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 2004-09-14

00:02:19Zagorcrap, the x window system keyboard events work just like the archos remote: repeat is equal to press-release-press-release
00:03:41amiconnA press-release about Archos? Did I miss something ;)
00:10:03amiconn[IDC]Dragon: How many screws are there for holding the Ondio together? I see 3...
00:10:20[IDC]Dragonyes, 3
00:11:01[IDC]Dragonand little clips at the top, but those come of easily if you "fold" up the case
00:11:35amiconnYup. My display is not glued to the cpu...
00:12:30amiconnMuch easier to open than the jbr...
00:13:42[IDC]Dragonyet, this is nice
00:13:57amiconnThe Ondio sp has a different mas... MAS3539F B4
00:14:28[IDC]Dragonwhat's that?
00:15:52amiconnUnfortunately I can't take pictures atm, my AA NiMH cells are all discharged (except those in the jbr, but the jbr is busy)
00:16:34[IDC]Dragonhow does the daughterboard look like?
00:17:17amiconnLike yours, but with the tuner chip and the surrounding components not there
00:17:37[IDC]Dragonwith footprints?
00:18:00amiconnThere is even text "ONDIO-FM V3.2" in the upper left corner
00:18:12amiconnYes, with footprints
00:18:31[IDC]Dragonsame as mine then
00:18:51[IDC]Dragonwhat MAS is that? Can it do other tricks?
00:19:12Zagorafaict the 3507 is discontinued and is replaced with the 3539. it only does layer3 decoding.
00:19:29[IDC]Dragonany version on the main PCB (mine is v2.22)?
00:20:09[IDC]DragonI thought they have a decode-only one with AAC, too
00:20:35Zagoryes, it's the 3559 for aac-only or 3529 for mp3+aac
00:21:09amiconn[IDC]Dragon: V2.22 ONDIO, yes
00:21:30[IDC]Dragonso the MAS is pin compatible
00:21:58[IDC]Dragonempty pads where the mic was?
00:22:15*Zagor can't find a X programming channel :(
00:22:55amiconnMAS 35x9F family: MPEG layer 2+3/ AAC/ G.729 decoder family. The 3539F can only do MPEG L2+3
00:23:40amiconn[IDC]Dragon: Yes, 2 empty holes where the mic was
00:25:10[IDC]DragonI can send you a MAS, if you're keen on recording
00:27:03amiconnThe challenge will be (after having a working mmc driver) to get this mas playing music.
00:27:29amiconnFor what I can tell from a quick glance at the datasheet, the setup procedure is quite different
00:27:39[IDC]Dragonmaybe it's the same, just without the recording application.
00:30:07 Quit Lucretia_ (Connection timed out)
00:30:14amiconnThe mas 35x9 does have a/d converters, as well as spdif in, and these are connected to the dsp core. So if someone writes some mas code...
00:30:34 Join Lucretia_ [0] (
00:30:55*Zagor goes to sleep
00:30:56 Quit Zagor ("Client exiting")
00:31:37amiconnIt also features AVC and MDB (although they call it "Micronas Bass" (MB) now
00:33:54amiconnAh. It seems that the standard decoder features are equivalent to those of the 3587. There is a new mode, "MP3 block input mode", for performance optimization
00:37:05 Quit Lucretia_ ("Leaving")
00:38:57*[IDC]Dragon also has to sleep
00:39:09 Quit [IDC]Dragon ()
00:41:38 Quit SmoothOperator ("CGI:IRC (EOF)")
00:47:35***Saving seen data "./dancer.seen"
00:49:42amiconnLinusN: Now I recorded a full album via SPDIF. Not a single bad frame...
00:50:37amiconnMy first test snippet (with deliberate disturbing of SPDIF transfer by shorting it) does have 11 bad frames though
00:51:17amiconnPlus: The time display of recordings done via SPDIF is rather odd
00:52:12amiconnLike it shows playtime 0:40 of a total of 0:29
00:53:05amiconnMaybe this is because the mas stops recording if there is no SPDIF data stream, but rockbox continues counting
00:53:56amiconnI stopped the recording of the album several minutes past the end of the CD
01:02:17 Quit ripnetUK ()
01:25:25 Part LinusN
01:31:16 Quit epictetus (Remote closed the connection)
01:31:18 Join epictetus [0] (
01:39:45 Join gromit`` [0] (
01:43:39 Quit GhUl ("Leaving")
02:03:48 Quit LePoulpe303 (Read error: 60 (Operation timed out))
02:35:16 Part amiconn
02:47:36***Saving seen data "./dancer.seen"
03:01:42 Quit mecraw_ ("Trillian (")
03:20:23 Join midk [0] (
03:47:26 Join pjf [0] (
03:49:05 Join plok [0] (
03:54:23*plok is away - Automatically set away. - messages will be saved.
