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 2004-09-07

00:00:40amiconnLinusN: Perhaps the tag handling in mpeg.c should be rewritten, because of several reasons:
00:01:06LinusNamiconn: yes
00:01:09amiconn(1) It has a static limit of 16 tracks
00:01:31amiconn(2) If this limit is exceeded, it seems to cause problems
00:01:56LinusN1) we want static limits
00:02:03LinusN2) that's the bug
00:02:25amiconn(3) It uses both an array of structs and a pointer array to these. Why it's designed that way is beyond me
00:03:02amiconn(4) There are frequent NULL pointer accesses while playback
00:03:20LinusN3) that's an old relic from the days the structs were allocated from the pool of 16 structs
00:03:34LinusN4) yes, that's another bug
00:03:47amiconn(5) the struct array for 16 tags consumes a whopping 24 K of ram (bss)
00:04:20LinusN"whopping" :-)
00:04:48amiconnFor what it does, I consider this "whopping"
00:05:09amiconnI wonder why it is necessary to convert & buffer the tags of all loaded tracks at load time.
00:05:39LinusNbecause we don't want to spin up again just to read the tags?
00:06:22amiconnAren't the tags loaded interspersed with the mp3 data, so that they are in ram already, between the tracks?
00:06:48LinusNthey aren't loaded into the mp3 buffer
00:07:01LinusNonly mp3 data is in the buffer
00:07:20LinusNand we need a copy of the tags, since the tag data will be gone when the buffer fills up the next time
00:07:36LinusN(if they were loaded into the mp3 buffer, that is)
00:07:51amiconnI'd rather suggest to only store the track start index at load time (this table could have many more entries, since these are simple pointers)
00:08:33amiconnThe tags could then be decoded when needed (e.g. when the previous track starts)
00:08:42LinusN"track start index"?
00:09:20amiconnYes, the buffer index where the track starts
00:09:37LinusNyou mean load the tags into the mp3 buffer?
00:10:07amiconnYes, although this is slightly more work for v1
00:10:27LinusNyes it is, and the swapping will be trickier
00:11:09LinusNsince you mustn't swap the tag data
00:11:20amiconnOf course
00:12:10amiconnMy first though was to place the already decoded tags into the buffer, but this won't work when the buffer wraps in the middle of the struct
00:12:24LinusNwe have discussed this approach several times, and found it pretty complicated
00:14:08amiconnIf we stay with the current approach (an array of structs) I still don't get why there are 16 pointers, allocating from a pool of 16 structs. This is complicated too, for no reason. We could use the struct array directly as a ring buffer
00:14:26 Quit edx ()
00:15:23amiconnOn the contrary, if we would buffer the tags within the mp3 buffer, the ram allocation for tags would be dynamic.
00:15:48LinusNamiconn: did you read my answer for your question 3)?
00:17:04amiconnYes, I read that, but I still don't get why this was done. Did the usused struct got used for something else?
00:17:54LinusNi don't remember why it was designed like that back then
00:18:21LinusNi think it might come from the days when we had a malloc
00:18:22amiconnFor the typical use as a music player, there are mostly 1..2 tracks in memory, so having 16 slots is wasting ram. On the other hand, if you have short clips, 16 slots might even be too few :(
00:19:46LinusNyes, i agree that we need a better way of handling this
00:20:20LinusNbut i think loading the tags into the mp3 buffer will lead to complicated code
00:20:21lImbuswhere's the problem to make this variable ?
00:20:37LinusNwe don't have dynamic memory
00:20:54lImbusuhh. that's some argument.
00:21:31lImbushow are the limits for "max files in dir browser" and playlist handled ?
00:21:52amiconnThese pointers are btw the reason for the NULL pointer accesses.
00:23:24LinusNlImbus: they are taken from the mp3 buffer at boot time
00:23:30LinusNamiconn: ah
00:25:50amiconnLinusN: You said that buffering the tags within the mp3 buffer complicates the bitswapping. This is only true if the tracks are swapped at load time. We could swap before playback instead (the next dma chunk)
00:26:33amiconnThis could in fact help portability - chances are that another architecture doesn't need bitswapping
00:27:38amiconnI consider bitswapping to be a part of the "mp3 driver", while handling the tags, track changes, reloads etc. is part of the playback engine
00:27:49LinusNfinding the next dma chunk will still be more complicated
00:28:18LinusNtoday, everything in the mp3 buffer is mp3 data, so we can just grab the next 64k
00:28:33amiconnWe don't...
00:28:43lImbusok, that's what I meant. isn't the case of having 16 tracks beeing smaller than 2 MB (or even 8MB) small enough to make that configurable like max dirbrowser ?
00:29:03lImbusFrihet must have chunks smaller than 128 kB
00:29:22FrihetI can look
00:31:32FrihetThe MP3 files average about 60KB
00:32:37FrihetThey are short phrases like "Hvordan har du de"
00:33:21LinusNamiconn: yes, there is the special cases when the end of the buffer is reached, and at track boundaries
00:33:55LinusNbut the tags-in-buffer approach complicates it even more
00:34:13amiconnLinusN: Yes, exactly. We already check for track boundaries, so having the tag data at the beginning of the next tracks shouldn't add much complexity
00:34:18LinusNi'm not saying that it is hard, i just mean that the code will get harder to understand
00:35:03LinusNand when the tags are big, we will waste buffer space, like when some silly person sticks album covers in the tag
00:35:35amiconnIf we pre-decode the tags and put the struct at the beginning of the track, the size will be constant.
00:36:04amiconnThis has to be copied when starting playback of that track, since otherwise it is gone at the next reload
00:36:37amiconnThe only special case would be when the buffer would wrap within the tag data, but that can be avoided by wrapping prematurely
00:36:39LinusNyup
00:39:00amiconnBtw: my interpretation of the posix rules for O_TRUNC is correct :)
00:39:07LinusNgood
00:40:03amiconnSo I'll try what you suggested, and hope it doesn't ruin my box' file system (of course I have a backup)
00:41:24*amiconn digs for the correct irc backlog
00:43:16FrihetI will leave bug 949268 with the experts and check in later in the week. If it's too big a deal, just let me know.
