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 2013-06-17

00:29:33fs-bluebotBuild Server message: New build round started. Revision dafc359, 217 builds, 21 clients.
00:33:09 Quit bertrik (Remote host closed the connection)
00:35:21fs-bluebotBuild Server message: Build round completed after 349 seconds.
00:35:22fs-bluebotBuild Server message: Revision dafc359 result: All green
00:35:36 Quit ender` (Quit: The curious thing about .Net is that it allows you to use any language you want, as long as it's C#.)
00:35:58pamaury\o/ it compiles
00:45:18 Join Scromple [0] (~Simon@
01:22:00 Quit kaputnik_ (Ping timeout: 245 seconds)
01:24:28 Quit pamaury (Ping timeout: 256 seconds)
01:25:39 Quit TheLemonMan (Ping timeout: 268 seconds)
01:40:30***Saving seen data "./dancer.seen"
01:43:00 Quit lebellium (Quit: ChatZilla 0.9.90 [Firefox 22.0/20130612084701])
02:01:16 Quit liar (Quit: huiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii)
02:02:57 Quit dfkt (Read error: Connection reset by peer)
02:03:11 Join dfkt [0] (OxO29A@unaffiliated/dfkt)
02:20:58 Quit soap (Remote host closed the connection)
03:13:44 Join soap [0] (~soap@rockbox/staff/soap)
03:15:06 Quit saratoga (Quit: Page closed)
03:25:41 Quit krabador (Quit: Bah...)
03:39:40 Join notben [0] (
03:40:32***Saving seen data "./dancer.seen"
04:06:27Ceninant500mb in 3 that slow for USB 2.0?
04:08:00 Join ikeboy [0] (
04:25:46 Quit guymann (Ping timeout: 245 seconds)
04:26:38 Join guymann [0] (~c@unaffiliated/guymann)
04:27:44 Join amayer [0] (
04:28:24ikeboyCeninant:It's slow for fullspeed usb 2.0, but probably not for most targets. 500MB is about 3MB/s, normal for mp3 player flash memory
04:29:01ikeboyOn fullspeed usb 2.0 you'll get 60MB/s=10gb in3 min. What device are you using?
04:41:43 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:41:43 Quit amiconn (Disconnected by services)
04:41:44 Quit pixelma (Disconnected by services)
04:41:44 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:41:46 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:41:47 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:47:56 Join TheSphinX_ [0] (
04:49:41 Quit TheSphinX^ (Read error: Operation timed out)
04:50:14 Join traps [0] (~dewlap@2001:5c0:1400:a::223)
04:53:41 Quit kilroy (Ping timeout: 245 seconds)
05:01:32 Quit ikeboy (Quit: CGI:IRC (EOF))
05:18:22 Quit alexbobp (Quit: leaving)
05:18:31 Join alexbobp [0] (
05:31:07 Quit TheSeven (Disconnected by services)
05:31:17 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
05:40:36***Saving seen data "./dancer.seen"
05:51:23 Quit amayer (Quit: Leaving)
06:01:54 Quit dfkt (Disconnected by services)
06:01:56 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
06:18:50 Quit notben (Quit: leaving)
06:43:31 Join kevku [0] (~kevku@2001:470:27:773:0:feed:c0f:fee)
06:56:53 Quit [Saint] (Remote host closed the connection)
06:58:07 Join [Saint] [0] (~saint@rockbox/user/saint)
07:12:11[Saint]Here's a question: If precut is applied, should we not alter the volume level displayed to the user?
07:12:33[Saint]should the volume level displayed not be "volume level + precut"?
07:14:22[Saint]for example, if you want line level, but have a preset with positive gain and precut applied, or are using precut as a volume cap.
07:19:21[Saint]It seems confusing, and non-obvious to me, that if I should apply a precut of -12dB (to cap the volume at line level on the Classic to avoid clipping when Ms [Saint] uses it without looking or allows dumbass workmates to use it even if they are looking) that the numeric volume displayed isn't inclusive of the precut value. it. "full volume" is still +12dB, even though it is *actually* 0dB.
07:19:47[Saint]s/value. it./value. ie./
07:40:40***Saving seen data "./dancer.seen"
07:46:34 Quit amiconn (Remote host closed the connection)
07:46:34 Quit pixelma (Remote host closed the connection)
07:46:48 Join melmothX [0] (~melmoth@unaffiliated/melmothx)
07:47:27 Join pixelma [0] (pixelma@rockbox/staff/pixelma)
07:47:27 Join amiconn [0] (amiconn@rockbox/developer/amiconn)
08:13:40 Join ender` [0] (
08:16:35 Quit kevku (Ping timeout: 245 seconds)
08:37:10 Join LinusN [0] (
08:47:22 Join liar [0] (
08:58:12 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
09:05:25 Join Zagor [0] (
09:05:25 Quit Zagor (Changing host)
09:05:25 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
09:07:46 Join kevku [0] (~kevku@2a01:d0:ffff:34a::8:3)
09:12:44 Join mortalis [0] (~kvirc@
09:20:59 Quit GodEater (Changing host)
09:20:59 Join GodEater [0] (~whoknows@rockbox/staff/GodEater)
09:32:58 Join einhirn [0] (
09:34:24 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:40:44***Saving seen data "./dancer.seen"
09:54:16 Part LinusN
10:06:44 Join mt [0] (~quassel@rockbox/developer/mt)
10:09:43 Join wodz [0] (
10:10:56*gevaerts grumbles
10:11:08gevaertsFull speed usb == 12 megabit/second
10:13:27 Join mt` [0] (~quassel@rockbox/developer/mt)
10:13:32copper08:03:49 UTC <ikeboy> On fullspeed usb 2.0 you'll get 60MB/s=10gb in3 min. What device are you using?
