimpleGuideToCompiling
11:37:55 Quit [Saint] (Client Quit)
11:38:47Zagorlooks like the bolding interfered somehow
11:38:51balintxmay now
11:38:59balintxyes
11:39:22balintxi just switched of WhatYouSeeIsWhatYouGet mode
11:39:22balintx+f
11:39:37Zagoryes. "raw view" and "raw edit" can be useful when diagnosing such issues.
11:44:09 Join [Saint] [0] (~St.]Lasci@203.184.3.101)
11:44:12amiconnfunman: Nope, the Player (and also the Ondio SP) have a MAS that doesn't include an MP3 encoder
11:44:30 Quit Steamcade (Quit: -=SysReset =-)
11:44:40amiconnBut even the recorders and the Ondio FM which do record with an MP3 capable MAS cannot use that for encoding from file
11:44:57amiconnThe reason is that there is no way to feed the data
11:45:03funmani'm not sure anyone would be crazy enough to convert wav files to mp3 on the SH cpu
11:45:20 Quit [Saint] (Client Quit)
11:45:35 Join [Saint] [0] (~St.]Lasci@203.184.3.101)
11:45:59amiconnAlso, there is no real reason to encode on device, because the encoder we have isn't as good quality as directly recording to mp3
11:46:59 Part lmjubuntu
11:47:15amiconnjhMikeS: Btw, MP3 also supports 8/11/12 kHz (MPEG2 lsf extensions, aka MPEG2.5)
11:47:38gevaertsamiconn: couldn't you in theory play wav to S/PDIF out and look that back to S/PDIF in?
11:47:53amiconnnope
11:48:00amiconnThe MAS isn't full duplex
11:48:13*gevaerts gets *two* recorders :)
11:48:16amiconnOnly one internal app can be run at a time
11:48:19jhMikeSamiconn: I believe that's correct but that encoder doesn't do 2.5
11:48:53 Quit [Saint] (Client Quit)
11:49:11 Join [Saint] [0] (~St.]Lasci@203.184.3.101)
11:50:04amiconnAha. The MAS does
11:50:23 Quit [Saint] (Read error: Connection reset by peer)
11:52:45amiconnhmm
11:52:58*amiconn wonders whether what he said is actually true
11:53:11amiconnThere may be a way to encode from file using the MAS
11:54:28jhMikeShow does the data get sent to it anyway?
11:54:47amiconnAndwhile the MAS does decode MPEG2.5, it does support it for encoding
11:55:00 Join [Saint] [0] (~St.]Lasci@203.184.3.101)
11:55:32 Join n1s [0] (~n1s@rockbox/developer/n1s)
11:56:52 Quit [Saint] (Client Quit)
11:57:04amiconnPlayback data goes through serial input B (I2S-alike)
11:57:42amiconnRecording goes through PIO output
11:57:50 Quit funman (Quit: leaving)
11:59:11amiconnSeems we cannot use it for encoding from file, because the encoder can only either encode from serial input A (which is hardwired to the internal ADC) or S/PDIF
11:59:51 Quit kachna (Ping timeout: 260 seconds)
12:00
12:02:06 Quit slck (Ping timeout: 248 seconds)
12:04:25 Join Buschel [0] (~~andree@p54A3E68B.dip.t-dialin.net)
12:05:29 Join slck [0] (Venci@Slackware.SlackPix.Com)
12:08:36 Nick YPSY is now known as Ypsy (~ypsy@geekpadawan.de)
12:11:12*jhMikeS is a little puzzled why left and right gain aren't both available in mono recording mode (you might still want to adjust left and right input gain independently).
12:14:41 Join [Saint] [0] (~St.]Lasci@203.184.3.101)
12:15:07 Quit [Saint] (Client Quit)
12:17:44CIA-6New commit by jdgordon (r27211): minor tweak to r27210 to make it the same as the svn parser... eat everything up to and including the \n after these tags
12:17:47 Join watto [0] (~watto@193.203.81.165)
12:19:24CIA-6r27211 build result: All green
12:34:31 Quit efyx (Remote host closed the connection)
12:43:46***Saving seen data "./dancer.seen"
12:52:02 Join bor_ka [0] (www-data@giant.haxx.se)
12:57:32 Join bubsy [0] (~bubsy@94-139-72-137.bb.cust.telefiber.no)
13:00
13:22:23 Join teru [0] (~teru@M016207.ppp.dion.ne.jp)
13:29:17 Join ucchan [0] (~ucchan@softbank126102048036.bbtec.net)
13:31:31CIA-6New commit by uchida (r27212): text viewer: ...
13:33:17CIA-6r27212 build result: All green
13:40:03 Quit yosafbridge (Quit: Coyote finally caught me)
13:40:12 Join yosafbridge [0] (~yosafbrid@li14-39.members.linode.com)
13:40:39balintxhelp
13:40:51balintxmake: *** No rule to make target `/home/balintx/rockbox/build/apps/plugins/bitmaps/native/amaze_tiles_9.9x9x16.o', needed by `/home/balintx/rockbox/build/pluginbitmaps/amaze_tiles_7.h'. Stop.
13:41:14balintxwas it worked for anyone?
13:41:21S_a_i_n_t"it"?
13:41:29S_a_i_n_tyou need to define what "it" is...
13:42:27n1ssvn has no "amaze tiles" so whatever patch you have applied is probably broken,
13:44:05 Quit sinthetek (Ping timeout: 264 seconds)
13:45:19 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
13:55:38S_a_i_n_tbalintx: What patch were you trying to install/
13:55:56S_a_i_n_ts///?/
13:56:19balintxsaratoga: this problem is solved
13:56:22balintxjust missing bmp-s..
13:57:52mitkHello commiters! Is there any chance to commit FS #11436 - Polish charset for 08-Nedore font?
14:00
14:05:36balintxwhere is the zip made by make zip ?
14:05:47balintxfound
14:05:47balintx: D
14:05:52S_a_i_n_tin your build directory
14:06:04S_a_i_n_tthe same dir you can "./configure" in.
14:08:49CIA-6New commit by teru (r27213): text editor: use plugin_get_buffer() to allocate buffer. if size of the file is bigger than the buffer, use the audio buffer.
14:10:09 Quit bor_ka (Quit: CGI:IRC (EOF))
14:10:23CIA-6r27213 build result: All green
14:13:05 Quit GodEater (Read error: Operation timed out)
14:22:16 Quit sinthetek (Ping timeout: 240 seconds)
14:26:08CIA-6New commit by uchida (r27214): text viewer: the manual updates.
14:27:42CIA-6r27214 build result: All green
14:30:20CIA-6New commit by rasher (r27215): Clarify that we use album artist if available.
14:30:38CIA-6New commit by ranma (r27216): Added some missing german translations
14:31:59CIA-6r27215 build result: All green
14:32:32pixelmaranma: the "voice" part should also be translated (I'd translate the "open" at least)
14:33:33pixelmaprobably just an oversight in the PictureFlow string :)
14:33:48CIA-6r27216 build result: All green
14:34:04 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
14:34:13 Join ruckus [0] (~Earworm@164.76.150.223)
14:35:12 Quit ucchan (Quit: Leaving...)
14:35:40 Quit Kitar|st (Read error: Connection reset by peer)
14:36:04 Join hebz0rl [0] (~hebz0rl@dslb-094-217-138-100.pools.arcor-ip.net)
14:36:27ranmapixelma: Oh, right, I managed to overlook that one :)
14:36:31 Join Kitar|st [0] (Kitar_st@BSN-142-63-17.dial-up.dsl.siol.net)
14:36:44ruckushey peoples. yesterday i updated to the most current build, and the metronome plug in has started acting weirdly. the screen used to go dark when i put it into "hold", but now it's shining brightly no matter what i do. is this a bug or a feature now? because if it's a feature, it's killing my battery really fast.
14:37:26S_a_i_n_tPerhaps there is a specific setting for the metronome backlight now...
14:37:29S_a_i_n_tjust a guess.