00:43:48***Saving seen data "./dancer.seen"
00:44:07 Quit Frihet ("ChatZilla 0.9.52B [Mozilla rv:1.6/20040114]")
00:44:43amiconnLinusN: As you probably found, I implemented the new screen_dump() without using assembler ;)
00:46:27 Join SmoothOperator [0] (~c35d2208@labb.contactor.se)
00:46:36 Quit SmoothOperator (Client Quit)
00:56:12LinusNamiconn: saw that, and was pleased :-)
00:56:17LinusNsleep time
00:56:18LinusNnite all
00:56:24 Part LinusN
01:00
01:10:38 Join Frihet [0] (~chatzilla@38.119.217.96)
01:13:07 Quit Frihet (Client Quit)
01:26:34 Join SmoothOperator [0] (~acbeb032@labb.contactor.se)
01:42:05 Quit SmoothOperator ("CGI:IRC (Ping timeout)")
02:00
02:23:05 Part amiconn
02:43:50***Saving seen data "./dancer.seen"
02:49:22 Join NegativeK [0] (~negk@68.114.40.130)
03:00
03:01:30 Quit elinenbe (" HydraIRC -> http://www.hydrairc.com <- irc client ownage!")
04:00
04:01:30 Quit lImbus (Remote closed the connection)
04:23:29 Quit NegativeK ("Leaving")
04:23:30 Join NegativeK [0] (~negk@68.114.40.130)
04:43:54***Saving seen data "./dancer.seen"
05:00
05:21:29 Join grogro [0] (~gromit@ALagny-151-1-23-64.w83-114.abo.wanadoo.fr)
05:39:50 Quit maikeul (Read error: 110 (Connection timed out))
06:00
06:27:50 Join midk [0] (~midk@c-24-18-39-204.client.comcast.net)
06:35:46 Join scott666_ [0] (~scott666@c-24-245-58-48.mn.client2.attbi.com)
06:35:51 Part scott666_
06:43:55***Saving seen data "./dancer.seen"
07:00
07:32:33 Join LinusN [0] (~linus@labb.contactor.se)
08:00
08:09:05 Nick midk is now known as midk|sleep (~midk@c-24-18-39-204.client.comcast.net)
08:20:21 Join hardeep [0] (1098@208.247.65.237)
08:20:48hardeephey, does anyone know where I can download some sample rvf files?
08:23:28hardeepaha, found some: http://rockbox.haxx.se/newvid/
08:25:15 Join Zagor [242] (~bjst@labb.contactor.se)
08:26:28LinusNZagor: today is payday for you, you owe me a lunch :-)
08:26:39Zagorsure
08:27:36dwihno:-)
08:36:28 Part LinusN
08:36:38 Join LinusN [0] (~linus@labb.contactor.se)
08:37:41dwihnoConTraktor
08:43:56***Saving seen data "./dancer.seen"
08:45:55Zagorwhoa, what happened to todays daily build?
08:47:24dwihno?!
08:47:48Zagorhmm, bug and patch lists are empty too. network blackout?
08:47:50dwihnoPerhaps somebody tried integrating the rombox build (Badger?)
08:48:23 Join [IDC]Dragon [0] (~d9ff8454@labb.contactor.se)
08:48:40[IDC]Dragon'morning from me logpeeker
08:48:46Zagor:)
08:49:09[IDC]Dragonhardeep: here are some really nice videos: http://rockboxvideo.redbreva.co.uk/
08:49:45[IDC]Dragonand a few links to mediocre ones: http://rockbox.haxx.se/mail/archive/rockbox-archive-2004-02/0189.shtml
08:52:08[IDC]Dragonhardeep: got enough bandwidth?
08:53:32[IDC]Dragondid anybody check if the daily ROMbox build works?
08:56:05 Join Bagder [0] (~daniel@1-1-5-26a.hud.sth.bostream.se)
08:56:42BagderZagor: you checked why the daily builds failed?
08:56:46dwihnoWhen you talk of the trolls ;)
08:56:58Zagorlooking at it now. the bug and patch lists are empty too.
08:57:10ZagorBagder: you check the builds, i'll check the lists ok?
08:59:27Bagdersomething truly weird happened
08:59:44Bagderthe size of the tarball package is a lot smaller today
09:00
09:00:03[IDC]Dragon(my question somehow got answered)
09:01:37BagderI can't see how the problems can be related
09:01:51BagderZagor: didn't you get some clues in a crontab mail?
09:02:02Bagderor is it dev/nulled?
09:02:44Zagori think i deleted it :(
09:03:40[IDC]Dragonbbl
09:03:58[IDC]Dragon(learned that acronym yesterday)
09:04:01 Part [IDC]Dragon
09:06:36Bagderhm
09:07:45LinusNBagder: bleeding edge and daily builds are done by differents scripts, right?
09:08:16 Quit Bagder ("Leaving")
09:17:51LinusNZagor: u there?
09:24:47Zagorrerunning daily builds...
09:26:20webmindmorning
09:26:31Zagorhi
09:29:07 Join amiconn [0] (~jens@pD9E7DE1A.dip.t-dialin.net)
09:29:25amiconnmorning
09:32:14LinusNhi
09:32:37LinusNamiconn: regarding the tag discussion
09:33:30LinusNwe need to keep the current and the upcoming tag in memory, otherwise the next-track wps info won't work
09:33:37amiconnOf course
09:34:08Zagordaily builds now work. the problem was "configure update" mangled the Makefile for some of the targets.
09:35:03 Join amiconn_ [0] (~jens@pD9E7F378.dip.t-dialin.net)
09:35:43 Join amiconn__ [0] (~d9092244@labb.contactor.se)
09:36:41 Quit amiconn (Nick collision from services.)
09:36:41 Nick amiconn_ is now known as amiconn (~jens@pD9E7F378.dip.t-dialin.net)
09:37:11 Join Bagder [0] (~daniel@1-1-5-26a.hud.sth.bostream.se)
09:38:52amiconn__Zagor: The tarball is now much larger than the last few days...
09:39:04Bagderlooks wrong
09:40:42Bagderand so it is
09:40:56Bagderit contains a rockbox-daily-20040907 dir
09:41:00Bagderwith contents
09:41:24Bagderlike rockbox-daily-20040907/rockbox-daily-20040907/apps/
09:41:41LinusNi guess someone has unpacked it in the build dir when we examined the build errors
09:42:03Bagderyes "somone" must have ;-)
09:42:10*Bagder can't type
09:42:27LinusNbut he can unpack a tar file :-)
09:42:38Bagder*I* didn't do that
09:42:46LinusNsorry :-)
09:42:53Bagdersince I can't type, I can't unpack either ;-)
09:42:59amiconn__LinusN: To flash that into the box?