10:13:51copperI've never seen a USB2 device read or write at more than 30, maybe 35 MB/s
10:13:56gevaertsThat too
10:14:24copperand I've bought some of those "ultra fast" USB flash drives
10:14:29gevaertsA full speed device still won't get more than about a megabyte :)
10:14:41copper08:03:49 UTC <n1s> yeah, there's no portable encoder though
10:15:04copperTAK works with wine, and caudec (my command line transcoder) supports it too
10:15:13coppershould work on OS X too
10:15:28copperit's perfectly usable on Linux
10:15:44coppernow what's needed is playback support, really
10:16:04copperI consider that this is no longer a chicken and egg type of problem
10:16:16 Quit mt (Ping timeout: 240 seconds)
10:17:00[Saint]I don't think the fact that an encoder exists was in question. I understood what was wanted was a portable encoder, for use as a recording option.
10:19:56copperalso, I shouldn't copy paste timestamps when they're wrong :P
10:25:37coppertranscoding flac to tak
10:28:30Zagoryuk, closed source
10:29:45wodzZagor: have you had a chance to look why build system reports 'all green' ?
10:30:00Zagorno :-(
10:36:00 Join kaputnik_ [0] (
10:44:59 Join kaitsu1 [0] (
10:45:49 Part kaitsu1
10:46:36 Quit pamaury (Ping timeout: 256 seconds)
10:50:17copperTAK playback support in Rockbox would be pretty epic
10:50:28copperand maybe give TAK a boost in popularity
10:50:46ZagorI don't think TAK deserves a boost until they open the source
10:50:51coppernow now
10:51:01copperdon't be an extremist!
10:51:39copperanyway, does it not depend solely on a Rockbox developer wanting to do it?
10:51:39Zagoruh, I'm not. closed source is the extreme view.
10:52:00copperor will such commits be rejected?
10:52:10Zagorcopper: yes it does. and I don't oppose it. I just personally don't think they deserve it yet.
10:52:25copperok then
10:53:29copperfrom what I remember, the reverse engineered decoder in TAK is pretty short (in number of lines)
10:53:57copperthe reverse engineered TAK decoder in ffmpeg
10:54:32copperand I assume tagging is not an issue since Rockbox already supports APEv2
10:55:10copperlet's hope that saratoga will be able to do it
10:55:16Zagorperhaps I'm out of tune, but are people really prepared to optimise their entire music collection for rockbox? seeing that no other devices can play TAK even if we add it.
10:55:31copperyou can't think like that
10:55:48coppereveryone just needs to implement format X in their own software
10:55:51ZagorI just don't understand what it is that you like so much.
10:56:00Zagorwell they can't, since it's locked down...
10:56:12copperthey can now that there is a decoder
10:56:20copperopen source decoder*
10:56:33Zagorthere is only a reverse-engineered decoder. it could break anytime.
10:56:53copperit's been more or less approved by TAK's author, and he even offered to help
10:57:07ZagorI would personally really think twice before many hours adding support for a format that obviously don't want us to support it
10:57:18copperhe does
10:57:29copperhe's just very protective of his baby
10:57:32Zagorthen he should release it
10:57:46Zagorit will NEVER take off as closed source
10:57:50copperit's been debated to death (with him) on HA, not gonna happen any time soon, unfortunately
10:58:10Zagorthat's his choice. and my choice is to therefore ignore him.
10:58:12copperbut he does have backwards compatibility in mind
10:58:21wodzThats braindamaged thinking, sorry.
10:58:47copperhe's actually a very nice guy
10:59:09ZagorI bet. but his code is 100% uninteresting to me in closed form.
10:59:10copperI don't agree with his decisions either
10:59:29wodzI am against investing limited developers time into some esoteric format without source. But if someone wants to - it is his choice
10:59:48copperfair enough
11:13:10 Join LinusN [0] (
11:17:23[Saint]I don't actually see what it would bring to the table. To be honest.
11:17:45[Saint]A lossless format with /slightly/ better compression, and worse performance than anything else we support.
11:17:50[Saint]...yay? :)
11:18:14[Saint]Add to that the fact that it is never going to be adopted widely, and, I think its largely wasted effort even discussing it.
11:20:10coppernot so slightly
11:20:15copperalso, why worse performance?
11:20:54[Saint]My understanding is that it is more CPU intensive than any of the other supported lossless formats.
11:21:25copperwhat? no
11:21:27coppernot at all
11:21:37copperTAK decoding is more or less on par with FLAC decoding
11:21:45coppercertainly nothing like APE decoding
11:22:18copperthere's a reason that TAK has been hailed as the (technically) king of lossless formats
11:22:19 Join lebellium [0] (
11:22:50[Saint]Hahahaha...that's fairly rich.