14:37:47 Join GodEater [0] (~bibble@cl-711.lon-02.gb.sixxs.net)
14:37:47 Quit GodEater (Changing host)
14:37:47 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
14:37:53S_a_i_n_tSome plugins have their own settings to override the backlight timeout
14:38:55CIA-6New commit by ranma (r27217): open Pictureflow -> öffne Pictureflow
14:39:06ruckusi know, but aside of the main screen, the metronome plug in doesn't have any other settings
14:39:17pixelmaI don't think metronome has but maybe you changed your "backlight on hold" setting accidentally. Stating which target you have might help too
14:39:39ruckuspixelma: target? you mean build ?
14:40:10pixelmawhich device (or player or however you want to call it)
14:40:49pixelmaand the exact revision number of your build is a nice info too
14:40:54 Quit S_a_i_n_t ()
14:41:02ruckusbut why does the target matter? it used to work before. and i just checked the setting for the "backlight on hold". it says "off"
14:41:29ruckus(ipod 30gig)
14:42:15pixelmaif it is a bug, there is a slight chance that it hits only a few targets and then all info is important
14:42:28 Quit bzed (Quit: meh.)
14:42:39ruckushopefully. it's the 5th gen ipod. not 5.5
14:43:49***Saving seen data "./dancer.seen"
14:45:20ruckus(r27188-100630)
14:47:23 Join bzed [0] (~bzed@devel.recluse.de)
14:49:42pixelmacan you remember roughly how old your build was before updating?
14:49:56ruckusless than a month
14:52:04ruckusi mean, there are no other problems with the thing. it's just the metronome acting weird with the screen light
14:52:15CIA-6r27217 build result: All green
14:57:30pixelmahmm... I can't imagine what should have changed in the meantime
14:59:18ruckussame here. :(
15:00
15:00:03ruckusi mean, there were some other changes done on the plug in, so someone definitely worked on it. the buttons were rearranged, but that's not a problem
15:02:10 Quit krazykit (Quit: Connection reset by beer)
15:02:50 Join kachna [0] (~kachna@r4ax178.net.upc.cz)
15:04:47 Part bubsy
15:05:48 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:09:58 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
15:11:08terupixelma, ruckus: changing TIMEOUT_NOBLOCK to 1 in button = pluginlib_getaction(TIMEOUT_NOBLOCK,plugin_contexts,PLA_ARRAY_COUNT); near line 916 in metronome.c seems to solve the issue for me with gigabeat x.
15:12:04ruckusi wish i had the petience to write code. :(
15:12:09ruckus*patience
15:18:43pixelmateru: something that slipped in during the big pluginlib action rework last month then?
15:19:29teruit seems like that.
15:19:54 Part LinusN
15:30:59 Quit Strife89 (Ping timeout: 240 seconds)
15:35:19CIA-6New commit by teru (r27218): metronome: fix issue that backlight doesn't turn off.
15:36:46 Quit Buschel (Ping timeout: 260 seconds)
15:37:11CIA-6r27218 build result: All green
15:38:22 Join Strife89 [0] (~Strife89@adsl-67-57-51.mcn.bellsouth.net)
15:42:24 Join panni_ [0] (hannes@ip-95-222-52-93.unitymediagroup.de)
15:43:25 Join mt [0] (~mtee@rockbox/developer/mt)
15:43:33 Join jxself [0] (~jxself@fsf/member/jxself)
15:43:59mtsaratoga: Here's what I've found out so far :
15:46:37 Quit saratoga (Ping timeout: 252 seconds)
15:48:07 Join jgarvey [0] (~jgarvey@cpe-065-190-066-089.nc.res.rr.com)
15:49:52mt1. (before imdct_half) I have scaled the output coefficients by 1<<16 to avoid overflows, and no scaling is done on the temp (input) buffer for the same reason.
15:49:53mt2. XNPROD31 in this case would produce a wrong result since the function fixmul31 it's using would scale the result down after multiplication by 31 which is wrong.
15:49:54mt3. a wma pro can have different subframe sizes, and every subframe size has a different mdct size with a different trig table.
15:50:00 Join fdinel [0] (~Miranda@modemcable235.127-131-66.mc.videotron.ca)
15:50:07mt*a wma pro file
15:50:29 Quit mitk (Quit: Leaving)
15:58:47 Part jxself ("Support freedom: fsf.org")
16:00
16:01:53 Join styroteqe [0] (1000@ool-457579de.dyn.optonline.net)
16:01:57 Part styroteqe
16:05:38 Join funman [0] (~fun@rockbox/developer/funman)
16:06:10funmann1s: have you seen http://www.rockbox.org/irc/log-20100701#10:07:00 ?
16:07:34n1sno, looking now
16:14:55 Quit teru (Quit: Quit)
16:15:59n1sfunman: about the channels, i've no idea, i vaguely remember someone reported that the encoded file had the channels swapped so i swapped them...
16:16:19funmanlet's swap them again and wait for the next report
16:16:38funmanok with you?
16:16:44n1sanyway, about encoders, i think it would be nice with a plugin that could encode with any of our encoding codecs instead of separate encoder plugins
16:17:04funmantrue
16:17:16n1syes, go ahead, would be nice if anyone could test but i don't think i have a stereo mic
16:17:27funmani and jhMikeS tested already
16:17:40n1sok
16:17:45n1sgtg now
16:19:21funmanbtw the weird to_mono() isn't inlined
16:21:30CIA-6New commit by funman (r27219): mp3 encoder (codec): left and right channels were inverted in mono
16:23:13CIA-6r27219 build result: All green
16:24:12 Quit kachna (Ping timeout: 276 seconds)
16:27:23 Quit bluebrother (Disconnected by services)
16:27:25 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
16:38:01 Quit hebz0rl (Quit: Ex-Chat)
16:39:35 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
16:43:53***Saving seen data "./dancer.seen"
16:43:54 Join AzureWurk [0] (opera@c-24-21-44-114.hsd1.wa.comcast.net)
16:44:05 Quit t0rc (Remote host closed the connection)
16:44:11 Quit ender (Quit: The sum of intelligence on the planet is a constant; the population is growing.)
16:50:41 Join ender` [0] (krneki@foo.eternallybored.org)
16:52:34 Join saratoga [0] (9803c6dd@gateway/web/freenode/ip.152.3.198.221)
16:54:36saratogamt: whats wrong with XPROD?
16:57:06 Join MethoS- [0] (~clemens@134.102.106.250)
16:57:14mtsaratoga: it's using MULT31 to do the fixed point multiplies .. which returns the result shifted down by 31 bits. While in wma pro there's a case where one operand is not scaled at all and the other is scaled by (1<<31), so if we want the output to be in Q.16 format the result should be shifted down by 15 not 31.
16:57:44mtsaratoga: thus there's an assumption there about the fixed point representation of the operands
16:58:15saratogamt: i think that should be ok, since the second input is a trig function thats always < abs(1), so its been scaled by 31 bits
16:58:29saratogaso the 31 bit shift cancels out the shift built into the trig constants
16:59:16saratogaso if you do ((x<<16) * (y<<31))>>31 you get (x<<16) back
17:00
17:00:29CIA-6New commit by alex (r27220): Correct English in text viewer manual entry.
17:00:46mtin wma pro it's not correct though, since the trig table contain very small values, while the left operand contains very large numbers which we can't scale up to avoid overflows ... i.e, the returned result is x<<0 not x<<16
17:01:24mtby the trig table I mean wma pro's, not the one we currently have
17:01:38 Join hebz0rl [0] (~hebz0rl@dslb-094-217-138-100.pools.arcor-ip.net)
17:02:08CIA-6r27220 build result: All green
17:03:53saratogasorry, which table are you referring to?
17:04:57mtsaratoga: I'm not sure if I explained the matter clearly, so here's a second take : x is very large, so it's in Q.0 (just an int cast of the floating point value) , the other operand (trig value) is Q.31 .. so after the multplication, the result would need to be shifted down 15 bits to have a 16.16 result.
17:05:33mtsaratoga: I converted ffmpeg's imdct_half and tables to fixed point for testing
17:06:19saratogai'm confused, if it starts in Q31.0 why would it end up in Q15.16?
17:07:45mtto avoid overflows
17:08:09mtand because I want the output in Q.16
17:08:18saratogawouldn't that need to be a separate step though?
17:08:28saratogaso you'd scale it to q15.16, then multiply it?
17:08:50saratogathough i guess you could write a function to do it in one step
17:09:08mtdo you mean scaling x to q15.16 before passing to XNPROD ?