09:43:03*LinusN points to Zagor
09:43:16*Bagder joins LinusN in the pointing
09:43:23amiconn__Ooops, I read "uclpack" instead of "unpack"
09:43:53Mode"#rockbox +o LinusN " by ChanServ (ChanServ@services.)
09:44:05LinusNamiconn__: want to kick amiconn?
09:44:14amiconn__LinusN: No.
09:44:18LinusNok
09:44:32amiconn__This is my home box, I just have to wait 15 min to re-connect to it
09:44:37Zagorwell i didn't unpack anything anywhere
09:45:11LinusNweird
09:45:31Bagderit might be because it was re-run on the same date
09:45:34Bagdersomehow
09:45:41Bagderlet's ignore it and hope it works better tomorrow
09:45:44 Join kurzhaarrocker [0] (~knoppix@p50877DD2.dip0.t-ipconnect.de)
09:45:45LinusNor we ran the update.sh script in the wrong dir
09:46:13BagderZagor: did you update the uclpack tool?
09:46:44Zagornope, didn't find the source then got busy with other stuff
09:47:00Bagderthe source is in the RomBox wiki page
09:47:14Zagorno, that's just one file. more is needed.
09:47:26Bagderaha
09:47:32Bagderyes, ucl-1.01
09:47:43BagderI'll fix it
09:47:46Zagorthanks
09:48:49Bagderit uses the /usr/local/bin one, right?
09:50:25Bagderit is updated now anyway
09:50:26Zagori think so yes
09:51:02Bagdergcc 3.0.4 makes a 5K larger ucl file than 3.3.1
09:51:32Bagderwe should upgrade the cross-compiler stuff on labb as well
09:51:54Zagoryeah
09:53:02Bagder"Output is 216 bytes larger than max (188400)"
09:53:11Bagder3.0.4 can't make a rombox for the v2
09:53:21Zagorboo
09:55:25Bagderannoying
09:55:32 Quit amiconn__ ("CGI:IRC 0.5.4 (2004/01/29)")
09:58:46amiconnBack for real :)
10:00
10:03:01 Join [IDC]Dragon [0] (~d90a3255@labb.contactor.se)
10:03:12[IDC]Dragonback again
10:03:28amiconnhi Jörg
10:03:49[IDC]Dragonno Ondio luck yesterday? (saw your bid)
10:04:04amiconnWas too expensive imho
10:04:16amiconnToday another Ondio is due (in the evening)
10:04:18[IDC]Dragoncheapskin!
10:07:48[IDC]Dragonamiconn: on a different one: nice screendump function
10:08:29[IDC]Dragonor, bad one before, 1.8K buried, ts, ts
10:09:00[IDC]Dragondo we have more like that?
10:09:28LinusNvery likely
10:09:39[IDC]Dragonwhere?
10:10:10LinusNthere is a possibility that we have more debug functions that have ended up in the official build
10:10:39LinusNand those are probably not very optimized
10:11:10[IDC]Dragonscary...
10:11:30LinusNnot at all, that's how things work
10:11:54LinusNthere is a first version of everything, then you improve it
10:12:05amiconnLinusN: I'd suggest eliminating most of the debug menu for the upcoming 2.3 release, leaving only stuff that could really help to spot problems
10:12:17LinusNamiconn: absolutely
10:12:22[IDC]DragonI was joking, don't take it too serious
10:12:33LinusN[IDC]Dragon: no offense taken
10:13:03LinusNwe can at least put the debug functions on a diet
10:13:17*[IDC]Dragon has a deja-vu: "upcoming 2.3 release"
10:13:43LinusNwe need to find the nasty recording bug that Paul experiences
10:13:55amiconnCandidates would be the mas register & codec debug...
10:13:56[IDC]Dragon3KB to go for "native FM" rombox
10:13:59LinusNwith lots of corrupt frames
10:14:39[IDC]DragonLinusN: can we reproduce that?
10:14:56LinusNamiconn: do you get corrupt frames?
10:15:16*[IDC]Dragon shoul try recording
10:17:08amiconnLinusN: Usually I almost use recording. I plan to do even more extensive recording tests, though
10:17:29amiconnI would need a reliable tool to spot corrupt frames for that
10:18:03[IDC]Dragonamiconn: have you looked at my frame parser in rvf_mux?
10:18:17amiconnmp3utility is a bit questionable, and mp3fixer is definitely buggy, at least the windows binary
10:18:28amiconn[IDC]Dragon: No, not yet
10:18:38[IDC]Dragonshould be easy to make a tool from it
10:19:15[IDC]Dragonit walks the headers, inspects them all
10:19:31[IDC]Dragonand recognises lost/regained sync
10:19:38amiconnDoes it check the frame checksum (if there is one)?
10:19:50[IDC]Dragonno
10:20:08[IDC]Dragonfeel free to add that ;-)
10:20:54LinusNthe code in mp3data.c and the vbrfix code should be a good base to build upon
10:21:14[IDC]Dragonso we got choices
10:28:28amiconn[IDC]Dragon: As you asked whether we have more functions that use a lot of ram: look at the ro(ck|m)box.map, especially the bss space each module takes
10:30:13[IDC]Dragon"ro(ck|m)box", hehe
10:30:58[IDC]Dragonyes, that's a start. but I don't know how much is "justified", and no names for static stuff
10:32:58amiconnOne of those little ram suckers is mpeg.o, because of the tag handling...
10:33:22[IDC]Dragonputting salt on LinusN's wounds?
10:33:42[IDC]Dragon(if that german saying translates correct)
10:43:14*[IDC]Dragon still thinks malloc would be nice
10:44:00***Saving seen data "./dancer.seen"
10:44:00Bagdermalloc for what?
10:44:27Zagormalloc is a waste of ram
10:44:28[IDC]Dragonfor not having worst-case static allocation all the time
10:44:52Zagorthe only result will be lovely out-of-memory handling code all over the place
10:44:58[IDC]Dragonit's a tradeof, not a waste
10:45:44[IDC]Dragonout of memory is an assert within malloc
10:45:45Zagorok i won't dismiss it right off like that. where do you suggest we'd benefit from it?