04:11:44 Quit DMJC (Remote closed the connection)
04:43:56 Join zorks [0] (
04:44:18 Quit plok ("I'm outta here!")
04:45:14 Quit zorks (Client Quit)
04:45:26 Join zorks [0] (
04:47:40***Saving seen data "./dancer.seen"
04:49:23 Quit zorks (Client Quit)
05:26:30 Join ashridah [0] (
05:28:18 Join maikeul [0] (
05:50:39 Quit gromit`` (Read error: 110 (Connection timed out))
06:05:14 Join Elemeno [0] (
06:05:18 Quit scott666 ("i'll be back...eventually...")
06:06:02ElemenoI'm having a weird problem with my Archos FM Recorder. Sometimes it will turn on just fine; other times when I press "On" the green LED will flicker rapidly a few times and nothing will turn on. Anybody ever experience this?
06:28:17 Quit Elemeno ()
06:46:05 Join _Headie [0] (
06:46:06 Quit Headie (Read error: 104 (Connection reset by peer))
06:47:43***Saving seen data "./dancer.seen"
06:54:27 Quit midk ("just STOP it arspy")
07:08:40 Join LinusN [0] (
07:10:06 Join plok [0] (
07:15:09*plok is away - Automatically set away. - messages will be saved.
08:07:56 Join moormaster [0] (
08:23:00 Quit moormaster (Remote closed the connection)
08:26:39 Join webguest13 [0] (
08:27:42 Quit webguest13 (Client Quit)
08:36:34 Join Zagor [242] (
08:47:44***Saving seen data "./dancer.seen"
09:24:13 Join amiconn [0] (
09:27:11amiconnhi all
09:53:34 Join Lynx_ [0] (lynx@
10:09:58 Quit plok ("I'm outta here!")
10:22:03 Join pillo [0] (
10:26:23 Join [IDC]Dragon [0] (
10:26:41[IDC]Dragonhi guys!
10:28:17[IDC]Dragonamiconn: r u there?
10:36:20[IDC]Dragonamiconn: I tries to run the (patched) SP firmware on my FM, no go
10:36:39[IDC]Dragonit got stuck during the boot
10:37:33[IDC]Dragonnd I found a difference in my and Archos USB mode:
10:38:02[IDC]Dragonthe MMC clock is about 3 times slower with Rockbox
10:38:23[IDC]Dragonhave no idea yet how this could be influenced
10:38:55[IDC]Dragontoday I will take that daughterboard off, to better reach the alive bridge
10:47:48***Saving seen data "./dancer.seen"
10:50:26 Quit ashridah ("out")
12:00:09amiconn[IDC]Dragon: Now I'm around-ish
12:05:29[IDC]Dragontop board is off now
12:05:40[IDC]Dragonbut I'm heading for lunch
12:19:02 Join ashridah [0] (
12:28:40 Quit Ka_ (
12:35:53NJoinKa_ [0] (
12:47:50***Saving seen data "./dancer.seen"
12:54:41 Join AciD [0] (
12:55:53 Join webmind_ [0] (
12:56:18 Join klr [0] (
12:57:49 Quit webmind (Read error: 104 (Connection reset by peer))
13:01:00 Quit webmind_ (Remote closed the connection)
13:02:59 Join webmind [0] (
13:06:21 Quit pjf ("Leaving")
13:06:32 Join webmind_ [0] (
13:06:42 Quit webmind (Remote closed the connection)
13:11:29 Quit klr ("CGI:IRC")
13:23:38[IDC]Dragonback again
13:26:45amiconn[IDC]Dragon: The hang on boot with the "wrong" firmware could be caused by the different mas.
13:27:30amiconnI2c addresses and register numbers are equal, but some bits present in one mas are not present in the other
13:28:06amiconnPlus, the dsp core config memory cells are mostly the same, but at different addresses.