17:09:38saratogai'm thinking it should be scaled into whatever format is needed to prevent overflow first, and then passed into the XNPROD code
17:11:11mthmm .. maybe I could just scale it down after XNPROD
17:11:36saratogabut just thinking about this, what the XNPROD functions do is a 64 bit multiply of 2 32 bit numbers, and then return the top 32 bits, so I don't think they should overflow unless both the inputs were > 0x8FFFFFF
17:11:37 Quit swilde (Quit: ERC Version 5.3 (IRC client for Emacs))
17:12:21saratogasince i think the only way to have overflow would be from the sum after the multiply (e.g. both products are really close to MAX_INT and they get summed)
17:12:46 Quit ruckus (Quit: Leaving.)
17:13:27saratogai guess theres really two things going on here, the need to have an extra bit or two of headroom so that the dot products can't overflow, and also the need to have a couple bits headroom so that rounding error in the FFT can't occasionally overflow
17:13:31 Join Stephenc___ [0] (www-data@giant.haxx.se)
17:14:00Stephenc___pixelma ping
17:14:16saratogai think as long as the input frequency domain samples to the fixed point mdct are less then, say 0x2FFFFFFF, they'll probably be fine
17:15:53saratogathe actual representation shouldn't matter, since a given binary number has the same chance of overflow no matter where we choose to put the decimal point, and the fixed format is just a mathematical convenience for us, not something with real meaning to a computer
17:16:14saratogadoes that sound right?
17:19:56mtI'm not sure. " a given binary number has the same chance of overflow no matter where we choose to put the decimal point", since moving the point implies scaling, the chance to overflow should be related to the place of the imaginary point.
17:20:18Stephenc___pixelma, doing the m3 themes (iCatcher and UniCatcher) the font isn't specified in the cfg for iCatcher, 08-Schumacher-Clean.fnt is included in the build but the rwps for the x5 used 09-Nedore.fnt, which would be better to use ?
17:21:22saratogamt: sorry that was unclear, what I mean is from the mdct's perspective, the decimal point on the input is always after the last bit (q31.0 - a plain integer), and we just say that its somewhere else so that the math works out in the rest of the codec
17:21:47 Quit Strife89 (Ping timeout: 265 seconds)
17:22:43saratogaso if you think of the xnprod functions as multiplying a whole number times a fixed point number, the scaling makes sense: INT * q.31 would be scaled as (INT * q.31)>>31
17:23:50mtSounds right
17:24:44saratogathe way i think about it, the actual fixed format is just the number you have to multiply by to get the floating point value back, but if you never multiply back to floating point, then the format isn't so important
17:25:19saratogait just matters that we tell the rockbox DSP code the right number of bits to scale so that the volume is correct, and that we don't ever overflow, or scale so badly that numbers round down to zero
17:26:48 Quit Stephenc___ (Quit: CGI:IRC)
17:30:16mtsaratoga: this sounds correct .. but the results based on this hypothesis say there's something wrong, and now I'm confused and can't explain the output. :( .. one minute I'll try to figure it out.
17:31:01saratogayeah something is definitely missing :(
17:31:21 Join Strife89 [0] (~Strife89@adsl-67-57-51.mcn.bellsouth.net)
17:31:34saratogabut i think its a problem with the algorithm, not the precision, since no amount of scaling seemed to fix it for me
17:32:32pixelmaAlexP: did you have time to look at the Onda manual patch? (and see my question about soft hold today in the morning?
17:33:08AlexPNot yet I'm afraid - I ended up going out last night, and am out tonight too. I will though, I promise!
17:33:56AlexPAnd for hold switch, how about just calling it hold?
17:34:07AlexPAnd not specifiy that it is a switch
17:34:39funmanis is to remove distinction between HWhold and SWhold ?
17:34:51pixelmahave to look at the exact phrasing, maybe that works
17:35:07AlexPfunman: IIUC :)
17:35:19pixelmafunman: just for the manual (description of the hold status tag)
17:35:46saratogamt: just thinking this through, have you tried comparing the rockbox IMDCT to the ffmpeg IMDCT at various points ?
17:36:01funmankeep in mind there are notable differences (depends which part of the manual you want to change)
17:36:06saratogafor example, it would be helpful to know if things go wrong before or after the FFT
17:36:58mtsaratoga: They go totally wrong before the fft .. the problem is that the prerotation uses very large trig values in comparison to what wmapro actually uses (in ffmpeg).
17:37:06pixelmafunman: or better: to include that part about the %mh tag for soft hold targets at all without duplicating code unnecessarily
17:38:57mtsaratoga: Another take at explaining the results .. if the trig operand is actually very small that the result of fixed point multiplication would always fit in 32 (or max 33) bytes) .. returning just high byte would result in an underflow, which is the case here.
17:40:00saratogahow could that happen though? the trig values are scaled to take up the whole 32 bits, and then the multiplication discards the lower 4 bytes
17:40:21saratogaeven the smallest trig value would still have something like 20 odds bits non-zero in it
17:42:07saratoga"20 odds bits" == "20 or more bits that were non-zero"
17:43:02 Part Zagor
17:43:23mtsaratoga: the trig values used here are as small as 0x5A82 (fixed point values)
17:44:33saratogathats still 15 bits though, does that really cause underflow?
17:45:19 Join Takato [0] (~chatzilla@83.5fec54.tvnetwork.hu)
17:45:35Takatohi there
17:45:48saratogawhere did you find that value anyway? the smallest I see in the lib folder is 0x1921fb
17:45:51mtmultiplying this by a coefficient which is say 371145 (which is a real example from a sample file), would result in a number that needs 33 bits to be represented. The high part returned by MULT31 would just contain the highest bit and discard the rest.
17:46:36mtsaratoga: I converted the tables needed by wma pro to fixed point .. yes these small values are not present in our table
17:46:42saratogaah!
17:46:50saratogacan you pastebin that table?
17:46:52Takatoi had a dream about rockbox
17:47:03saratoga0x371145 or just 371145 ?
17:47:15mtdecimal not hex
17:47:26mtso just 371145 (sorry wasn't clear there)
17:48:37saratogamt: in 16 bit precision, 371145 is just 5 (e.g. -66dB below peak), so I think its reasonable that it might underflow
17:49:56 Quit pamaury (Remote host closed the connection)
17:50:23mtsaratoga: here's the table : http://www.pastebin.org/370629, it's actually quite large since it's actually made up of 6 tables for the 6 wma pro block sizes, and the entry point in the table is calculated based on the block size.
17:51:37saratogamt: its cutoff since its so huge, could you just give me the 1024 or 2048 block table? the others are just resampled versions of it
17:51:45mtsaratoga: in the floating point decoder, 371145.25 is the value contained in the buffer before passing to the imdct_half function.
17:51:54funmanmt: pastebin cuts large pastes
17:52:07mtheh should have paid attention there !
17:52:24pixelmaAlexP: e.g. "if the main unit hold is on" (that would be the old minus the "switch")...or more like "if the main unit's keys are locked"?
17:53:05pixelmaor something else... ;)
17:53:21saratogamt: basically what I'm getting at here is that the range of input samples that matter isn't so big (since we only need to output 16 bit samples) and the range of values in the trig table also isn't very big, so there should be a combination of precisions here that makes underflow only happen to samples that would be 0 in 16 bit precision anyway
17:53:35AlexPpixelma: Yep, or if you want a direct replacement you could use something like "key lock" maybe
17:53:35mtsaratoga: this is the 2048 block size table : http://www.pastebin.org/370642
17:53:42*pixelma should just commit something in broken English and wait for AlexP to put it nicely
17:53:50saratogamt: ah I see the issue
17:54:00AlexPpixelma: heh!
17:54:10saratogathe table isn't normalized to q.31 precision
17:54:37saratogaits in q15.16 precision
17:54:54 Quit Takato (Quit: ChatZilla 0.9.86 [Firefox 3.6.6/20100625231939])
17:54:54saratogaso the shifts in XNPROD would be wrong for that table
17:54:57 Quit MethoS- (Read error: Connection reset by peer)
17:55:02AlexPpixelma: But yeah, "if the main unit's keys are locked" is fine, or "if the main unit key lock is enabled" or "if hold on the main unit is enabled"
17:55:13AlexPon/enabled/whatever :)
17:55:45saratogafunman: FWIW I stared at the stupid AMS datasheet for like an hour yesterday trying to figure out what we're doing wrong the m200v4, but I still don't really get it
17:56:08mtsaratoga: Actually it is in q.31
17:56:30mtexample : 5a82 is actually 0.000011 in ffmpeg
17:56:40funmansaratoga: i compared the init code in disassembly between clipv1 and m200v4: perhaps the common ground misset is what makes the audio quite low
17:56:55saratogamt: oh is it still cutoff?