10:46:14[IDC]Dragonthe tag handling, for example
10:46:16amiconnI'm tempted to rewrite the tag handling (dynamically within the mp3 buffer, as I discussed with LinusN). I wouldn't need malloc for that
10:46:30Zagorbut with an assert we'd crash whenever we run out of memory!?
10:47:23[IDC]Dragonwe have the mp3 buffer to munch from, we won't run out of mem for the other stuff
10:47:53Zagorit's pretty icky munching from the mp3 buffer while music is playing
10:48:19*[IDC]Dragon thinks vaguely about a way to have the unallocated as a linked buffer list
10:48:35Bagderit is possible
10:48:44Bagderbut icky
10:48:49[IDC]Dragonwhenever mem is needed, it should be taken from the played end
10:49:00Bagderbut...
10:49:03[IDC]Dragonif not possible, from the newes unplayed
10:49:20Bagderso it would need to reread data?
10:49:28[IDC]Dragonyes, something like that
10:50:14amiconnAn ugly thing with malloc is memory fragmentation (without having an mmu of course). How would you handle that?
10:50:34dwihnoMMU?
10:50:49Bagderit fragments even with MMU
10:50:59[IDC]Dragonas we wade through, it will defragment again
10:51:13amiconnBagder: Yes, but you can remap the unused pieces into one region
10:51:45LinusNwe have had this discussion countless times
10:51:55[IDC]Dragonthere's different algorithms for reducing fragmentation
10:52:01[IDC]DragonLinusN: sorry
10:52:23LinusNdon't worry, this time we might come to a different conclusion
10:52:31Bagderreducing is the word, it can't be avoided
10:52:36[IDC]Dragonsounds optimistic
10:53:15BagderI still haven't read a good argument why it is good in current rockbox
10:53:28LinusNBagder: agreed
10:53:32[IDC]Dragonwell, by profession I'm working on an embedded system for mobile phones
10:53:53[IDC]Dragonand we malloc like crazy
10:54:01[IDC]Dragonno problems yet
10:54:14BagderI'm sorry, but that doesn't really answer my question
10:54:28BagderI too usually use malloc a lot
10:54:53[IDC]Dragonwe're probably not even aware any more in how many places we put memory aside
10:55:45[IDC]Dragonlike, to communicate the recording pathname to the mpeg task, there are 2 static buffers
10:56:33[IDC]Dragonreally needed only until the taskswitch
10:57:18BagderI believe the cost of adding code for handling segmented mpeg playback, malloc and everything is much greater than the amount of memory currently "wasted" due to our lack of malloc - but this is only my suspicion
10:57:35[IDC]Dragona valid one
10:59:13Bagderbtw
10:59:18[IDC]Dragoncertainly I also don't know, it's just by gut feeling I woudn't build an embedded system of this size without malloc
10:59:24BagderI think we should release 2.3 soon
10:59:30[IDC]Dragon;-)
11:00
11:01:38LinusNalso consider the amount of debugging time we have gained by not having to hunt memory leaks
11:02:08Zagoror oom crashes
11:02:17LinusNor defragmentation problems
11:02:51LinusNdynamic memory is a two-edged sword imho
11:02:57 Join lImbus [0] (~manuel@kernel.cycos.net)
11:03:05LinusNi avoid it if i can
11:03:12*[IDC]Dragon shuts up on the subject
11:03:16LinusN:-)
11:04:41[IDC]Dragonhow about a call for 2.3 feature freeze?
11:04:54[IDC]Dragonor are we not freezing any more?
11:04:58Bagdergood idea
11:05:10[IDC]Dragonwe appear frozen already
11:06:16[IDC]DragonI'll probably folk (temporarily) for Ondio, but that's a different story
11:06:31Bagderwe should fix the gcc on the build box before release so that we can offer rombox on as many configs as possible
11:07:08[IDC]Dragons/folk/fork
11:09:39[IDC]Dragondo we need some more field testing with gcc3.3.1?
11:10:32Bagderwe need to upgrade the rockbox.haxx.se's gcc installation to 3.3.1 from 3.0.4 to make smaller files
11:10:42[IDC]Dragonyes
11:11:20[IDC]Dragonand perhaps the build results could use some more testing, because of the compiler change?
11:11:27Bagdertrue
11:11:41Bagderjust in case
11:14:32[IDC]Dragonbtw, gcc3.3.1 now works on my box, too
11:14:45[IDC]Dragonthanks to BC's devkit
11:15:07Bagderbbl
11:15:09 Quit Bagder ("Leaving")
11:31:11 Quit NegativeK (Read error: 113 (No route to host))
11:59:08 Join dwihno_ [0] (~dw@81.8.224.89)
12:00
12:07:40 Quit dwihno (Read error: 60 (Operation timed out))
12:10:27 Join dwihno [0] (~dw@81.8.224.89)
12:17:21 Join Bagder [0] (~daniel@1-1-5-26a.hud.sth.bostream.se)
12:24:39 Quit dwihno_ (Read error: 110 (Connection timed out))
12:32:05lImbusBagder: can you tell me more about the rockbox-funds ?
12:32:31Bagderit is the donated money via the "donate" paypal button on the site
12:34:07lImbusmhmm. would like to donate a small bit too, but mind to open a paypal stuff for only that.
12:34:30BagderI don't think you need to do that
12:34:42Bagderyou can pay with a credit card if you wish
12:36:04lImbusmhmm. don't have one. that's the main issue I think.
12:36:32lImbusdoesn't it have a SWIFT and BIC like european international transfers now ?
12:36:46BagderI don't know, it might
12:41:28 Quit dwihno (Read error: 110 (Connection timed out))
12:44:01***Saving seen data "./dancer.seen"
12:44:31lImbuscan't find anything concerning that.
12:47:27Bagderme neither
12:53:17Lynx_|Imbus: are you german?