11:23:25[Saint]Ok, I was mistaken about the performance, it seems...but "king of lossless" is a bit rich.
11:24:02[Saint]"King of lossless formats adopted by 7 people", perhaps :)
11:25:15copperthat's why I said technically
11:25:57[Saint]The compression doesn't seem /that/ much better...unless there has been some drastic improvements lately.
11:26:06[Saint]~4% or so
11:26:57gevaerts[Saint]: can't be worse performance than ape -5000
11:27:06[Saint]Well, this is true.
11:27:19 Quit mt` (Read error: Connection reset by peer)
11:27:30gevaertsBeing worse than that is actually mathematically proven to be impossible I believe :)
11:27:52 Join mt [0] (~quassel@rockbox/developer/mt)
11:28:46Zagorperformance would very likely improve if he released it
11:28:47[Saint]TAKp5 appears to be roughly twice as costly on CPU time as flac8
11:29:06copperyou mean p4
11:29:27[Saint]I actually don't, but, OK.
11:29:42copperthere is no p5
11:30:44copperand actually TAK -p4 is faster than FLAC -8
11:30:52[Saint]Ah. Well. I guess I shouldn't be too surprised sourcing this from HA... ;)
11:31:21copperthe max setting is -p4m
11:31:38[Saint]This chart lists p4m, perhaps a typo.
11:31:52copperp4m is correct
11:32:01[Saint]now *that* was a typo :)
11:32:05[Saint]*5m, even.
11:32:10copperFLAC -8: 222x
11:32:15copperTAK -p4: 328x
11:32:21copperTAK -p4m: 148x
11:32:40[Saint]flac1.2.1 -0 8.801.938.201 67,99% 233 315x 383 192x 180 409x 234 314x
11:32:40[Saint]flac1.2.1 -5 8.113.749.200 62,68% 453 162x 584 126x 194 379x 257 285x
11:32:40[Saint]flac1.2.1 -8 8.071.892.842 62,35% 1596 46x 1668 44x 204 359x 262 280x
11:32:40DBUGEnqueued KICK [Saint]
11:32:40[Saint] TAKp0 62,04% 262 280x 408 180x 251 293x 290 253x
11:32:40[Saint] TAKp2 7.821.652.842 60,42% 397 185x 542 135x 285 257x 311 236x
11:32:40***Alert Mode level 1
11:32:40[Saint] TAKp5m 7.722.405.330 59,65% 3097 24x 3149 23x 347 212x 380 193x
11:32:46copperbut the 'm' parameter is not very useful
11:32:58copperthe gains are very marginal
11:33:19[Saint]those are: Compressed Compression PT PRate GT GRate PT PRate GT GRate
11:34:27CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
11:34:27*gevaerts recommends #tak
11:34:47[Saint]Reasonably old, but, I sincerely doubt the performance has changed much in ...well, ever.
11:35:03[Saint]Nor do I have any reason to believe it will.
11:36:28[Saint]copper: are you *sure* 4m is the highest?
11:36:47[Saint]I find many, many, many references to p5{e|m}
11:36:53copper-p# select encoder preset #: 0-4 (fastest to strongest, default is 2).
11:37:00 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:37:01coppermaybe in an old version
11:38:50 Join krabador [0] (~krabador@unaffiliated/krabador)
11:38:56[Saint]Anyhoo, indeed. If someone wants to implement this, in a sensible fashion, I'm certain no one will stand in their way. But I doubt it would help to bolster the popularity of either project.
11:40:46***Saving seen data "./dancer.seen"
11:42:41***Alert Mode OFF
11:55:05 Part LinusN
12:04:14 Quit krabador (Quit: Bah...)
12:08:38 Quit kevku (Ping timeout: 260 seconds)
12:32:55 Join kevku [0] (~kevku@2a01:d0:ffff:34a::8:3)
12:36:49 Quit Kohlrabi (Read error: Operation timed out)
12:37:56 Join Kohlrabi [0] (
12:41:53 Quit [Saint] (Remote host closed the connection)
12:43:00 Join [Saint] [0] (~saint@rockbox/user/saint)
12:43:24 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
12:45:43 Quit fs-bluebot (Ping timeout: 252 seconds)
12:46:29 Quit bluebrother^ (Ping timeout: 276 seconds)
12:47:01 Join fs-bluebot [0] (
12:47:35 Quit mt (Ping timeout: 256 seconds)
12:48:00 Join mt [0] (~quassel@rockbox/developer/mt)
12:50:54copper<saratoga_> TAK is such a dead-simple but also really efficient lossless format we should think about porting the ffmpeg decoder
12:51:14copperit seems though that the ffmpeg decoder is a lot slower than the reference decoder (Takc.exe)
12:52:20copperThe difference between the official FLAC and TAK decoders on a 74 min file is small (7.94s vs 9.96s), but with ffmpeg, it is rather large (9.3s vs 15.3s) (FLAC and TAK respectively)
12:52:26 Join kingratface [0] (
12:52:38kingratfaceHey all
12:52:49kingratfaceI'm having a problem with my sansa clip+
12:53:21funmancopper: it would be nice to profile ffmpeg and see where the slow-down is. maybe that part is already rewritten efficiently in rockbox?