13:34:03 Nick _Headie is now known as Headie (
13:34:11Zagorit's interesting that they chose to use a new chip. there must be a substantial price difference for that to pay off.
13:34:38 Join benschi [0] (
13:36:16[IDC]Dragonamiconn: yes, I intended that as a MAS compatibility test
13:36:34[IDC]Dragonyou could try the countercheck
13:37:36amiconnI'll try that, in the evening.
13:38:40amiconnBtw, your measurement of the mmc clock from the sh confirms my statement that the serial can do 1.5 MBit/s maximum @12 MHz CPU clock
13:40:14amiconn3 MBit/s would also be possible, but the data sheet says "no continuous data transfer possible" at that speed
13:41:14[IDC]Dragonwe can try
13:41:16Zagorweird wording. what is continous data transfer? 100KB? 1MB? 10MB?
13:41:29[IDC]Dragonit doesn't need to be continuous
13:41:37Zagorthat's my point
13:41:43[IDC]DragonZagor: back to back, I'd say
13:42:02Zagortwo sectors?
13:42:08LinusNtwo bytes?
13:42:15LinusNtwo bits?
13:42:32[IDC]Dragonbytes are integer, i'd say
13:42:53[IDC]Dragonbut there may be gaps for the DMA ot whatever
13:43:06LinusNstill, how long is a "continuous" transfer?
13:43:08[IDC]Dragonagreed, the phrase is fuzzy
13:43:27[IDC]Dragonwe drive the clock, so the SH governs that
13:43:38amiconnWe can try if the 3 MBit/s setting gives an advantage.
13:43:41[IDC]Dragonit's not like we could be flooded
13:44:08[IDC]DragonLinusN: how fast id the MAS transfer?
13:44:17amiconnFor card init, we have to stay below 400 KBit/s (from mmc speds)
13:45:07amiconn[IDC]Dragon: Mas transfer should be 750 kBit/s for 3507d and 1MBit/s for the 3587f. Rockbox uses 750kBit/s for both
13:45:49[IDC]Dragonfile I/O is a lot faster now
13:46:18[IDC]Dragonwas it like 8MBit/s?
13:49:03amiconnata i/o is around 4 MByte/s with optimized routines
13:49:36amiconnSo mmc will be a lot slower
13:49:54[IDC]Dragonimpressive 32MBit/s then
13:50:10[IDC]Dragonan order of magnitude below
13:50:32[IDC]Dragonwe'll have a hard time with the wav playback ;-)
13:51:11[IDC]Dragonbut at least we're not worried about spinups
13:51:41[IDC]DragonArchos also seems to have the buffering code still in place
13:53:05[IDC]DragonI measured phases of activity at the song begin, then silence on the MMC
13:53:07 Quit benschi ("CGI:IRC")
13:54:32[IDC]DragonLinusN: have you read that we need anothe bitswap? :-(
13:55:29dwihnoLuckily we have Jens to optimize the living crap out of the swap ;)
13:57:02LinusN[IDC]Dragon: yes
13:57:54Zagorondio rockbox won't be a speed monster...
13:58:29[IDC]Dragonthe SH-2 has an option for MSB first in its SCI
13:58:54Zagorthe grass is always greener on the other side :)
13:59:28[IDC]Dragonmaybe the SH-1 has an undocumented bit at that place? ;-)
14:01:13Zagorthat would be ironic
14:23:43[IDC]Dragonthe SH-2 has an extra register behind the RDR register, the only taken bit is bit3, controlling the direction
14:24:13[IDC]Dragonit's called SDCR, serial direction control register
14:24:38[IDC]Dragonin our SH-1, these addressed are unused
14:24:50[IDC]Dragonhmm, maybe we should try?
14:25:37Zagorcan't hurt
14:31:47[IDC]Dragonjust for a note: set bit 3 at address 0x05FFFEC6 to 1 and see if mp3 still plays
14:32:35Zagorthe irc log is a nice memo pad :)
14:37:13amiconn[IDC]Dragon: [13:51:43] <[IDC]Dragon> Archos also seems to have the buffering code still in place
14:37:49amiconnThe archos guys seem to be rather lazy when it comes to updating software
14:38:38amiconnE.g the recording transfer loops are identical in both the jbr and the ondio fmr firmwares
14:40:31amiconnThe recording transfer loop seems to be present in the ondio sp firmware too; I wonder why
14:42:25[IDC]Dragontransfer loop?