12:56:04Lynx_https://www.paypal.com/de/
12:59:18lImbusnot really. but that seems to be solution, as soon as I am willing to create another account for that. I simply don't want to have a special account on nearly every internet page
13:00
13:17:21 Quit Bagder ("Leaving")
13:42:10 Join dwihno [0] (~dw@81.8.224.89)
13:45:18 Join yousan [0] (~yousan@peta5.u-aizu.ac.jp)
13:45:47 Quit yousan (Client Quit)
13:53:41Zagorwhee, the iriver uses the isd-300 usb/ata bridge
13:54:01Zagorit's only renamed "cypress cy7c68310" since cypress bought isd
13:54:44kurzhaarrockerHow much is in the rockbox fund? Can we buy one too? :)
13:54:57Zagorwe will
13:55:17*kurzhaarrocker was thinking about buying cypress
13:55:25Zagorhaha ok
13:57:08 Join Bagder [0] (~daniel@1-1-5-26a.hud.sth.bostream.se)
13:58:32dwihnoCypress hill! \o/
13:59:23kurzhaarrockerOne Cypress should do. Or do we need a whole forest?
14:00
14:01:31LinusNwe just need some seeds
14:01:50LinusNand then we just wait a few generations :-)
14:02:28[IDC]Dragonif only buying a box would be enough...
14:02:46Zagor[IDC]Dragon: have you got the ondio yet?
14:02:59[IDC]DragonZagor bought an AV3x0 long time ago, still no Rockbox on it
14:03:14[IDC]DragonZagor: not yet :-(
14:03:29[IDC]Dragonlooks like it won't reach here today
14:04:33LinusNit took 2 days for my fm recorder to go from USA to Sweden, but it took almost a week to get it through customs...
14:05:00[IDC]Dragonno customs this time, it's from germany
14:05:18kurzhaarrockerYou should ship those things without good music on int, LinusN
14:05:27kurzhaarrocker-t
14:05:57LinusNit was new, with only those default french crap songs on it
14:06:15[IDC]Dragonrepairing mp3 players must be a good source of music
14:06:21LinusNindeed
14:06:30[IDC]Dragonmaybe I should open up my shop
14:06:42LinusNthe "send me your songs for analysis" scheme works good for me :-)
14:06:55[IDC]Dragon:-/
14:07:42kurzhaarrockerWith french songs its different. Quarantine. They must be relabledt to "Freedom Songs".
14:11:16Zagor[IDC]Dragon: i admit i bought the av a bit early :)
14:13:04Zagor[IDC]Dragon: if you like, the rockbox fund can sponsor the ondio
14:14:18[IDC]Dragonwhoa, what a nice fund
14:14:43[IDC]Dragonto be fair, let's wait awhile if this proves fruitful
14:14:55Zagorok
14:15:14[IDC]Dragon(before my wishes soar too high)
14:16:08[IDC]Dragonin case of success, we can direct new Ondio Rockbox user to the donate button
14:16:21[IDC]Dragons/user/users
14:16:31Zagorthere's more than one? ;)
14:16:41[IDC]Dragonnot yet ;-)
14:16:59[IDC]Dragonread: "Ondio Rockbox users"
14:21:24kurzhaarrockerI hope you are talking about the ondio recorder?
14:22:14[IDC]DragonI ordered a recorder, yes
14:22:45kurzhaarrockergood.
14:23:56[IDC]Dragonthat can be a nice thing about a flash recorder: no moving parts, no noise in the internal mic
14:24:41[IDC]Dragon1 GB MMC cards are less than 100 Euro nowadays, I was surprised
14:25:16[IDC]Dragon(don't tell me I can get a 40 GB disk for that)
14:25:30kurzhaarrockerwell Archos proved that it is possible to have unbearable noise even without moving parts. I hope they improved.
14:25:55[IDC]Dragonis it so bad?
14:26:30kurzhaarrockerRockbox improved it a lot.
14:26:42kurzhaarrockerIt was because of the crappy lcd programming.
14:26:48[IDC]Dragonwhen? with what hardware?
14:26:57kurzhaarrockerThe normal Recorder 6
14:27:32kurzhaarrockerWith the original archos hardware you'd hear and record every lcd refresh. With rockbox it was ok.
14:27:47kurzhaarrocker(still not the best recording device I ever had...)
14:28:00[IDC]Dragonalways, or gradually improving?
14:28:21kurzhaarrockerI don't remember any of that trouble with rockbox anywhen.
14:28:33dwihnocrappy lcd programming?
14:28:42[IDC]DragonI'm asking because our LCD access is a lot faster since video
14:28:54kurzhaarrockerdwihno not in rockbox of course :)
14:29:03[IDC]Dragonabout factor of 4
14:29:51 Join wildstoo [0] (~d512f815@labb.contactor.se)
14:29:53dwihnokurzhaarrocker: nah, rockbox lcd code is da shiznat! :)
14:29:55kurzhaarrockerI must admit that I probably haven't used recording with Rockbox after January 2004 because I use a special build with trigger included.
14:30:15LinusN[IDC]Dragon: that was one of the first things we noticed when we developed rockbox, that the noise was gone
14:30:53[IDC]Dragonnice. Have you ever held a scope to the LCD clock with Archos s/w?
14:31:00LinusNnope
14:31:02wildstoohi. i recently acquired an Archos Studio 20 with an 80GB HD upgrade.. problem is, WinXP will not format drives that are greater than 32GB as FAT32.
14:31:12LinusNwildstoo: tough luck :-)
14:31:29[IDC]Dragonuse Paragon or other tools to format
14:31:34Bagderwindows is a fine OS *really*
14:31:37LinusNi think some of the partitioning tools out there can help you with that
14:31:39Bagder;-)
14:31:40wildstooanyone know of any (free) tools that will allow me to format a partition on a USB drive as FAT32 under WinXP?
14:31:49dwihnowildstoo: get that software
14:31:54[IDC]DragonParagon did it for me
14:32:07LinusNi use paragon as well
14:32:10[IDC]Dragonthe free demo version was sufficient
14:32:33wildstoook i'll look at Paragon, thanks
14:33:04dwihnowhat was it
14:33:04dwihnoparagon partition something
14:33:04dwihnoit is neat
14:33:04DBUGEnqueued KICK dwihno
14:33:04dwihnoand you can do it with the demo ver
14:33:55wildstoo$50 :(
14:34:14[IDC]Dragonhttp://kb.paragon.ag/paragon/templ/169.jsp?catId=3283
14:34:15wildstoouh.. site says that demo version doesn't let you do any actual operations
14:34:29wildstoojust play with the UI and see how things MIGHT work if you registered
14:34:45 Quit wildstoo (Client Quit)
14:34:52 Join wildstoo [0] (~d512f815@labb.contactor.se)
14:35:04wildstoolol clicked a link and got navigated away
14:35:10wildstooshould have seen that coming really
14:35:10kurzhaarrockerDo you have knoppix?