12:53:46copperI would have no idea on how to do that
12:54:00kingratfaceI get a sort of pause and shakey noise whenever I try play stuff
12:54:18copperkingratface: what format?
12:54:35kingratfacemp3 atm, lemme try a flac file
12:55:19kingratfaceokay, it's only mp3
12:55:52kingratfaceroughly half the audio is replaced by silence
12:55:54coppermaybe your MP3 file is wonky. Did you encode it yourself?
12:56:12kingratfaceI've tried with multiple MP3s
12:56:16funmancopper: on Windows, I have no idea
12:56:16kingratfaceand they work on the firmware
12:56:29copperfunman: I run linux!
12:56:32kingratfacethe default firmware, I mean
12:57:00funmancopper: then you can use perf
12:58:06copperwhat package is that part of?
12:58:10kingratfaceIt could be a hardware thing
12:58:21kingratfacethe battery has been crappy recently
12:58:25copperkingratface: you didn't answer my question :)
12:58:43funmancopper: linux-tools
12:58:48kingratfaceNo, I didn't encode them myself
12:58:58kingratfaceThey're all lame V0s I think
13:00:00kingratfaceI've tried using a dev build, still doesn't work
13:00:17 Join krabador [0] (~krabador@unaffiliated/krabador)
13:01:13 Join linuxstb [0] (~linuxstb@unaffiliated/linuxstb)
13:01:23kingratfaceI can play the same file fine on the sandisk firmware, so surely it's a problem with RB
13:01:41copperfunman: got it, what perf command should I run?
13:01:49 Join Derp [0] (
13:01:59kingratfaceI've googled around a bit but can't find any fixes, someone seemed to be having a similar problem in 2011 but never got anywhere with it
13:05:11funmancopper: no idea :)
13:05:51copperyou want "trace"?
13:06:10 Quit mt (Ping timeout: 245 seconds)
13:06:11funmani want nothing, just giving you suggestions
13:06:27funmanprofiling which functions take the most CPU time can be interesting
13:06:43coppermeh, I have no idea how to use it
13:07:54 Join mt [0] (~quassel@rockbox/developer/mt)
13:08:36kingratfaceI'm at work atm so have don't have long, does anyone have any idea how I could fix this? Otherwise I guess I'll just use the sandisk firmware for a while
13:09:12 Quit mt (Read error: Connection reset by peer)
13:09:23funmanwithout the file hard to tell
13:09:43kingratfaceWithout what file?
13:09:54copperthe wonky MP3 file
13:09:59 Part linuxstb ("Leaving")
13:10:02kingratfaceThere is no wonky MP3 file
13:10:08kingratfaceI've tried with multiple MP3s
13:10:16kingratfaceand they work fine on other audio players
13:10:17coppernone of which you encoded yourself
13:10:20copperand got from shady sources
13:10:32 Join mt [0] (~quassel@rockbox/developer/mt)
13:10:39copperencode one yourself with LAME and try that
13:10:51kingratfaceThey're all decent What.CD rips
13:11:04coppermany users are clueless
13:11:26kingratfacedude, the encodes are fine
13:11:35coppercan you upload one of those MP3s online? Dropbox or whatever
13:11:48copperone that you have tested and that produces the problem
13:12:02kingratfacetake any MP3 from there
13:12:20copperplease tell me which one specifically that produces the problem
13:12:33copperone that you have actually tried
13:12:39kingratfaceI've tried with a ton of MP3s
13:12:41kingratfacebut fine, wait
13:13:04funmankingratface: well i tried a ton of MP3s as well and they worked for me
13:13:19kingratfaceI've been using rockbox for over a year on this thing
13:13:29kingratfaceMy MP3s have worked fine until now
13:14:01funmanafter a rockbox upgrade?
13:14:52kingratfaceI switched to a dev build after I started getting the problem to see if that'd fix it
13:15:05kingratfacebut up until then I hadn't changed the install in months
13:15:19 Quit kugel (Remote host closed the connection)
13:15:25 Join kugel [0] (~kugel@
13:15:25 Quit kugel (Changing host)
13:15:25 Join kugel [0] (~kugel@rockbox/developer/kugel)
13:16:54copperkingratface: playing that file right now on my Clip+
13:17:06kingratfacedoes it work?
13:17:10copperfor now
13:17:14kingratfaceokay, so great
13:17:15copperI'll let it play until the end
13:17:20kingratfaceit's not a problem with the file then
13:17:27copperfile looks fine
13:17:47copperfoobar2000 says it's fine too
13:18:10copperI'm using the latest Rockbox dev build for the Clip+
13:18:47kingratfaceso I guess the hardware must be fucked up somehow
13:19:44copperupgrade to the latest dev build just in case
13:20:53kingratfaceI already have
13:21:06copperok it played back fine until the end
13:21:30 Join mt` [0] (~quassel@rockbox/developer/mt)
13:21:56kingratfaceit's running dafc359-130616 atm
13:24:42 Quit mt (Ping timeout: 260 seconds)
13:25:30kingratfaceI reset the settings and it's still screwed up
13:25:36kingratfaceanyway, I've really gotta run
13:25:41kingratfacethanks for the help
13:25:50kingratfaceI guess I'll just get a replacement
13:25:57kingratfacethe battery is this one is crap anyway
13:25:59kingratfacecheers folks
13:26:03 Quit kingratface (Quit: leaving)
13:39:29fs-bluebotBuild Server message: New build round started. Revision c1eafa1, 217 builds, 20 clients.