14:42:40amiconnReading the mas parallel out
14:42:59[IDC]Dragonwhy shoult it differ?
14:43:13amiconnBecause the ondio sp has no recording?
14:43:23[IDC]DragonSP, ah, ok
14:43:39amiconnBtw, the 8-bit data buffer for mas parallel access is present on the sp board!
14:44:17[IDC]Dragonhmm, what else can we read from the MAS?
14:44:54[IDC]Dragonhow is the direction pin(#1) wired?
14:45:17amiconnDidn't measure, and can't do so atm. The ondio is @home.
14:45:40[IDC]Dragonno seti@home?
14:46:50amiconnNo ;)
14:47:53***Saving seen data "./dancer.seen"
14:52:53amiconnBtw: How did you power the ondio for measuring levels? If you use the usb cable, it goes straight into usb mode...
14:53:41 Join midk [0] (
14:56:06[IDC]Dragonwith test clips and a benchtop power supply
14:57:22[IDC]Dragonit's current limiter also helps to reduce the risk of damage when accidently shorting something
15:16:25 Quit midk (Remote closed the connection)
15:27:54 Join midk [0] (
15:33:43 Join ripnetUK [0] (
15:34:15 Quit pike (Read error: 54 (Connection reset by peer))
15:34:28 Join pike [0] (
15:47:52Zagorthe "symbolic buttons" model is working nicely. i've added it to browser, wps and menus now
15:49:46Zagorthe code is actually simpler than before
15:50:16LinusNeven better
15:50:18amiconnZagor: Nice. Is it smaller in binary form too?
15:51:44Zagoryeah, a few hundred bytse
15:51:58LinusNme likes it
15:55:03dwihnosymbolic buttons?
15:55:24Zagoryes, using WPS_FFWD instead of BUTTON_RIGHT|BUTTON_REPEAT in wps.c for instance
15:56:29Zagornothing major, just avoids some #ifdefs (but actually adding a few as well)
15:56:48LinusNgotta go, cu guys
15:56:51 Part LinusN
15:56:58Zagorhowever it makes it easier to get an overview when adding new keypads to the brew
15:57:12dwihnoZagor: ah, okay. Smart.
15:59:03Zagorsome things, like the quickscreens and the pitch screen, are still hardcoded to BUTTON_ macros since they are so specific to a certain model
15:59:24Zagori need to test this on a player though. the sim is not quite good enough.
16:01:11amiconnZagor: Perhaps the quickscreens should be surrounded with #ifdef BUTTON_F2 etc. to only include them if the corresponding button is present
16:01:39Zagoryes, they are
16:05:26 Quit midk (Read error: 104 (Connection reset by peer))
16:06:38 Join midk [0] (
16:24:22 Join methangas [0] (
16:33:07 Quit lImbus (Remote closed the connection)
16:39:48 Join mecraw_ [0] (~lmarlow@
16:47:54***Saving seen data "./dancer.seen"
16:50:16 Join iRiverRockBox [0] (
16:52:25iRiverRockBoxHi, I am interested in the iRiver RockBox porting project.
16:52:38iRiverRockBoxI don't know much about the Archos RockBox, but I was wondering have well the current RockBox handles "gapless" playback.
16:54:09dwihnoif the mpegs are encoded in a good fashion
16:54:26dwihnoI use −−nogap when I encode using LAME - great result!
16:54:49Zagorwe feed the data to the decoder gapless, so it the data is gapless so will the playback be.
16:55:01Zagorbut we don't "massage" the data in any way
16:58:12ZagoriRiverRockBox: are you interested in helping with development?
16:58:39iRiverRockBoxI was just wondering if the iRiver playback engine could use a hybrid hardware/software decoding of the mp3s, by software decoding the last frame and removing the silence part/first of the frames in a song. All other frames could be decoded by hardware.
16:59:07Zagorthe iriver does not have any decoding hardware, so we have to do everything in software there
16:59:40ashridahiRiverRockBox: iriver's latest firmware already deletes silence inside audio files, according to the release notes
16:59:50Zagorand yes, this improves our possibilities of doing various "tricks" do avoid gaps
16:59:55ashridahiRiverRockBox: iriver's firmware lacks song prebuffering, which is where the gapless issue remains
17:00:20Zagorashridah: how do you mean lacks song prebuffering?
17:00:41Zagordoesn't load and play at the same time?