17:57:57mtno the pastebin contains the complete table
17:58:06mtI just meant that the second value in the table
17:58:11mtwhich is 5a82
17:58:20mtis from 0.000011 in floating point
17:58:32saratogamt: the largest value should be 1.0 right?
17:58:34funmanIIUC the ground is currenty common for left/right speakers, would that mean twice more current goes into the same wire than if the grounds were separate?
17:59:00saratogawhich in q.31 would be 0xFFFFFFFF right?
17:59:04mtsaratoga: in wma pro no, I haven't seen such large values yet :)
17:59:11saratogafor the trig table?
17:59:18saratogait must have cos(0) as a value
17:59:39saratogafunman: i think the issue is regulator being used, not where the power goes
18:00
18:00:04saratogasince to use a AA battery there must be an upconverter that takes ~1-1.5v and generates 1.5-2v out
18:00:27saratogatheres a half dozen in the player, i think we just use the wrong one
18:02:38 Quit scorche (Disconnected by services)
18:02:47 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
18:04:49mtsaratoga: I'm pretty sure the table values are correct, I have compared them to ffmpeg's .. and the code to convert the tables to fixed point is pretty straight forward : http://www.pastebin.org/370685 .. this is ffmpeg's mdct_init function, with just some printf's to dump the fixed point tables.
18:05:11 Quit Strife89 (Ping timeout: 260 seconds)
18:05:11 Quit tmzt (Ping timeout: 260 seconds)
18:05:43 Join tmzt [0] (~ircuser@99-157-224-139.lightspeed.bcvloh.sbcglobal.net)
18:07:20 Join rado1 [0] (~jon@66-214-153-109.dhcp.mtpk.ca.charter.com)
18:08:07 Join Strife89 [0] (~Strife89@adsl-67-57-51.mcn.bellsouth.net)
18:16:28 Join pamaury [0] (~quassel@p5DDEE131.dip.t-dialin.net)
18:16:29 Quit pamaury (Changing host)
18:16:29 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
18:40:30 Quit bmbl (Read error: Connection reset by peer)
18:43:20 Join pamaury_ [0] (~quassel@p5DDECFE7.dip.t-dialin.net)
18:43:54***Saving seen data "./dancer.seen"
18:43:59 Quit pamaury (Ping timeout: 260 seconds)
18:54:31 Join Jerom [0] (~jerome@95.171.137.241)
18:55:23 Nick pamaury_ is now known as pamaury (~quassel@p5DDECFE7.dip.t-dialin.net)
18:55:32 Quit pamaury (Changing host)
18:55:32 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
18:57:00 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:58:39saratogamt: in that pastebin, what is the value of scale?
18:59:00bertriksaratoga, do you have an m200v4?
18:59:07saratogabertrik: no
18:59:18saratogajust looked at the datasheet
18:59:19funmanbertrik: check the forums, there is a tester
18:59:34bertrikI looked up some m200v4 init code and I think domonoky tried some of the init values, but no luck as far as I remember
19:00
19:00:41bertrikI guess there just aren't that many people who have an m200v4.
19:01:03saratogai assume theres not much interest in swapping batteries anymore
19:01:17pixelmahuh?
19:02:14mtsaratoga : http://www.pastebin.org/370899
19:03:10saratogamt: why is that in there? its screwing up the values
19:03:24mtsaratoga : That's ffmpeg's code
19:03:35mtscale is provided by the decoder too
19:04:04saratogayou'll need to correct for it
19:04:19saratoga1/0.000043 is about 14.5 bits
19:04:36saratogaso you need to shift it back by 14-15 bits
19:04:46saratogabut i wonder what thats supposed to do
19:06:25saratoga 1.0 / (1 << (BLOCK_MIN_BITS + i - 1)) / (1 << (s->bits_per_sample - 1))
19:06:27saratogawhat the hell
19:06:32mt:(
19:06:53saratogaso what that does i think is make the mdct conserve energy
19:07:35mtI'm trying to make the fixed point imdct_half as close as possible to ffmpeg's first so that it would be easier to spot the problems, then I'll try to move back to our code.
19:08:04saratogathe issue here is that prescaling everything is not going to work in fixed point
19:08:18saratogabut you can make this work pretty easily i think
19:08:30saratogaleave the scale in, and then shift everything by 15 bits
19:08:43saratogathen change your fixed to float function to shift things back by 15 bits
19:08:45saratogathen it should work
19:08:54saratogabasically heres what I think is going on
19:09:13saratogain a normal mdct the codec ignores parseval's theorem for efficiency sake
19:09:35saratogaso the output energy is scaled by the number of samples in the transform
19:09:57saratogaffmpeg apparently lets you rescale the output so that energy is conserved, but it looks like it does it by scaling the trig coefficients
19:10:19saratogabut we can't do that since sclaing them basically changes the fixed representation, which makes everything underflow
19:10:47saratogabut if you scale the trig vales, then shift them (== change the fixed format) it'll avoid the underflow while keeping the scaling
19:11:36saratogayou just have to remember when you compare to FP at the end to shift everything by 15 bits again
19:12:57 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
19:13:33 Quit liar (Ping timeout: 252 seconds)
19:16:36mtsaratoga: The fixed point trig tables contain the scaled values of the floating point trig values, or did you mean that I should create fixed point trig tables of the unscaled trig values ?
19:17:06saratogamt: you can scale them, but if you do, you must then shift them so that 1.00000 is very close to 0x7FFFFFFF
19:17:18saratogaor 0xFFFFFFFF depending on twos complement i guess
19:18:09saratogabasically you need to arrange the trig values so that the largest is represented by close to 2^31, and the smallest by -2^-31 and 0 by something in the middle
19:19:49mthmm, I'll need to redo not too few stuff this way .. I'll continue today using what I have now and then redo the tables and the calculations tomorrow.
19:21:05mtCurrently there's only something wrong with the post rotation anyway so I'll try to fix that and make sure there's nothing wrong somewhere else.
19:21:21saratogamt: did you ever have to look at how floating point calculations are implemented in hardware in school?
19:21:32mtyeah
19:21:54saratogasame idea here right? multiply by some random value, keep track of the scale, rescale as needed :)
19:23:05 Quit DerPapst (Quit: Leaving.)
19:24:34 Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
19:25:23 Quit hebz0rl (Quit: Ex-Chat)
19:25:32 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
19:26:51 Quit flydutch (Quit: /* empty */)
19:28:45mtsaratoga: yeah I see .. I misread the first time
19:29:06saratogaok
19:29:14saratogaif you get stuck send me a patch and I can try and look at it
19:29:45mtalright .. thanks :)
19:35:17 Quit wombat23 (Ping timeout: 245 seconds)
19:35:34mtsaratoga: yeah scaling up by 16 seems to do it for n = 8K (scale = 0.000043)
19:40:34 Join janwil [0] (~janwil@137.224.225.4)
19:45:42janwilhello, can someone help me fixing my sandisk c240 v1 ?
19:46:31funmannobody can if you don't tell the problem
19:47:02janwilfor some reason sansapatcher doesn't recognise the device
19:47:27funmanperhaps it's a v2, no matter what the case says. check the firmware version in the OF
19:47:29janwilsame for e200tool
19:49:11funmansaratoga: codec lib mdct is based on tremor, right?
19:49:13janwilhow can i check this exactly, lsub just says c40
19:49:45bertrikon the player itself, probably somewhere around settings / system settings
19:50:00funmanhttp://download.rockbox.org/daily/manual/rockbox-sansac200/rockbox-buildch2.html#x4-70002.1
19:51:12mtfunman: The old one is based on tremor, the new one is based on ffmpeg (re: mdct)
19:51:57funmancan you think of a better name than 'union magic' for the 64 bits numbers?