13:40:48*[Saint] cannot think of a hardware problem that only affects mp3
13:40:50***Saving seen data "./dancer.seen"
13:41:13[Saint]The fact the issue spontaneously occurred is also suspicious.
13:41:23 Join gapan [0] (
13:41:28[Saint]Perhaps SanDisk is more lenient with fs corruption then we?
13:41:58gevaertsAlmost certainly
13:42:09gevaertsOur FAT implementation isn't very robust
13:42:27[Saint]I was going to ask him to check his filesystem, but I came in late.
13:42:41[Saint]I wouldn't be surprised if it was trashed.
13:43:01*pamaury wonders if we should warn the user when the FS is found to be corrupted
13:43:25[Saint]I brought that up once upon a time myself.
13:43:36[Saint]I think it would be a good idea. But no one thought of a sane way to do so.
13:44:25[Saint]Can someone give me some input on my question earlier regarding volume level and precut?
13:44:46[Saint]should we, or should we not, adjust the level displayed to include the precut level?
13:45:29pamauryyeah, online checking of file system consistency is not exactly easy if you don't have a lot of memory to waste. There are some obvious cases you could catch but it would need a lot of experiment to check if this is useful at all
13:45:57[Saint][17:19:22] <[Saint]> It seems confusing, and non-obvious to me, that if I should apply a precut of -12dB (to cap the volume at line level on the Classic to avoid clipping when Ms [Saint] uses it without looking or allows dumbass workmates to use it even if they are looking) that the numeric volume displayed isn't inclusive of the precut value. it. "full volume" is still +12dB, even though it is *actually* 0dB.
13:45:57fs-bluebotBuild Server message: Build round completed after 388 seconds.
13:45:58fs-bluebotBuild Server message: Revision c1eafa1 result: All green
13:48:45[Saint]I had a look, and doing "volume+precut" when precut and EQ are enabled seems trivial.
13:49:10[Saint]The question I'm wanting to know the answer to, is "IS this a sane thing to do?"
13:56:29 Join linuxstb [0] (~linuxstb@unaffiliated/linuxstb)
14:17:48 Quit wodz (Quit: Leaving)
14:21:21 Join linuxstb_ [0] (~linuxstb@unaffiliated/linuxstb)
14:21:32 Join Cen1nant [0] (
14:21:59 Join amayer [0] (
14:22:02 Join Derpina [0] (
14:23:00 Quit linuxstb (Disconnected by services)
14:23:04 Nick linuxstb_ is now known as linuxstb (~linuxstb@unaffiliated/linuxstb)
14:23:12 Quit Derp (Read error: Connection reset by peer)
14:23:33kugelpamaury: how did you find the key(s)?
14:23:44pamaurybrute force
14:24:04 Quit JdGordon (Ping timeout: 264 seconds)
14:24:05kugelsuch a huge key?
14:24:40 Quit Ceninant (Ping timeout: 264 seconds)
14:24:45pamauryyeah, it looks huge but it is generated from a much smaller one in reality. It's all implemented into imxtools/sbtools/sbtoelf if you want to see the details
14:25:03kugelhow large is the real key?
14:25:20 Join JdGordon [0] (~jonno@
14:25:21 Quit JdGordon (Changing host)
14:25:21 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
14:26:14pamaurylet me check...
14:26:26funman for(int i = 0; i < 0x10000; i++)
14:26:30funman16 bits?
14:27:35pamaury48 bits
14:27:52pamaurylaserfuse[0] and lower 16-bits of laserfuse[1]
14:28:11pamaurybut in the two cases where I already brute forced the key, laserfuse[0] was 0
14:28:28pamauryso it was really a 16-bit search...
14:30:08pamauryactually this is not exact: on device, the three "laserfuse" are used but what really matters is laserfuse[0] | laserfuse[1] so having two fuses for it is just security illusion
14:30:53pamaurysame for laserfuse[2]: the whole content matters but you quickly realise it only has 65536 interesting values, they really screw up
14:34:57funmanpamaury: how much time do you think you spent on this?
14:35:23funmanso we can estimate a bruteforce speed in bits per hour (for the first key you found because i guess next ones are 'free')
14:35:59pamauryon what exactly ?
14:36:50 Join mt [0] (~quassel@rockbox/developer/mt)
14:37:03funmanon finding the key?
14:37:20pamaurywith the tool it's instant
14:37:29funmanwell on writing the tools if you prefer
14:38:51 Quit mt` (Ping timeout: 245 seconds)
14:38:56pamauryah...hum, much longer ;) I spent many hours reversing the stmp tool to understand the format, then I had to spent like 10 hours of brute force to find the first key, then I disassembled the ROM and then I realise the key was smaller than expected
14:50:59pamauryI should time my reverse engineering sessions, that would depressing though
14:51:39funmanpamaury: well for me too, that's why i do not count
14:51:57funmani guess it's still less than 2^1024 seconds though
14:54:29Derpinahey saint, im back! how else can i speak to you other than on IRC? i dont think i'll have IRC when booting into linux and H need instructions from you on the steps. Googletalk/whatsapp/facebook/ whatever?