17:00:50ashridahit can load and play at the same time
17:00:55ashridahit just doesnt' do it for the NEXT song
17:01:19iRiverRockBoxYeah, correct prebuffering is of course necessary... but with mp3 their is still the "inside a frame" silence problem.
17:01:47iRiverRockBoxmp3 has fixed frame size... ogg doesn't.
17:01:58ashridahiRiverRockBox: oggs don't have the fixed-frame-size problem, and the gap is still present for oggs
17:02:05ZagoriRiverRockBox: correct, but with software decoding we can cut silence in the decoded pcm data
17:02:06*ashridah has a vast majority of oggs in his collection
17:02:47ashridahiRiverRockBox: but yes, again, the latest firmware already cuts out silence inside the files, according to the release notes for it (i really should get around to testing that, i know a few songs with mid-song gaps, which supposedly it also deletes)
17:03:04iRiverRockBoxI actually thought that the Motoral MFC5249 offered Mp3 hardware decoding.... I may be wrong though :)
17:03:21Zagorit doesn't
17:04:10Zagorashridah: sounds odd to remove midsong gaps. how would it know what gaps to remove? some tracks contain silence on purpose.
17:04:49ashridahlike i say, i haven't tested it.
17:04:54ashridahbut yes, it does sound odd
17:05:19ashridahnot impossible to conceive if it does threshold testing at anything but the start/end, but i doubt they went to that much trouble
17:06:01ashridahiRiverRockBox: it's not really particularly economical to cram an mp3 decoder onto a general purpose cpu.
17:06:52ashridahiRiverRockBox: you'd be far better served by cramming a crapload of runtime-configurable FPGA.
17:07:01*ashridah fondly remembers the ATMEL chips he briefly played with
17:09:33 Join Bagder [0] (
17:10:26Zagorbagder do you have any idea where to look/ask about X keyboard programming? none of the channels here are very interested in programming
17:10:50Bagdertried checking for a related newsgroup?
17:11:12ashridahZagor: heh. catching raw X11 events eh?
17:12:06Zagorashridah: yes, but the key events are lousy. i don't get true press/release events. instead I get repeated press/release/press/release/press/release when a key is held down
17:12:50ashridahZagor: that is strange. xev doesn't seem to suggest that behavior
17:13:19Zagordoes for me
17:13:20ashridahi've got the xlib programming manual, if you don't have it already (i'm not entirely certain it's a legal copy, i forget where i picked it up)
17:13:56ashridahoh, hangon, keyboard.
17:14:00*ashridah smacks head
17:17:49Zagorit even does that for shift :(
17:19:23*ashridah isn't anything remotely related to an expert X11 coder
17:19:46ashridahZagor: like i say, i've got an x11 programming manual in html if you're interested tho.
17:20:44Zagorthanks, but i've found lots of them on the web
17:21:06Zagorhaven't found the answer to my question though :( maybe there is no way
17:21:08ashridahfair enough.
17:21:26ashridahit's probably why most people avoid x11 apis like the plague
17:22:12Zagorwell nothing on top of x11 will ever be able to give me anything different unless there is a mode setting
17:22:29Zagorand I do want X11. network transparency is king.
17:23:02ashridahdepends how far the abstraction is taken.
17:23:07Zagorif I could just ask "is key X pressed or unpressed right now"
17:23:14ashridahproblem is, further in you go, less likely it is to suit your needs
17:23:14iRiverRockBoxOK, thanks for clearing up my mp3 decoding questions. I found websites which state that the MFC5249 is designed for mp3 decoding, but which does not mean that it has a dedicated hardware for decoding. Thanks :)
17:23:42BagderZagor: x11 is event-based, I don't think you can do that ever unless you keep the state info yourself
17:24:13ZagorBagder: you're right. but i'd like to get the proper key events, not fake repeats :(
17:24:27BagderI understand
17:24:28 Join mecraw [0] (~lmarlow@
17:24:34ashridahyou can't set the keyboard to a raw mode or something?
17:24:37ashridahor perhaps use DGA?
17:24:44 Quit mecraw_ (Read error: 54 (Connection reset by peer))
17:24:47ashridah(goodbye network transparency tho, i bet)
17:25:12Zagorand since I do quite a lot of stuff remotely I'd hate to lose it
17:26:43iRiverRockBoxI have another question about the random function of RockBox - I shortly looked at the code. Which input parameters does it utilize? the clock etc.? and where is the random seed saved? iRiver's original firmware developers have major problems with generating a working random function.