14:35:27wildstoo(on CGI IRC, firewalled here)
14:35:38Zagortry ranish partition manager: http://www.ranish.com/part/ it's definitely free
14:35:45wildstoonope
14:35:49wildstoobut i could get it
14:41:20Bagdersyncing 20GB music takes a while
14:41:42kurzhaarrockerJust imagine you had to do it in real-time, Bagder
14:41:50Bagderhehe
14:42:03Lynx_wildstoo: are you sure windos xp can't do that?
14:42:16BagderLynx_: yes, many people have claimed this
14:43:11LinusNXP can do it, but it won't let you
14:43:17kurzhaarrockerLynx_ only Win95 / Win98 can do it. Upgrading sometimes is not an improvement.
14:44:02LinusNmicrosoft want to promote NTFS, so they limit the usage of FAT32
14:44:04***Saving seen data "./dancer.seen"
14:45:13Bagderrockbox is so lame that hasn't switched to ntfs yet
14:45:18wildstoowell, FAT32 is a bit crap in many ways
14:45:26 Part Zagor
14:45:44wildstooroll on WinFS ;)
14:46:18BagderI read they won't include that
14:46:31CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
14:46:31*kurzhaarrocker wants a portable mp3 player with 5 1/4" floppy.
14:47:04[IDC]Dragonmaybe you find an ATA floppy, just connect it then
14:47:42 Join dwihno_ [0] (~dw@81.8.224.89)
14:47:47 Quit dwihno (Read error: 60 (Operation timed out))
14:49:40wildstooit doesn't matter what drive label you give the drive, does it?
14:49:53wildstoodoesn't have to be called JUKEBOX or ARCHOS or something, does it?
14:49:54Bagdernope
14:49:55LinusNno
14:50:35[IDC]Dragonjust reading the Ondio manual, the firmware doesn't look bad
14:50:38[IDC]Dragonhttp://www.minidisc.org/manuals/archos/archos_ondio.pdf
14:50:51[IDC]Dragonmaybe it doesn't need Rockbox?
14:51:02wildstoo*sigh* 80GB HD over a USB connection = about 24 hours to format it
14:51:07Lynx_wait a minute, i just remember if formated my box from xp home two weeks ago...
14:51:27[IDC]Dragondo a quick format
14:51:39Lynx_sure does work for me, xp offers ntfs and fat32
14:51:49[IDC]Dragonelse this is no-go, tha batteries won't last 24 hours
14:51:54Lynx_maybe the drive has to be fat32 to begin with, though
14:52:05wildstooi have it connected to mains
14:52:23wildstooXP offers FAT32 if the drive is less than 32GB in size
14:52:25[IDC]Dragonstill, this supplies less power than the disk needs
14:52:29wildstoomine is 80GB
14:52:37Lynx_wildstoo: ah, that's it...
14:52:38[IDC]Dragonso you will drain the batteries
14:53:21wildstoowell i copied stuff off it for about 12 hours yesterday and it didn't die
14:53:24wildstooand it's still alive now
14:53:39[IDC]Dragonvery surprising...
14:54:01[IDC]Dragonwhat model?
14:54:08*kurzhaarrocker is glad that he has a 2 1/2" -> 3 1/2" adapter...
14:54:45wildstooum.. Archos Studio 20
14:55:12[IDC]Dragondo the studios have a stonger charger?
14:55:19wildstooif you mean the HD itself i have no idea of the model :)
14:55:22[IDC]Dragonstronger
14:55:51wildstoono idea.. it doesn't even show a battery symbol when it's connected to USB anyway
14:55:54wildstooso there's no way to tell
14:58:57wildstoogets pretty toasty when it's running all night connected to mains
14:59:07wildstoo:)
15:00
15:11:08 Quit Bagder ("Leaving")
15:11:32 Quit wildstoo ("CGI:IRC (EOF)")
15:13:34 Join paultje3181 [0] (~83d3e218@labb.contactor.se)
15:14:01paultje3181hi, how can I view JPG's using rockbox?
15:14:15paultje3181player version, today's build
15:14:48kurzhaarrockerThe player doesn't have a graphical lcd and can not display jpg's :(
15:14:55paultje3181pity
15:15:04paultje3181no bmp's as well?
15:15:17kurzhaarrockerNo graphix at all, only text.
15:15:21paultje3181damn
15:15:36paultje3181I should've bought the recorder :P
15:15:43kurzhaarrocker:)
15:16:15paultje3181I'll try to get text working then
15:16:33paultje3181haven't tested it yet, but rockbox sure is great
15:23:43paultje3181a smaller font isn't possible as well I guess?
15:24:20kurzhaarrockerYou guessed right
15:24:43[IDC]Dragonyou can move the box further away
15:24:48paultje3181:P
15:25:12paultje3181I meant to get more text on the display :D
15:25:52paultje3181what I pity, now I can get RSI from reading Harry Potter on my jukebox :p
15:26:30kurzhaarrockerlol!
15:26:48paultje3181well, I'm off reading :P
15:26:57paultje3181thnx for the info!
15:27:01lImbuslol @ [IDC]Dragon
15:27:02kurzhaarrockerno problem
15:27:38 Quit paultje3181 ("CGI:IRC (EOF)")
15:27:46[IDC]Dragonfunny guy
15:27:46 Nick dwihno_ is now known as dwihno (~dw@81.8.224.89)
15:28:04[IDC]Dragonjpeg on a player, rotfl
15:28:20kurzhaarrockerMaybe we should have advised him to read the faq instead of harry potter on the player...
15:28:55kurzhaarrockerWell with those user definable characters there in fact could be a way to visualize jpgs...
15:29:09[IDC]Dragoncertainly, and video
15:29:48[IDC]Dragonif you redefine them at exacly the time they are displayed
15:30:34kurzhaarrocker... and by synchronizes shaking you might even increase the display area...
15:30:46[IDC]Dragonah, yes
15:31:23[IDC]Dragonand by syncing to external RGB light pulsed you'll get color
15:31:34kurzhaarrockeryay! :D
15:31:58[IDC]Dragonmaybe 3D if you move it back and forth, too
15:32:27kurzhaarrockerNow let's find a hard drive that survives that treatment.