14:59:39 Join Derp [0] (
14:59:41 Quit Derpina (Read error: Connection reset by peer)
15:25:42 Quit mt (Read error: Connection reset by peer)
15:25:45 Quit GeekShadow (Quit: leaving)
15:26:19 Join GeekShadow [0] (
15:26:19 Quit GeekShadow (Changing host)
15:26:19 Join GeekShadow [0] (~antoine@reactos/tester/GeekShadow)
15:27:12 Join mt [0] (~quassel@rockbox/developer/mt)
15:29:20Derpis anyone here at all?
15:31:10 Quit kevku (Ping timeout: 245 seconds)
15:37:39 Join mt` [0] (~quassel@rockbox/developer/mt)
15:39:23 Quit mt (Ping timeout: 276 seconds)
15:40:54***Saving seen data "./dancer.seen"
15:44:35 Quit mortalis (Quit: KVIrc 4.3.1 Aria
15:46:44 Quit mt` (Read error: Connection reset by peer)
15:50:38 Join Derpina [0] (
15:50:39 Quit Derp (Read error: Connection reset by peer)
15:51:10 Join mt [0] (~quassel@rockbox/developer/mt)
15:53:10 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
15:55:12 Quit Derpina (Read error: Connection reset by peer)
15:55:23 Join Derp [0] (
15:59:26 Quit mt (Ping timeout: 240 seconds)
16:08:11 Join mt [0] (~quassel@rockbox/developer/mt)
16:11:18 Join wodz [0] (
16:11:44wodz,43157.msg219567/topicseen.html#msg219567 <- Thats quite interesting. Maybe worth to include in official build?
16:16:37 Quit wodz (Quit: Leaving)
16:22:54 Quit krabador (Quit: Bah...)
16:25:39 Join madcat1990 [0] (~madcat199@
16:27:06 Quit Derp ()
16:52:13 Join ikeboy [0] (
16:53:34ikeboyFor the record the usb 2.0 spec allows up to 480MBit/s which is 60MB/s
16:53:51 Join kevku [0] (~kevku@2001:470:27:773:0:feed:c0f:fee)
16:55:23copperikeboy: I don't know the specifics, but does that account for overhead?
16:55:46copperand again, I've never seen any real world device go that fast, including "super fast" expensive USB flash drives
16:56:17copperikeboy: is that protocol speed or data speed?
16:57:30ikeboyJust because no one has made a "fullspeed" usb 2.0 doesn't mean that isn't fullspeed
16:57:52copperwhy were you talking about this again?
16:58:26ikeboysomeone asked whether 500MB in 3 min was slow for usb 2.0
16:59:06copperyup, pretty slow for USB2, but normal for internal flash and microsdhc cards
16:59:47copperikeboy: from your own wikipedia link: "USB 2.0: Released in April 2000. Added higher maximum signaling rate of 480 Mbit/s (effective throughput up to 35 MB/s or 280 Mbit/s)"
16:59:56copper"effective throughput up to 35 MB/s"
17:00:06copperwhich is precisely what I said
17:00:28copperand again it's not "full speed" it's "high speed"
17:01:14copperso yeah 480 Mbps is the protocol speed, not the actual data throughput
17:01:23copperseems like there's quite a bit of overhead
17:01:36 Join mt` [0] (~quassel@rockbox/developer/mt)
17:01:40gevaertsikeboy: 60MB/s raw bitrate is *far* from actual speed
17:02:36gevaertsThere are actually some tables in the USB spec about theoretically achievable packet rates IIRC, but that's still before UMS
17:03:19 Quit Zagor (Quit: Clint excited)
17:03:31madcat1990Too bad we don't have the documentation for every device, it'd make USB So much better..
17:03:58madcat1990Though, considering, the only problem I've ever had with RockBox's USB on my Mini2G was bad cables that made it act weird
17:04:16 Quit mt (Ping timeout: 256 seconds)
17:04:24copperI'm waiting for more testing before I can report that I fixed my Fuze+ USB problems
17:04:36gevaertsSomeone worked this out. High speed USB gives you a maximum of 53248000 bytes per second, but that's before you've counted the SCSI layer
17:05:14gevaertsSo yes, I'd expect that 35 MB/s to be a more or less realistic maximum
17:05:48gevaerts(which is still twice as fast as the fastest disk I've ever heard of in a rockbox context)
17:08:32coppermy iPod Classic does 22MB/s
17:09:50gevaertsNot bad!