17:27:24BagderiRiverRockBox: we just seed it with the time
17:27:28Bagderthen we keep the seed
17:28:12Zagorthus we can resume a shuffled playlist in the same order
17:28:31Bagderand that's why we call it shuffle, not random. We truly shuffle ;-)
17:28:32ashridahBagder: that's a problem then. the iriver devices don't have a built in persistent clock
17:28:42iRiverRockBoxI wonder why is so hard for the iRiver developers? :) They have had more than a year to fix it.
17:28:49BagderI doubt that is a problem
17:28:59Zagoryou don't need it persistent, you can just use current elapsed ticks
17:29:11Zagorhumans are fuzzy and won't react the same tick every time
17:29:14ashridahBagder: well, no, of course not, but it lends yourself to being boxed into a particular range.
17:29:30ashridahbut yeah, at 150MHz or something for the ihp's, i doubt that's an issue :)
17:30:00BagderI think lack of clues is the true reason
17:30:03ZagoriRiverRockBox: i don't think it's hard at all for them. they are probably just busy with other things.
17:30:18Bagderthere's no money in fixing that
17:31:21ripnetUKhave we found a coldfire disassembler yet? i managed to decrypt the firmware, and a strings suggests it decrypted ok... did we ever find out if it runs from rom or if its copied to ram first? (iriver)
17:31:44BagderI good guess is that it copies it to ram
17:32:03ashridahdamnit. 1:30am. i've gotta crash
17:32:05 Quit ashridah ("sleep")
17:32:09iRiverRockBoxThey don't care about the consumers. The H3xx has the same problem and it is brand new. It was scheduled and they did implement it.
17:32:25iRiverRockBox It was scheduled and they did NOT implement it.
17:33:45*Bagder rembers Archos' bold statements about working on .wav playback
17:33:51Zagorwell ask ipod owners how long they have waited for gapless playback...
17:35:30Bagderhehe, ipod owners don't care for features
17:35:38Bagderits made by Apple, its enough
17:36:12Zagorwhat formats does the iriver record in by the way?
17:36:21iRiverRockBoxwav and mp3
17:36:52iRiverRockBoxwav 44khz - mp3 32kpbs-320kpbs..
17:37:10Zagorcbr or vbr?
17:37:23iRiverRockBoxCBR... I don't think it is LAME encoding..
17:37:35Zagorhehe, definitely not
17:40:01Zagorfinding a free integer mp3 encoder might be a problem
17:40:34iRiverRockBoxI think FLAC encoding could be really cool :)
17:41:01Zagoryes, but as I understand it flac compression is pretty hard on the cpu
17:41:11Zagordecoding is pretty fast though
17:42:16ripnetUKi dont understand - the wiki says that the iRiver has a UDA1380TT - Philips Stereo audio coder-decoder for MD, CD and MP3
17:42:19iRiverRockBoxYeah, you might unfortunately be right :) Playback should be no problem though... digital out + FLAC rocks!
17:42:22ripnetUKbut you say its all software?
17:42:53ZagorripnetUK: yes, that chip name is misleading. it's basically a D/A converter
17:43:06ripnetUKi see :)
17:43:43Zagorit's just designed for MD, CD and MP3 players
17:44:20iRiverRockBoxyeah, just like MFC5249 is "designed" for MP3 :)
17:44:33Zagoryeah, something like that
17:46:44 Quit [IDC]Dragon ("CGI:IRC")
17:48:20Zagorflac-compressing a 780MB wav CD took 13 minutes on a PII-333 according to the flac homepage. using default parameters.
17:49:16Zagorit's difficult to judge the relative speeds of the p2 and our coldfire, but by those numbers real-time flac compression doesn't sound impossible
17:50:06Zagoroh, i forgot. the encoder uses floating point...
17:50:18iRiverRockBoxdamn :)
17:55:40iRiverRockBoxWhere did you find this floating point encoding info - I must be blind :)
17:56:25Zagorthe features page. "Decoding requires only integer arithmetic".
18:02:28iRiverRockBoxAre you sure about this floating point encoding - I am still in doubt... it is not a good reference but check
18:05:38Zagor# egrep "(double|float)" flac-1.1.0/src/libFLAC/* | wc -l
18:05:43Zagor 39
18:05:58Zagorlooks like it :)
18:06:27iRiverRockBoxYeah :)
18:07:49iRiverRockBoxDoes any other lossless formats support integer arithmetic encoding? still open source?