15:33:57kurzhaarrockerAfter watching "lord of the rings" with that method you don't need a fitness center any more.
15:35:18[IDC]Dragoni'm off
15:35:23 Part [IDC]Dragon
15:56:14 Part LinusN
15:56:24 Part kurzhaarrocker
16:00
16:14:33 Join edx [0] (edx@p548792DE.dip.t-dialin.net)
16:18:25 Join dwihno_ [0] (~dw@81.8.224.89)
16:30:57 Quit dwihno (Read error: 110 (Connection timed out))
16:44:06***Saving seen data "./dancer.seen"
16:54:24 Join methangas [0] (methangas@0x50a476c8.virnxx10.adsl-dhcp.tele.dk)
16:54:24 Quit lImbus (Read error: 54 (Connection reset by peer))
17:00
17:06:59 Nick dwihno_ is now known as dwihno (~dw@81.8.224.89)
17:16:35 Join dwihno_ [0] (~dw@81.8.224.89)
17:24:32 Join lImbus [0] (~manuel@kernel.cycos.net)
17:34:39 Quit dwihno (Read error: 110 (Connection timed out))
17:39:08 Join midk [0] (~midk@c-24-18-39-204.client.comcast.net)
17:40:30 Quit midk|sleep (Read error: 104 (Connection reset by peer))
17:42:44 Join dwihno__ [0] (~dw@81.8.224.89)
17:51:51 Join epictetus [0] (uifils@otto.microway.com)
17:54:15 Join dwihno [0] (~dw@81.8.224.89)
17:59:27 Quit hardeep ("[BX] Connection reset by panasync?")
18:00
18:04:38 Quit dwihno_ (Read error: 110 (Connection timed out))
18:10:43 Quit dwihno__ (Read error: 110 (Connection timed out))
18:12:17 Join mecraw [0] (~lmarlow@69.2.235.2)
18:32:22 Join bagawk [0] (Lee@ACC0E21B.ipt.aol.com)
18:35:23Lynx_i just notices something weird: I had the flashed box charging while off. then I turned it on without unplugging it. now if i press off again it displays "battery charging", but that does not go away, and if i do something like move the cursor it smears through the message.
18:35:37Lynx_like only the parts the cursor changes are refreshed.
18:35:56midkLynx_, this was recently fixed
18:36:03midkupgrade to the latest daily build
18:36:03Lynx_ok
18:36:15Lynx_thx
18:36:33midknp
18:39:44Lynx_hehe, the installer says "..while installing rockbox on your computer"
18:41:03bagawk:P
18:41:15bagawkthe installer is a little wierd
18:41:28midk:)
18:41:33bagawkit also adds an item to your add/remove programs list
18:41:34Lynx_didn't after upgrading, the box ask if it should rolo the new firmware?
18:41:47bagawkyes do that :)
18:41:54Lynx_hehe, didn't even see that. and, does removing work? ;-)
18:41:59Lynx_well, it's not asking
18:42:02bagawki have no idea
18:42:34midk?
18:42:47bagawki think i just did it, without my archos connected, and it just said it was removed, with absolutly nothing done
18:42:54midkjust rolo into the ajbrec.ajz
18:44:10***Saving seen data "./dancer.seen"
18:46:33Lynx_hmm, same with just unzipping it. maybe that feature was changed..
18:47:03midkit sometimes doesn't work
18:47:06midkjust do it manually
18:47:35Lynx_i did...
18:48:04midknow reflash.....
18:50:28 Join bagawk_ [0] (Lee@ACC06A2A.ipt.aol.com)
18:50:45Lynx_thx, but I know that much ;-)
18:51:13Lynx_with the video plugin, is there still some value one can adjust to remove flickering?
18:51:44midknope
18:51:57midklet it warm up and it doesn't flicker though
18:52:07midkjust a little band runs across the screen..
18:52:19 Quit bagawk (Read error: 104 (Connection reset by peer))
18:53:05 Nick bagawk_ is now known as bagawk (Lee@ACC06A2A.ipt.aol.com)
18:55:18bagawkmidk: not always for me
18:55:29bagawkmost of the time i see nothing
18:55:44bagawk:)
18:56:04Lynx_is there some site where a rvf stuff is collected?
18:56:15Lynx_like futurama and such...
18:56:17bagawkthere are a few on the rocbkx site
18:56:31bagawkhttp://rockbox.haxx.se/newvid
18:56:41bagawkthere are only a few good ones
18:56:41midknot really
18:56:57midkhttp://rockbox.haxx.se/newvid has a few things
18:57:06bagawkdoom is ok, futruama is great, baby, and a warm reception in LA suck
18:57:17bagawkand grantz graf is kinda cool sometimes
18:58:05Lynx_grantz graf?
18:58:21Lynx_ah...
18:58:43Lynx_and there is a wiki entry on the conversion tools
18:59:08 Join pike [0] (amiga@h234n1fls22o1064.bredband.comhem.se)
19:00
19:02:39Lynx_cu all
19:02:45 Quit Lynx_ (" Like VS.net's GUI? Then try HydraIRC -> http://www.hydrairc.com <-")
19:02:45pikeHello Felix, Markus, Lee and anyone else I couldn't find the real name for :)
19:08:52bagawk?
19:09:09bagawkdo i know you?
19:09:11bagawklol
19:09:47pikenah, I came here from your website
19:09:57pikedid a little reading before
19:10:00pike:p
19:10:28pikeso you guys involved in anyway with this? http://rockbox.haxx.se/twiki/bin/view/Main/IriverInfo
19:16:00pikejust found this... shame I dont speak japanese http://www.susono.com/~hiroya_d/iRiverJ/Font/
19:18:07methangasuse babelfish
19:18:19methangasif you wanna find out what it can do
19:18:33methangasallthough i don't really have a solution to the firmware :P
19:19:10pikesolution to what?
19:20:00methangasgetting it translated
19:21:01pikeaha.. I was just curious what it did. so anyone in here involved in getting rockbox onto ihpX series ? or it's still in the "need more info" stage ?
19:21:58lImbusi think they planned to buy a device
19:22:26lImbus'they' found out it's the same usb-bridge
19:22:27pikemost people in here swedes, or only the founders?
19:23:00pikethey just cut the price further on ihp120 afaik.