17:10:46ikeboy says Theoretical Bandwidth iafter overhead 48MB/s but tested is 25/35
17:11:53copperI wonder how fast the flash-based iPods go
17:13:40ikeboy seems to be their fastest usb 2.0 drive benchmarked at 30MB/s
17:14:41copperikeboy: well let me try my USB 3 HDD on a USB2 port
17:15:06gevaertscopper: there's some old data on
17:15:14copperit reads at over 50MB/s on USB3
17:15:50gevaertsOn a typical DAP you're not going to get anywhere near those speeds
17:16:17gevaertsEspecially flash ones. They just don't *need* very fast flash, and faster == more expensive
17:16:24pamauryI think 48MB/s is theoretical raw performance, but on top of that you need to add UMS and SCSI which are not exactly light protocol. I would be interesting to do the exactly calculation
17:16:42copper536870912 bytes (537 MB) copied, 15.772 s, 34.0 MB/s
17:18:40 Quit lebellium (Remote host closed the connection)
17:18:40gevaertspamaury: I think the actual protocol overhead strictly speaking isn't too bad. A header every 64K, and one packet for the request, so let's say two 512 byte packets extra per 64K, i.e. about 3%
17:19:09gevaertsWhat kills you is lack of queueing
17:19:33 Join lebellium [0] (
17:19:47gevaertsYou can work around that up to a point with clever caching, but it's not magic
17:20:06pamauryyeah by protocol I mean: first send all data, then do all read/write, then send back status. So indeed so queuing means you cannot interleave as much as possible
17:20:40gevaertsI wouldn't be surprised if you could approach those 53248000 bytes per second using several devices
17:21:00 Join n1s [0] (~n1s@rockbox/developer/n1s)
17:21:27pamauryyeah, let's plug 10 hard disk on the same link at next devcon !! :)
17:22:16gevaertsFor PP502x (the one I know most about), you won't get more than about 15MB/s whatever you do. That's what I could get using throwaway transfers, where you just have to reply and not care about the data at all
17:24:51gevaertsBy the way, if you want to test pure USB performance, don't use the ramdisk code in usb_storage.c. That's slower than a hard drive using DMA on some devices :)
17:29:14ikeboyThe person who asked the original question didn't mention he was using a rockbox target or even a mp3 player:) for all we know, he could have had a usb 2.0 hard drive
17:29:35ikeboyCeninant500mb in 3 that slow for USB 2.0?
17:31:02ikeboyAnd when i said fullspeed I didn't mean the technical definition I mean the fastest theoretically possible
17:31:48copperdon't use "full speed" in the context of USB, since that's an actual standard name
17:32:01copperUSB names suck, I'll give you that
17:32:08gevaertsThey could have done better
17:32:15copper"full speed", "high speed", "super speed"
17:32:22gevaertsAlso, don't call it "USB 2.0 speed"
17:32:31copperah? why not?
17:32:34gevaertsBecause low speed and full speed are valid USB 2.0 speeds
17:32:55copperwhat's next though
17:32:57gevaertsUSB 3 is different for that, they refer to the USB 2 spec for anything non-superspeed
17:32:59copper"mega speed"? :P
17:33:12gevaertsSo you *can* say "USB 3" meaning superspeefd
17:35:54 Join mt [0] (~quassel@rockbox/developer/mt)
17:39:15 Quit ikeboy (Quit: CGI:IRC)
17:39:38 Quit mt` (Ping timeout: 276 seconds)
17:40:58***Saving seen data "./dancer.seen"
17:55:52 Quit mt (Read error: Connection reset by peer)
17:57:14 Join mt [0] (~quassel@rockbox/developer/mt)
18:00:35 Join johnlocke [0] (dcff014d@gateway/web/freenode/ip.
18:00:47johnlockei need help with my clip+
18:00:52johnlockeis anyone out there?
18:01:01evilnickDon't ask to ask, just ask
18:01:13evilnickIf someone here can help, then they will
18:01:29johnlockei was following the guide at
18:01:45johnlockeup till the step with sudo fdisk -l
18:02:22johnlockeand only got a disk with no partition at 4MB capacity instead of the 979.75MB
18:02:39johnlockeam i doing something wrong?
18:03:14evilnickWhat OS are you using? Did you follow the steps after the instructions?
18:03:25evilnick"If it doesn't works, try to repeat it several times."
18:03:31johnlockeDisk /dev/sdc: 4 MB, 4231680 bytes 1 heads, 9 sectors/track, 918 cylinders, total 8265 sectors Units = sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disk identifier: 0xfffbfffd
18:03:57evilnick"You must make sure that the player was off before plugging the USB cable to the PC."
18:04:00johnlocketried more than 10, still the same
18:04:30johnlocke"If it doesn't works, try to repeat it several times." by "it doesn't works", does he actually mean the situation im in or something else?
18:04:39johnlockei disconnected the battery
18:04:50evilnickMy guess is "anything that isn't exactly as described"
18:05:13johnlockewell thats helpful
18:07:07johnlockeanyone else?
18:13:09copperhow do you guys brick the clip+?
18:13:39copperUSB problems?
18:16:27johnlockecheck this
18:17:27 Join krabador [0] (~krabador@unaffiliated/krabador)
18:20:37johnlockeits funny how theres a long list of people in this channel but no ones talking
18:20:54Torneno, it's entirely normal
18:24:35johnlockenormal, but still funny
18:24:51copperI can't stop laughing
18:25:57copperI wonder how much (if at all) bad quality microsdhc cards affect Rockbox functionning properly
18:26:22copperthat sentence doesn't sound right…
18:28:01copperany chance that Rockbox can miss some special case errors from the reader and go wonky because of it?
18:29:01johnlockei got a feeling the problems im having are related to storage issues since it was pluged in the the PC for file transfer just before it went wonky, then dead
18:32:52 Quit melmothX (Ping timeout: 240 seconds)
18:34:53johnlockeif anyone is looking, should i just short the pins in red or the ones in blue as well?