18:07:59Zagori don't know
18:08:56Zagorrealtime compression has a pretty low priority, i must say
18:10:34iRiverRockBoxyeah, there is still a long way to go also :)
18:13:51Zagorhmm, why is the player simulator much slower than the recorder sim?
18:23:52*Zagor just found XAutoRepeatOff()
18:24:14Zagorbut it is global for the entire display...
18:25:16 Join webguest19 [0] (
18:27:09 Quit webguest19 (Client Quit)
18:45:01 Quit iRiverRockBox ("CGI:IRC (EOF)")
18:47:57***Saving seen data "./dancer.seen"
18:49:07 Quit Lynx_ (" HydraIRC -> <- :P")
18:50:24 Part pillo
18:53:18 Quit midk (Read error: 104 (Connection reset by peer))
19:08:31 Part Zagor
19:09:24 Join midk [0] (
19:24:40 Join PaulS [0] (
19:25:11 Join R3nTiL [0] (
19:25:56PaulSI saw some conversation earlier about the iRiver remote. You have my official permission to traverse the link I put on the Wiki to iRiverLounge. If you're uncomfortable with that I can post new stuff on the Wiki from my notes.
19:29:05PaulS(I suppose I can only give you permission to read _my_ posts on the lounge ... maybe. :-)
19:36:28 Part PaulS
19:37:55 Join Zagor [0] (
19:51:42 Quit ripnetUK ()
19:58:06 Quit R3nTiL ()
20:27:16 Join mecraw_ [0] (~lmarlow@
20:27:17 Quit mecraw (Read error: 104 (Connection reset by peer))
20:47:59***Saving seen data "./dancer.seen"
21:07:35 Quit epictetus (Ping timeout: 14400 seconds)
21:20:23 Join zeekoe [0] (
22:03:15 Join scott666 [0] (
22:06:25 Join scott666_ [0] (
22:16:36 Join [IDC]Dragon [0] (
22:16:52amiconnhi Jörg
22:17:19amiconnFound my wiki additions?
22:17:26[IDC]Dragonremoving the tuner board was very useful
22:17:39[IDC]Dragonnot yet, /me looks
22:19:11[IDC]Dragoncute, our twins
22:20:28[IDC]Dragonmine also has that roughly shaped plastic, but it stayed at the LCD
22:20:44[IDC]Dragoncame off the CPU first
22:22:08amiconnMy tuner board looks rather funny. All naked
22:22:23[IDC]DragonI found a collision when checking the signals at the bridge
22:22:48[IDC]Dragonyea, I wonder why they did it that way
22:23:13[IDC]Dragonnot swapping the jacks, making the board obsolete for the SP
22:23:26[IDC]Dragonjust for the power button?
22:23:34amiconnThe on/off button is also mounted on the tuner board
22:24:00 Join CGI808 [0] (
22:24:11 Quit CGI808 (Client Quit)
22:24:12 Quit scott666 (Read error: 110 (Connection timed out))
22:24:20amiconnMy RAM is rather sparely marked
22:24:48amiconnAnd the chip under the mmc slot cover is the flash ROM I guess?
22:25:04[IDC]Dragonas long as it works... we know what it is.
22:25:30[IDC]Dragonalso a strange layout, so far from the CPU
22:25:46[IDC]Dragonthey needed to route the bus all the way there
22:27:21amiconnOkay. Did you also notice my addition to the "new components" table? I'd like to check if the LT3440 and the NC7S08 are also on my board. I hope you can tell me where I should look, as I guess these are marked with a short code only
22:29:24[IDC]Dragonthey're there
22:29:58[IDC]DragonLTNP, in the power section
22:29:58amiconnOkay, I'll adapt the table
22:30:41[IDC]Dragon7S08 close to the 12 MHz osc.
22:33:02amiconnTell me about the usb bridge measurements
22:33:22[IDC]Dragonwell, I found two colliding outputs
22:33:49[IDC]DragonI need to make TxD1 an input while in USB mode
22:34:44[IDC]Dragonthis is visible on the scope, when the level goes only halfway
22:35:19amiconnyup, I understand
22:37:09amiconnBtw, I do now understand how the software power off is done on the player & recorder v1: The off button is also connected to a port pin, which is GP in normally. For power off it is set to GP out & low, simulating a keypress
22:37:36[IDC]Dragonah, ok
22:38:38amiconnDid you already try making TxD an input for usb mode?