19:52:56janwilfunman: hmmm, can't find it i think. rbutil says: VID: 0781 PID: 7451, Sansa c240 / [System] USB VID: 0781, PID: 7451 / Sandisk firmware i put on it is 01.01.00P / Sericlas Sanca c200B0612020984
19:53:25janwils/Sericlas/Serials/
19:53:54funmanjanwil: read the link i gave you
19:54:13 Join wombat23 [0] (~beuteltie@adsl-99-39-2-249.dsl.pltn13.sbcglobal.net)
19:54:49mtfunman: I thought the name magic was not good too .. maybe : INT64 or BUF64 ?
19:55:36funmanINT64 looks good, i think it's not a reserved name
19:56:16bertrikjanwil, on my c240v1 the firmware version can be found at menu Settings / System Info / Version
19:56:20mtyeah I don't think so
19:56:34mt-don't
19:56:59janwilfunman: i think i know why it don't work but have realy no idea how to fix this. Think the filesystem on the flash partiton is broken, it only reads the sdhc external thingy
19:57:49 Quit bmbl (Read error: Connection reset by peer)
19:58:07funmanis there a MTP setting on c200 ?
19:58:18bertrikfunman, on some firmware version, yes
19:58:52bertrikMSC is missing/hidden on some c200v1 firmware versions I mean
19:58:53janwilfunman: my firmware version shows the partitions yes
19:59:07 Join Buschel_ [0] (~~andree@p54A3A710.dip.t-dialin.net)
19:59:31 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
20:00
20:00:15janwilhehe fun, i now already have the bootloader on the system but it shows the rockbox logo saying no partition found
20:01:21bertrikhm, this may help: http://www.rockbox.org/wiki/SansaFAQ#Why_isn_t_installation_working_o
20:02:12pixelmathe PID is a c200 in UMS mode (as DeviceDetection says) so that's not the problem
20:03:26janwilbertrik: i think the problem i have is the filesystem on the chip in the c200 is causing the problem. When i put the sansa firwmare in it it works like a charm and i can put files on the external memorry but without it acts wierd
20:04:00bertrikSo what is the problem now exactly?
20:05:12janwilbertrik: i can't access the filesystem so can,t update the device correct
20:05:14bertrikYou said it read the sdhc external thingy, but as far as I know the original firmware does not support SHDC
20:06:07janwilbertrik: hmmm i now use a very old 1 GB micro sd card
20:06:54 Nick evilnick_B is now known as evilnizzles_B (0c140464@rockbox/staff/evilnick)
20:07:01 Quit Topy44 (Read error: Connection reset by peer)
20:07:53janwilbertrik: shall i recreate the partition table?
20:08:25janwilbertrik: fdisk -l did gave an error
20:08:31 Join kachna [0] (~kachna@r4ax178.net.upc.cz)
20:08:46gevaertsjanwil: is it partitioned?
20:09:00gevaertsoh, the internal flash?
20:09:05janwilgevaerts: i only see a 1 Gig partitoin on /de/sda1
20:09:14janwiland that is the sd card
20:09:32*gevaerts reads some backlog
20:09:43*bertrik happily leaves this to gevaerts
20:10:26gevaertsthat's weird
20:10:46janwilhmmm wheb removing the sdcard it shows the sandisk error not enough space. Please free memory.
20:10:55janwils/wheb/when/
20:11:04gevaertsah, so it's the OF behaving stupidly
20:11:18bertrikCan you still format it from within the OF?
20:11:22gevaerts(probably)
20:11:32janwiltreid but doesn't show much
20:12:01gevaertsbertrik: I suspect not. I suspect the OF is confused, and from what I know, once it's confused about its flash, it can't recover on its own
20:12:01janwilthat is why i was trying to use the e200tool to put a new c240erase.bin in it
20:12:11gevaertsand that doesn't work?
20:12:26janwilit can't find the device
20:12:42gevaertsare you running e200tool with root privileges?
20:12:48janwilgevaerts: yeps
20:13:09 Join DerPapst [0] (~Alexander@p5797C277.dip.t-dialin.net)
20:13:20janwilgevaerts: the wanring with not enough space, etc. end after a while and than i am in the sansa software
20:13:27janwilgevaerts: system info is wierd :)
20:13:44gevaertsyes, that's the symptoms I know where you need c240erase
20:14:05gevaertsDoes the c240 show anything on its screen while you're trying to run e200tool? Are the keypad lights on?
20:14:13janwilgevaerts: memory 33554441MB / Free 16976001MB
20:14:49gevaertsyes, it won't recover from that without c240erase
20:14:57janwilgevaerts: whell it starts in miniusb mode (recovery mode)
20:15:13janwiland keeps saying lun0 unlocked / lun0 locked / etc.
20:15:19gevaertsah, careful there. You need manufacturer mode, not recovery mode
20:15:27funmanmt: btw there is only *i*mdct there, no mdct; right?
20:15:43janwilgevaerts: oh ooops
20:16:00gevaertse200tool won't work with recovery mode :)
20:16:09janwilhmmmmm
20:16:32janwilok, i knew i did some wrong, but how to get the manufacturere mode
20:16:48gevaertshttp://www.rockbox.org/wiki/SansaE200Unbrick#Manufacturing_Mode
20:17:29 Nick Ypsy is now known as YPSY (~ypsy@geekpadawan.de)
20:19:06mtfunman: Yeah just the inverse
20:19:21janwilgevaerts: hmmmm i now says something like upgrade firwmae: erase sd media erase sd nvp and repeating that very often
20:19:23 Quit Buschel_ (Ping timeout: 276 seconds)
20:19:38gevaertsjanwil: let me grab my c200 to verify those instructions...
20:20:36janwilgevaerts: k, the e200tool stil doesn't find the device
20:21:09 Join petur [0] (~petur@rockbox/developer/petur)
20:21:10janwillsusb doesn't say something either
20:21:21gevaertsjanwil: try holding "select" (the middle button) instead of "down"
20:21:55janwilgevaerts: hey that is other ;)
20:22:02 Join Topy44 [0] (~Topy44@f049117056.adsl.alicedsl.de)
20:22:05gevaertsThe result (after plugging in or turning it on) should be a black screen and blue illuminated buttons
20:22:06janwilgevaerts: no it sees a device 0781:0720 Sandisk ocrp
20:22:19janwilgevaerts: nice init finds it
20:22:23gevaertsgood
20:22:26janwilshall i try to use the c240 recover?
20:22:30gevaertsyes
20:23:39janwilhmmmm no idea if it is doeing anything it doesn't say found now and is counting back :(
20:23:52gevaertse200tool is?
20:23:56janwilyeps
20:24:04janwillike not found
20:24:28 Quit Rob2222 (Read error: Connection reset by peer)
20:24:41 Join Rob2222 [0] (~Miranda@p4FDC9475.dip.t-dialin.net)
20:24:47janwilok
20:25:00janwiltried again and now its erasing the filesystem
20:25:04gevaertsnot good. Try unplugging, turning the c200 off, and try again. Turn on "hold", hold "select" (the middle button), and plug in. The immediately run ./e200tool recover c240erase.bin
20:25:05janwillooks good i think
20:25:09gevaertsok
20:25:10gevaertsperfect
20:25:25janwil122016 kB left
20:25:31janwilyay :)
20:25:38 Join wodz [0] (~wodz@inet20909nc-3s.nat.umts.dynamic.eranet.pl)
20:26:07gevaertsDon't skip the last step (reformat in the OF). I'm not *entirely* sure the formatting c240erase does is fully correct, and letting the OF format once makes sure it is
20:26:09janwili totaly love the device :) (it is a new one, the other one is somewhere with broken sound. have to let someone solder it
20:26:12 Quit wombat23 (Ping timeout: 260 seconds)
20:26:19janwilgevaerts: yeps
20:27:26gevaertsjanwil: just to verify, you had to hold "select", right?