19:23:07pikethey = irivernordic
19:24:42lImbusa few of the founders/core developers are swedish. then there are germans, and a few from everywhere
19:24:42lImbuserr. I actually think all of the founders are swedish
19:25:12pikeso am I
19:25:37pikeH120: förr 3795kr - nu 3499kr and thats directly from them, I think it can probably be found for as little as 3000 sek
19:26:37pikethats the unit I have myself, but I'm no viz with electronics, just know how to handle a primitive soldering iron :p
19:28:55lImbustell it Zagor or Bagder. I dunno what they planned
19:29:11 Quit pike (Read error: 104 (Connection reset by peer))
19:29:58 Join pike [0] (~amiga@h234n1fls22o1064.bredband.comhem.se)
19:30:05lImbuswelcome back
19:30:14pikexchat crashed.. :p
19:30:53 Join radu [0] (~radu@face-00000002.user)
19:32:55pikeso when I read the documentation, the mp3 decoding hardware in the archos only supports mp3, this will probably not be an issue on iriver?
19:40:43lImbusyeah, you may be right. I know iriver can do other formats. so either the dsp does or they do in hardware (must have > 100 mHz)
19:40:54lImbuserr, in software, of course
19:41:13pikecpu/dsp faster than 100mhz?
19:42:06lImbuscpu
19:42:07pikeiriver ihpX handles wma, wav and mp3 afaik. no drm though
19:42:39pikedream would be lossless and aac support, hehe. but I'll say no more. too early to make wishes
19:42:57lImbusisn't wav lossless ?
19:43:20bagawklImbus: yes but not compressed
19:43:24pikeyeah, but I mean lossless compression, like flac, ape, you know
19:43:35lImbusah, of course
19:43:55lImbusif they already to decoding in cpu, then we could give it a try. as well as ogg
19:44:09lImbusogg depending on memory
19:44:32pikeog yeah, firgot vorbis
19:44:35pikeforgot*
19:44:44pikeit's already supported
19:44:51pikebut a real battery drainer
19:46:19bagawkbye
19:46:22bagawk: )
19:47:10 Quit bagawk ("umount /dev/brain")
19:47:14pikecrappy decoder for vorbis methinks
19:47:35pikedoes rockbox support cuesheets handling now?
19:53:03 Join hdn [0] (~TapeTape@dyn-83-157-167-218.ppp.tiscali.fr)
19:53:06hdnlut
19:54:07lImbuspike: what's a cuesheet ?
19:54:32pikeit's an invetion that CDRWIN makers created
19:55:05pikeit's a textfile basically desrcibing the layout for a cd, an instruction how it should be burned, with track separations etc
19:55:37pikepast years have been used for "mixed" audiocd's, to provide "seamless" playback
19:55:50pikewinamp supports it via a plugin
19:55:54pikefoobar has the support builtin
19:55:55midkif you rip a cd as one huge mp3 file a cuesheet specifies times for each track, and names i believe etc
19:56:04pike^yep
19:56:13midkat 3:17 it's track 2, 7:53 track 3 etc
19:57:04pikehttp://pike2k.mine.nu:8010/201-va_-_dont_look_now_mixed_by_way_out_west-boss.cue
19:57:36 Part hdn
19:57:48pikethat cuesheet is faulty
19:57:54pike... the ending shouldnt be there.
19:58:14pikemp3 scene has issues with them, or better yet, the tools they use
20:00
20:00:57lImbusok, I see. it takes consideration of pre and postroll
20:02:06pike"gaps" yeah
20:02:20pikenow most cuesheets dont include this info, but some do
20:09:50 Join webguest35 [0] (~0c01350a@labb.contactor.se)
20:10:00 Quit webguest35 (Client Quit)
20:10:20 Join webguest83 [0] (~0c01350a@labb.contactor.se)
20:10:38webguest83hello
20:11:54 Quit webguest83 (Client Quit)
20:44:13***Saving seen data "./dancer.seen"
21:00
21:03:44 Join Elemeno [0] (~4givn@dpc691997050.direcpc.com)
21:04:05Elemenohey, has anyone ever replaced the LCD screen on the FM recorder?
21:04:34midkyes.
21:04:57midkoh, right, you were trying to replace yours.
21:04:58midk:)
21:05:19Elemeno:-) Yep
21:05:29midkhaving trouble?
21:05:39ElemenoI received the new one today, put it in, and it works!
21:05:39ElemenoBut... I have a question
21:06:20ElemenoWhen I took the original one out, it seemed like it was glued or it somehow stuck to the MP3 player. so I'm wondering how I should reattach it
21:06:48midkyour new one is loose?
21:06:52midkdid you check the page then?
21:07:43Elemenowhat I mean is the original screen, it seemed like the entire back was somehow glued or something to the white square
21:08:00 Join _aLF [0] (Alexandre@mutualite-3-82-67-66-128.fbx.proxad.net)
21:08:01_aLFhi
21:08:07Elemenoand it's just a practical question, I guess, of how to keep the new screen in place. does that make any sense?
21:08:11Elemenohi
21:10:47Elemenomidk: Basically, I'm trying to figure out how to physically reattach the screen to the Archos once it's plugged in and working
21:11:25midki know what you mean
21:11:30midkdid you check the doc i gave you?
21:12:31ElemenoI think so. I know there was one page I visited that talked about replacing the LCD screen, but it was for a different model
21:13:58midkshould be the same method, as it's the same screen
21:14:11midkyou could just roll up a piece of tape and stick it back there :)
21:14:33Elemenoyup. I'm sure I could get creative. I was just wondering if there was "a way"
21:14:48ElemenoI'm pretty happy to just have my screen working again:-)
21:16:55Elemenonow I just have to fix the mic, but I wouldn't think that would be too bad
21:17:22midknot really a way :)
21:21:27 Quit Elemeno ()
21:52:14 Join scott666_ [0] (~scott666@c-24-245-58-48.mn.client2.attbi.com)
21:52:25 Part scott666_
21:57:51 Quit midk (Remote closed the connection)
22:00
22:31:52 Quit methangas (" The IRC Client of the Gods! -> http://www.hydrairc.com <- HydraIRC")
22:44:16***Saving seen data "./dancer.seen"
23:00
23:31:52 Quit lImbus (Remote closed the connection)
23:47:56 Quit Hadaka (Read error: 238 (Connection timed out))

Previous day | Next day