22:39:09[IDC]Dragondoing that now
22:41:04 Quit methangas (" HydraIRC -> <- *I* use it, so it must be good!")
22:48:02***Saving seen data "./dancer.seen"
23:04:57[IDC]Dragonhmm, still not working.
23:05:25[IDC]Dragonthe collision is gone, but I measure different frequencies
23:09:25 Quit zeekoe ("bed time")
23:10:24amiconn[IDC]Dragon: In usb mode?
23:12:13 Join webguest76 [0] (
23:12:50webguest76rockboxx is awesome!! you guys are cool
23:12:51[IDC]Dragonthe Rockbox power-down timer is active in USB mode, too :-(
23:13:30Zagor[IDC]Dragon: really? that's a bad bug.
23:14:19amiconnAny chance to get Linus here? I have a very interesting question...
23:15:30Zagorhe usually pops in a little after midnight
23:16:13webguest76is there any way to have custom bitmaps for the screens without compiling the source myself
23:16:23Zagorwebguest76: no
23:16:24webguest76or if it isn't is there any freeware compilers?
23:16:28[IDC]DragonZagor: I menat the Ondio, don't worry
23:16:30amiconnCurrently analyzing the archos recording (yes I'm desperately hunting the recording issues)
23:16:39Zagorwebguest76: yes, gcc is free. it's the one we use.
23:17:14webguest76is that for linux?
23:17:22Bagderyes, linux too
23:17:29Zagorwebguest76: it's available for both windows and linux. see here:
23:17:43webguest76cool, thanks!
23:24:35webguest76if you use windows for compiling, what IDE do you use, if any?
23:25:05webguest76so do you use cywin
23:25:08amiconn[IDC]Dragon, Zagor, Bagder: The archos recording isr does something that the mas datasheet doesn't tell us...
23:25:42Zagorwebguest76: most core developers use linux
23:25:43webguest76wait nevermind
23:26:01Bagder... and the rest run make in cygwin
23:26:02Zagoramiconn: any idea what it is?
23:26:07[IDC]Dragonamiconn: aha?
23:26:25amiconnYes: it checks PB14 (IRQ6) in 2 places
23:26:52amiconnLinus' logic analyzer will most likely prove _very_ helpful...
23:27:29amiconnIt also uses a timeout on checking PRTW in case the mas doesn't want to tell us something
23:28:40amiconnIt has no 30 byte transfer limit
23:31:40amiconnThe PB14 check seems to be an end-of-data check
23:35:52 Join DigitalDiva [0] (
23:40:37 Join _aLF [0] (
23:41:39[IDC]Dragonamiconn: interesting twists with the recording
23:42:17amiconnI think this will result in an all-new recording routine, being rock-solid, as the name Rockbox implies...
23:42:53 Quit webguest76 ("CGI:IRC (EOF)")
23:44:10amiconn[IDC]Dragon: Btw, I did a long-term recording via spdif, and wanted to compare how the archos fw performs in this case. It didn't let me do this... "recording prohibited"
23:44:46Zagorright, you need to unlock it. can't remember how.
23:45:28amiconnYou'd need a copy protection bit stripper
23:45:43DigitalDivai know how u do it in the Archos fw
23:45:54Zagorno, the firmware allows recording if you enter a specific artist name or some such
23:46:50DigitalDivato disable the protectoin, go to the Archos fw and input HFSCMSOFF to disable the protection
23:47:20amiconnDigitalDiva: Ahaa! Where? Artists? Title?
23:47:40*amiconn tries that
23:47:42DigitalDivain the recording mode
23:47:52DigitalDivatype HFSCMSOFF as the ID3 tag
23:54:00amiconnGot it. You have to type that as the track title, then it tells you that SCMS is off. The track title does not change.
23:54:31amiconnHowever, now it says "Press play to start". No matter how long or hard I press, the recording doesn't start
23:56:10amiconnProbably a bug
23:57:29amiconnIndeed a bug - now I managed to start the recording by shortly removing the spdif source signal (!)
23:57:56DigitalDivaI wonder why Archos no longer provide spdif on their newer devices

Previous day | Next day