20:27:38janwilgevaerts:
20:27:39janwilyeps
20:27:55gevaertsok, I'll fix the wiki and figure out who put the wrong info there
20:27:59janwilgevaerts: think the wiki has to be changed because there it says c200 submenu button
20:28:09gevaertsyes, that's clearly wrong
20:28:14janwilgevaerts: :)
20:28:22janwil103136 togo
20:28:51gevaertsyes, it takes a while :)
20:29:01*gevaerts has run that at least half a dozen times
20:30:19janwilgevaerts: looks likt there was some general info with the center button first and that was changed to e200 and c200. Maybe it work with some types like the other explanation
20:31:05 Join delt [0] (1000@mail.deimos.ca)
20:31:09deltHello
20:31:13gevaertsI doubt that, actually. I don't think there's that much variation between c200v1s
20:31:28janwilhi delt
20:32:05deltWhat capacity SD card can i use with a sandisk e260? (and does it matter what original firmware is on it?)
20:32:08janwilgevaerts: they are very hard to get here :( out of shop in allot of stores, only having newer versions
20:32:19deltin rockbox of course
20:32:32bertrikdelt, up to 32 GB as far as I know
20:32:37deltawesome
20:32:57 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
20:33:06janwilbertrik: nice the c200 versiosn only supports up to 8 GB i think
20:33:18deltthat's not depending of the firmware if i'm using rockbox, right?
20:33:26gevaertsjanwil: no, they should all work
20:33:38delt(looking at: http://answers.yahoo.com/question/index?qid=20081021112426AAxPOVB )
20:33:46bertrikjanwil, the c200v1 and e200v1 only support SD (not SDHC), rockbox supports SDHC on all c200 and e200 models
20:34:05 Join wombat23 [0] (~beuteltie@adsl-99-39-2-249.dsl.pltn13.sbcglobal.net)
20:34:12bertrikthe c200v1 and e200v1 original firmware I meant
20:34:20deltthanks for the info!
20:34:42janwildelt: but 32 GB SDHC is stil expensive :)
20:35:01janwil60 mb left
20:36:16deltok, if i buy a few 16gb or 8gb cards i can change them without needing to connect the player to a PC right?
20:36:47delt(i suppose i have to turn off the player to change it though)
20:37:18gevaertsdelt: no need for that
20:38:08janwilgevaerts: thats what i like for that systems
20:38:48deltheh... what about if it's playing a mp3 from the sd card and you remove it?
20:39:02delt(or do you need to eject it by software)
20:39:13janwildelt: then it freezes, whell the c200 did but you just can choose other file
20:39:18 Quit wombat23 (Ping timeout: 240 seconds)
20:39:30gevaertsit shouldn't freeze!
20:39:37janwilgevaerts: i mean the music ;)
20:39:43pixelmait doesn't freeze if there is no other problem
20:39:53gevaertsyes, it might not finish that particular track of course
20:40:05deltawesome
20:40:19deltthanks for the info =)
20:41:02 Join wombat23 [0] (~beuteltie@adsl-99-39-2-249.dsl.pltn13.sbcglobal.net)
20:41:21delti see a 32gb sdhc card for $55 canadian on ebay
20:41:42janwildelt: nice
20:41:44deltnot too bad
20:42:07 Join Luca_S [0] (www-data@giant.haxx.se)
20:42:12janwildelt: just bought a 8 GB SDHC for around 20 euro
20:42:53saratogaprobably fake if its that cheap on ebay
20:43:13 Join fml [0] (~chatzilla@p5DD2CC7C.dip.t-dialin.net)
20:43:16deltuh wait, this is a standard size sd card... the sansa c200 uses microsd right?
20:43:33janwildelt: yep
20:43:57***Saving seen data "./dancer.seen"
20:44:05deltbah i'll probably buy a few smaller ones at a store. less trouble
20:44:17fmlWhat do you think about this change to the manual (generalizing description of %d(N) wps tags): http://pastebin.ca/1892955 Can I commit it?
20:46:00janwildelt: that what i did :)
20:47:22gevaertsjanwil: after c240erase is done, you'll still have to go through the "regular" manufacturer mode recovery as described in http://www.rockbox.org/wiki/SansaE200Unbrick#Manufacturing_Mode
20:47:26*gevaerts is off now
20:47:42janwilgevaerts: yeps
20:47:49janwilgevaerts: bye and thanks
20:55:45CIA-6New commit by alle (r27221): Generalize the description of the %d(N) WPS tags
20:57:30CIA-6r27221 build result: All green
20:58:23 Nick scorche` is now known as scorche (~scorche@rockbox/administrator/scorche)
20:58:56 Quit sneak (Remote host closed the connection)
20:59:36 Join sneak [0] (debian-tor@gateway/tor-sasl/sneak)
21:00
21:05:05deltother question, what about the e200? Is microSD functionality similar to the c200?
21:05:11delt(sansa)
21:05:48delt(wikipedia) "Rockbox firmware also supports the e200 series just as it supports the c200 series, and adds the same features"
21:06:13delti'd guess so =)
21:06:15janwil:)
21:06:37bluebroth3re200 was even supported before c200 was
21:07:09deltheh =) but the player came out after...?
21:07:46delte200 == 2006, c200 == i think 2004 or 2005, not sure
21:08:18saratoga2006 or 2007
21:08:30 Part AzureWurk
21:08:36funmanyou can port rockbox to a 2002-old device if you want to
21:08:38saratogai think both came out at the same time, the hardware is nearly identical
21:09:12 Part watto
21:09:25deltexcept for screen size, & maybe other details
21:10:58deltbtw rockbox kicks ass, just wanted to let you know :D
21:11:00janwildelt: is it much bigger than c series?
21:11:14janwildelt: yeps, they do
21:11:20 Join AzureWurk [0] (opera@c-24-21-44-114.hsd1.wa.comcast.net)
21:11:38janwildelt: only the solder point of the headphone is a bit broken by hardware
21:11:39deltc200 128x98 pixels, e200 220x176
21:11:53bertrikthe e200 is quite a bit heavier than the c200
21:11:55deltyeah, i read that somewhere... can easily be fixed with a soldering iron
21:12:31janwildelt: i hope so, i have one with broken headphone and a new one, just trying to get rockbox on that one :)
21:13:46delti have a noname "4gb mp4 player" from hong kong that actually has 2gb (works fine if you make a 2gb partition on it) .... needs headphone jack fixed too
21:14:37delt(its only use right now is as a usb flash drive)
21:15:42 Quit wombat23 (Ping timeout: 245 seconds)
21:15:46janwildelt: i like the easyness and smallness of this thing with a nice big display c series that is
21:16:09janwilbertrik: ok think i fixed everything
21:16:13amiconndelt: c200 is 132x80
21:16:30deltoh, both c200 and e200 have that proprietary sandisk connector (30 pin?) and not a standard usb connector right?
21:16:38janwilbertrik: now it installs the bootloader perfectly and it find the internal memory :)
21:16:52janwildelt: yeps
21:17:02deltamiconn: ah, wikipedia is mistaken then (as often)
21:17:03janwiloh btw don't know about e series, sorry
21:17:27deltnot very hard to add one if you know how to use a soldering iron =)
21:18:03deltvery hard for an ipod nano because there's practically no space to put it :/
21:20:01pixelmajanwil: I wouldn't call the c200 display "nice"... suitable for that size maybe but nothing more
21:20:17 Quit Jerom (Quit: Leaving.)