18:35:04 Join melmothX [0] (~melmoth@unaffiliated/melmothx)
18:35:46coppersounds hardcore
18:38:11johnlockedo you know stuff?
18:38:31coppernot really
18:39:51 Join Strife89 [0] (~Strife89@2602:306:250e:8d79:225:d3ff:fed6:15a)
18:41:33 Join mt` [0] (~quassel@rockbox/developer/mt)
18:43:22 Quit dfkt_ (Remote host closed the connection)
18:44:06 Quit mt (Ping timeout: 246 seconds)
18:45:00 Quit johnlocke (Quit: Page closed)
18:59:59 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
19:00:32 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
19:05:27 Quit Zarggg (Quit: Zarggg)
19:06:18 Join Zarggg [0] (
19:08:45 Quit scorche|sh (Changing host)
19:08:45 Join scorche|sh [0] (~scorche@rockbox/administrator/scorche)
19:22:42 Nick Gareth__ is now known as Gareth (~gareth@2607:ff38:2:83::3)
19:30:20 Join mt [0] (~quassel@rockbox/developer/mt)
19:32:45 Quit mt` (Ping timeout: 246 seconds)
19:41:01***Saving seen data "./dancer.seen"
19:48:43 Quit thomasjfox (Quit: Konversation terminated!)
19:59:58 Quit pamaury (Ping timeout: 268 seconds)
20:00:00 Join Epicanis [0] (
20:09:46 Quit kevku (Ping timeout: 260 seconds)
20:11:00 Join kevku [0] (~kevku@2001:470:27:773:0:feed:c0f:fee)
20:28:13 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
20:33:26 Quit mt (Ping timeout: 245 seconds)
20:43:44 Join kaputnik__ [0] (
20:47:06 Quit kaputnik_ (Ping timeout: 260 seconds)
20:58:52 Quit simabeis_ (Ping timeout: 264 seconds)
20:59:15 Quit Strife89 (Quit: Vamoose.)
21:00:19 Quit Belzebub (Ping timeout: 248 seconds)
21:04:57 Join simabeis [0] (
21:07:26 Join einhirn [0] (
21:08:00 Join Belzebub [0] (torrentow@gateway/shell/
21:11:51 Quit einhirn (Ping timeout: 240 seconds)
21:14:58 Join saratoga [0] (123e1cf8@gateway/web/freenode/ip.
21:15:43saratogacopper: the ffmpeg flac decoder is *really* fast though, so 50% slower then it is still very fast
21:15:58saratogaprobably the official TAK decoder uses more SSE stuff on windows
21:19:55saratogabeing closed source is annoying, and I told the developer a couple years ago that he'd either open source it himself or have it reverse engineered, but he didn't listen and here we are :)
21:22:16 Quit akaWolf (Ping timeout: 264 seconds)
21:41:04***Saving seen data "./dancer.seen"
21:41:51coppersaratoga: #rockbox-community doesn't share your enthusiasm about any of it
21:42:12copperand they also blame the decoding speed
21:43:19gevaertscopper: please go and misrepresent other people's opinions somewhere else
21:43:58copperhow was that not accurate?
21:44:41coppersaratoga replied to my assessment of the ffmpeg decoding speed
21:44:51copperhence the relevance
21:45:22copperoh screw it
21:47:54saratogaits tough to generalize the performance on x86 devices to that on ARM, but given the relative numbers to flac and that the format seems algorithmically similar, I expect decode performance to be quite good
21:48:18 Quit thomasjfox (Quit: Konversation terminated!)
21:55:09 Quit kaputnik__ (Ping timeout: 252 seconds)
22:01:35 Quit krabador (Quit: Bah...)
22:02:17 Join krabador [0] (~krabador@unaffiliated/krabador)
22:04:22 Quit krabador (Client Quit)
22:06:11 Join stoffel [0] (
22:09:34 Join kaputnik__ [0] (
22:17:54 Quit b1101 (Quit: b1101)
22:21:16coppersaratoga: I guess that would require more work than just porting the decoder
22:21:37copperoptimizations I mean
22:21:39saratogagenerally decoders need a lot of optimization to work well on portable systems
22:41:44 Join MethoS- [0] (
22:51:38 Join dfkt [0] (OxO29A@unaffiliated/dfkt)
22:56:06 Join TheLemonMan [0] (~LemonBoy@unaffiliated/thelemonman)
23:04:37 Quit n1s (Quit: Ex-Chat)
23:08:01 Quit kevku (Ping timeout: 260 seconds)
23:09:40 Quit stoffel (Remote host closed the connection)
23:22:41 Quit MethoS- (Read error: Connection reset by peer)
23:23:23 Join MethoS- [0] (
23:28:00 Quit melmothX (Quit: #)
23:39:18 Quit Bagder (Changing host)
23:39:18 Join Bagder [241] (~daniel@rockbox/developer/bagder)
23:39:33 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
23:41:08***Saving seen data "./dancer.seen"
23:44:03 Quit shamus (Read error: Connection reset by peer)
23:45:03 Join shamus [0] (
23:49:29 Quit Scall (Ping timeout: 245 seconds)
23:50:57 Join Scall [0] (~chat@unaffiliated/scall)

Previous day | Next day