21:20:23 Quit Luca_S (Quit: CGI:IRC (Ping timeout))
21:20:39janwilpixelma: hehe maybe that is because i didn't see a e series
21:20:50pixelmame neither
21:20:52janwilpixelma: sansa clip is too small for my taste :)
21:21:08 Join Jerom [0] (~heidi@95.171.137.241)
21:21:33delthoo yeah!! http://cgi.ebay.ca/ws/eBayISAPI.dll?ViewItem&item=200489130395
21:22:41janwildelt: http://www.highend.sk/shopping_cart/product_picts/1/sandisk_sdmx4_3.jpg
21:24:07 Join Luca_S [0] (www-data@giant.haxx.se)
21:26:00 Join Buschel [0] (~~andree@p54A3F3E2.dip.t-dialin.net)
21:28:47 Quit Luca_S (Client Quit)
21:31:16janwilhmmm new feature, pitch detector :)
21:31:17janwilnice
21:33:05 Part AzureWurk
21:37:43janwili am going away
21:38:01janwiltnx for al the help and keep up the good work :)
21:38:45 Quit janwil (Quit: bye)
21:40:15 Join AzureWurk [0] (opera@c-24-21-44-114.hsd1.wa.comcast.net)
21:46:55 Join lestatar_ [0] (~chatzilla@cpe-72-229-41-214.nyc.res.rr.com)
21:48:39 Part AzureWurk
21:49:11 Quit lestatar (Ping timeout: 260 seconds)
21:49:23 Nick lestatar_ is now known as lestatar (~chatzilla@cpe-72-229-41-214.nyc.res.rr.com)
21:51:37 Join AzureWurk [0] (opera@c-24-21-44-114.hsd1.wa.comcast.net)
21:53:52 Join stripwax [0] (~Miranda@87-194-34-169.bethere.co.uk)
21:54:19 Part AzureWurk
21:55:27 Join AzureWurk [0] (opera@c-24-21-44-114.hsd1.wa.comcast.net)
22:00
22:04:48 Join wombat23 [0] (~beuteltie@adsl-99-39-2-249.dsl.pltn13.sbcglobal.net)
22:09:11 Quit FF (Quit: CGI:IRC (Ping timeout))
22:09:42 Join Tarvis [0] (~Tarvis@97.100.203.106)
22:09:55Tarvishey, I have a question about ipod 5g hardware
22:10:09Tarviswill a 20gb hdd work with a 5.5g mainboard?
22:10:16Tarvisthose were sold in 30, 60gb
22:10:23Torneyes, should work
22:10:34Tornethe ipod isn't really bothered what sie the disk is
22:10:39TarvisI lucked out on ebay
22:10:45gevaertsif you get one with the right connector, it will work
22:10:48TarvisI found a bunch of good 20gbs for about $15, buy it now
22:11:07Tarvisgevaerts: the standard ZIF cable isnt good enough?
22:11:32gevaertsit is, but not all 1.8" drives are ZIF
22:12:04TarvisI see
22:12:11TarvisWere iPod 1-4G HDDs zif?
22:12:19Tarvisthe page says compatible with ipods 1-4th generation
22:13:46gevaertshave a look at http://www.rockbox.org/wiki/HardDriveReplacement
22:15:55Tarvishmm
22:16:07TarvisI dont think there's much room for an adapter then
22:16:35Tarviszune 30gb HDDs are the same model as an ipod video one, it seems
22:16:39TarvisMK3008GAL
22:16:48wodzjhMikeS: ping
22:16:55Tarvisunless dimensions are different
22:19:04Tarvisdoes anyone know about that?
22:19:23TarvisI've seen that theyre the same model but ebay says it won't work in an ipod, though they are both zif
22:19:48Tornethey should work..
22:21:00 Quit wombat23 (Ping timeout: 260 seconds)
22:21:55wodzunless HDD firmware is weird
22:22:05Tarvisshould work
22:22:17TarvisYou would figure the hdd model number would be different if so
22:22:24Tarvisbut the dimensions look the same
22:23:01Tarvisand photos look the same except the 'zune' one has orange squares on the sides of the zif port
22:23:16 Join MethoS- [0] (~clemens@134.102.106.250)
22:27:07 Join Transformer [0] (~Transform@ool-4a59e397.dyn.optonline.net)
22:28:43Buschelwodz: hi again. can you test the latest patch in FS #11445 for speed?
22:29:54wodzyes
22:29:58 Part Transformer
22:30:02Buschelgreat :)
22:30:26CIA-6New commit by bieber (r27222): Theme Editor: Renderer now adds a newline after EVERY logical line in the parse tree
22:30:31Tarviswow, I'm dumb
22:30:38TarvisI intended to put that in -community
22:30:40Tarvissorry
22:31:10CIA-6New commit by nls (r27223): Factor out sane code list bisection code and use instead of insane code, clean up some tests for readability's sake. Gives a small speedup on h300 ...
22:31:49CIA-6r27222 build result: All green
22:33:25CIA-6r27223 build result: All green
22:34:26*wodz is compiling
22:36:02bieberAnyone know how images work in progress bars in the WPS?
22:36:52n1sbieber: it just shows a portion of the image, no?
22:36:56bluebroth3rbieber: it shows only the left part of the image
22:37:03bieberOkay, thanks
22:37:13deltOH - IMPORTANT!! Does that mikmod (tracker module/sequencer format) patch still work, and has it been updated in the last.... 2 years or so :D :D
22:37:24bieberJust in proportion to the percentage played I take it?
22:37:29n1sdelt: dunno, check the tracker entry?
22:37:36deltiirc the name of the author seemed to be vietnamese or thai
22:38:03deltum, ... "tracker entry" on rockbox.org website?
22:38:16*bluebroth3r grumbles at the additional loops needed to build universal binaries on Mac when using libs :/
22:38:19n1syes, see "Patches"
22:38:38*delt *checks on rockbox.org*
22:40:27deltoh, "tracker" as in cvs/svn/mercurial/etc... =-} (as opposed to ScreamTracker, ImpulseTracker, FT2, SchismTracker, ....)
22:41:10gevaertsuhm, no :)
22:41:19n1sno, as in bug/patch tracker :)
22:41:20gevaertscvs, svn, and mercurial are not trackers
22:41:30deltoh, thats what i meant, sry
22:41:30bieberbluebroth3r: Sorry the renderer isn't going quite according to schedule. I'm thinking I should be able to get everything knocked off my to-do list for it next week, and move on to graphical editing
22:42:01delt"patch tracker" which is most often used with a versioning system such as cvs, svn, .....
22:42:06wodzBuschel: nero_192.m4a 233.18% realtime 53.25MHz
22:42:14deltas opposed to " " " "
22:42:44bluebroth3rbieber: let me check the schedule ...
22:43:34Buschelwodz: the latest measurements from yesterday resulted in 55 MHz, correct?
22:43:40wodzyes
22:43:58 Join Topy [0] (~topy@2a01:608:ffff:2619:2e0:4dff:fe83:a45e)
22:44:00***Saving seen data "./dancer.seen"
22:44:01bluebroth3rhmm, the timespan between midterm and pencils down is shorter than I remembered :/
22:44:31n1sis FS #11438 clip specific? i was pretty sure we didn't have music in doom on any player.
22:44:32deltpencils down as in, "test finished, stop writing NOW" ...?
22:44:53n1sdelt: the end of google summer of code
22:44:59bluebroth3rbieber: do I understand you correctly that you want to finish the renderer next week, then graphical editing afterwards?
22:46:09deltclipv2 ie. the sandisk player? −−-> I know it doesn't work (yet) but I wanted to point out that DOOM works now for Clipv2 !!!
22:46:53Tarvishow would that work?
22:46:53TarvisDOOM in monochrome?
22:47:43Tarviswell, not monochrome
22:47:43n1sTarvis: doom itself works but there's no music
22:47:43Tarvisbut there isnt much you can do with black, yellow, and blue
22:48:59n1syou can emulate geyscale if you flip them fast enough
22:48:59TarvisI see
22:48:59TarvisWow, that's insane
22:48:59Tarvishttp://www.youtube.com/watch?v=1rVV_0xKmhw
22:48:59 Quit fml (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
22:49:01delt4 lines of text about 12 pixels each, that would be around 72x48 pixels?
22:50:39Bagder128x64 I believe ?
22:50:43bieberbluebroth3r: That's what I'm shooting for. I get the feeling that I really should have reversed the time-frames for rendering and graphical editing in my original schedule, as once the renderer is all together it shouldn't be much trouble at all to work in some graphical editing
22:51:34bluebroth3rbieber: that's fine with me. IMO the graphical editing part is less important than the rendering anyway :)
22:51:47bieberCool
22:51:56bieberHave you gotten a chance to look at the renderer so far?
22:52:06bluebroth3rnot at the code itself, no.
22:52:07delttarvis: heh.... that looks like an old TV picking up static =) is that a demo running? (the guy's hands arent touching the player)
22:52:14Tarvisprobably
22:52:29deltwell, except when he's starting the game
22:52:29bluebroth3rI've made a new Windows binary a bit earlier today in case someone wants to test. At the usual location.
22:52:30TarvisHes running around and aiming quite well for just a couple buttons
22:52:57bieberOh, bluebroth3r, do you have a Mac as well?
22:53:34bluebroth3rbieber: yes, and I'm currently looking into the lib stuff for building universal binaries. Unfortunately that is always a bit nasty.
22:53:36deltimpressive nonetheless for a... what, 4 color screen, w/the top one fixed to yellow?
22:53:58bieberI got a friend to compile the theme editor and run it, and he said the project panel and renderer were just altogether not working
22:54:22bieberWhich could have been a case of him just grabbing the wrong source package, or something could just be going horribly wrong in OSX compiles
22:54:27bluebroth3rI was planning to move ipodpatcher / sansapatcher for Rockbox Utility into libs too, but simply haven't gotten around.
22:55:05bluebroth3rI've build an i386 binary earlier today and opened a simple file. It did show some text (no backdrop due to wrong filenames though, but that's not an issue of the editor)
22:55:18Tarvisdelt: the top being yellow is probably just the camera's refresh rate
22:55:29Tarvisit probably flickers to blue too to convey a brightness
22:55:41Tarvisit probably looks much smoother than it does in the video
22:55:50gevaertsTarvis: the clip's screen is yellow at the top and blue for the rest
22:55:58Tarvisoh, I see
22:56:07gevaerts16 rows of yellow, 48 rows of blue
22:56:08deltnah, on wikipedia it says: 4-line OLED screen (one line yellow, three blue.)
22:56:27n1sit's a bitmap screen.
22:56:31delt"organic" LED? wondering......
22:56:50funman'lines' probably refer to the OF using 16 pixels high font
22:57:10n1syes but it's an incorrect description of the display
22:57:20deltmy thoughts exactly as i read 16:52 < Bagder> 128x64 I believe ?
22:57:20delt16:52 < Bagder> 128x64 I believe ?
22:57:34deltuh oops, forgive my stuck middle mouse button :/
23:00
23:00:26 Quit bmbl (Quit: Bye!)
23:01:36 Quit Topy44 (Quit: Leaving)
23:02:13CIA-6New commit by pixelma (r27224): Theme tags description: include the part about 'hold' for soft hold targets as well (available there since hm... r13716), put more universally to get ...
23:03:40CIA-6r27224 build result: All green
23:04:04deltbtw - is there a way to boot into original firmware on the sansa e200? like, flip on "hold" right after turning it on, ...?
23:04:29bertrikhold left I think, should be in the manual
23:04:38deltthanks
23:04:56delt(havent looked at the manual yet, will do when i get the player)
23:05:18*bluebroth3r wonders why it's libskin_parser and not libskinparser
23:06:02n1sit parses libskins?
23:08:40bluebroth3rprobably :)
23:09:36*bluebroth3r gets tempted to dump the whole Makefile and start over with the one from mkamsboot
23:10:39 Quit jgarvey (Quit: Leaving)
23:12:05pixelmabluebroth3r: got a little question about LaTex capabilities. I wonder if the following would be possible to solve the \Actions for targets where this is actually defined twice (e.g. remote targets) in text descriptions in an elegant way: have an \ActionAllSomething in the text and a rule somewhere to combine \ActionSomething and \ActionRCSomething if the latter exists, otherwise just use one of them?
23:14:56deltah, FOUND IT! FS #8806
23:17:39 Quit stripwax (Quit: http://miranda-im.org)
23:18:46CIA-6New commit by Buschel (r27225): Submit FS #11445. Speed up of faad (aac) decoder via several optimizations like refactoring some requantization routines, moving several arrays and ...
23:18:54*Buschel crosses fingers
23:19:14funmanwould be nice to have test_codec results after each build round ;)
23:19:37Buschelin fact
23:20:26CIA-6r27225 build result: 4 errors, 6 warnings (Buschel committed)
23:20:32Buschelnah :/
23:20:37bluebroth3rpixelma: should be doable with something like \if\ActionRCSomething\newcommand{\ActionSomethingAll}{\ActionSomething \ActionRCSomething}\else\newcommand{\ActionSomethingAll}{\ActionSomething}\fi
23:21:02bluebroth3rthough I have never done something like that before :)
23:21:52jhMikeSwodz: pong
23:22:56wodzjhMikeS: I am reading audio-iaudio.c and you seem to be the author
23:23:35wodzand audio-iriver.c
23:24:03n1sfunman: just need to hack up something with usb serial and hook a lot of targets up to a server somewhere :)
23:24:47wodzdo I understand correctly that audio_set_output_source() routes samples from ADC back to DAC and forms 'digital monitor'?
23:25:05wodzjhMikeS: ^
23:25:52 Quit Dhraakellian (Ping timeout: 248 seconds)
23:27:38pixelmabluebroth3r: would that mean that I have to prepare every \ActionAllSomething beforehand or could the necessary strings set together from the \ActionAllSomething so that LaTex knows it has to take the Something part and use it with \Action and \ActionRC (maybe dependent on HAVEREMOTEKEYMAP)? My goal would be the latter, just wondering if it's worth some research and trial and error for me
23:27:50jhMikeSwodz: It's been awhle. what about it?
23:28:21jhMikeSoh, yeah, output source is loopback when the same as the input
23:29:11pixelmabluebroth3r: I just don't know if that's possible as that's going back and forth a bit while parsing I guess
23:29:45 Nick evilnizzles_B is now known as evilnick_B (0c140464@rockbox/staff/evilnick)
23:30:15wodzjhMikeS: what makes me wondering is that on irivers inputs are connected to IIS1 and playback out to IIS2. I can't find setup for IIS1
23:30:17jhMikeSwodz: really, just *an* input, if loopback is appropriate. in the case of S/PDIF or the like, output is taken from the S/PDIF
23:31:09jhMikeSwodz: iirc it's in the pcm driver
23:31:27bluebroth3rbieber: looks ok to me: http://www.alice-dsl.net/dominik.riebeling/mig/albums/rockbox/themeeditor-r27225-PenPaper.png
23:31:48bieberExcellent, thanks
23:32:09bieberExcept the transparency doesn't seem to be working
23:32:39bluebroth3rhaven't tried doing anything with that though. Just loaded the wps and made the screenshot :)
23:32:46wodzjhMikeS, in pcm-coldfire.c it seem to setup only 'play' IIS
23:36:13jhMikeSwodz: right, I was just checking again. it may have been that the defaults may have been appropriate values.
23:36:53 Quit Jerom (Read error: Connection reset by peer)
23:36:58wodzjhMikeS: are codecs in irivers and iaudios in IIS master mode or slave?
23:37:35 Quit evilnick_B (Quit: Page closed)
23:39:56CIA-6New commit by Buschel (r27226): Fix red and yellow. Needed to remove one codebook from IRAM due to full IRAM on iPod4G.
23:41:29 Quit powell14ski_ (Ping timeout: 264 seconds)
23:41:36CIA-6r27226 build result: All green
23:41:52Buschelthat's how I like the builds to finish :)
23:42:09jhMikeSwodz: on one of them, it gets switched for record mode. I remember it was easier to manage that way. Think it was iRiver.
23:43:15 Join Dhraakellian [0] (~ntryon@cpe-67-240-248-41.rochester.res.rr.com)
23:44:50bluebroth3rbieber: the cfg editor looks a bit strange though. Seems it doesn't skip comments at all
23:45:01bluebroth3ryep, can reproduce the issue on Linux as well.
23:45:37bieberThe idea is that the editor can embed its own data as comments by prefacing a key with a #
23:46:09bieberAt the moment, though, I'm thinking it might be better just to back to a plain text editor for cfg docs, the separate inputs just seem too cludgy
23:46:19bluebroth3rwell, at least the theme I've tried has a comment block at the beginning that states the license
23:46:33bluebroth3rwhich looks somewhat weird split up in that editor :)
23:46:39 Quit petur (Quit: Zzzzz)
23:47:21bluebroth3rso maybe that approach needs a bit refinement
23:48:46bieberDefinitely
23:49:16bieberIf it stays the way it is, perhaps I could just identify comment lines and put the whole thing on one long line-edit
23:49:58CIA-6New commit by bieber (r27227): Theme Editor: Fixed line numbering bug in parser. Implemented playlist display in renderer: playlist will use info for next track for all tracks ...
23:51:31CIA-6r27227 build result: All green
23:55:48 Quit esperegu (Read error: Connection reset by peer)
23:57:05 Quit wodz (Quit: Leaving)

Previous day | Next day