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 2020-10-22

00:41:52***Saving seen data "./dancer.seen"
01:59:26efqwupdate regarding battery bench:
02:00:09efqwif I shut down rb and actually put the microsd into my computer, I do get the logs
02:00:24efqwbut I don't see anything if I `cat` the file in the console
02:00:39efqwso I assume everything was written when rb quits?
02:00:44efqwthis is odd
02:04:53 Join Stanley00 [0] (~stanley00@unaffiliated/stanley00)
02:06:07Stanley00I also tried battery_bench on my M5, with in-ear phone, it reported about 9h playing to go from 100% to 5%
02:06:21Stanley00*in-ear headphone
02:07:20efqwyeah, I have some low-impedance stuff plugged in as well to simulate regular usage
02:29:28_bilgus__are you guys using the test audio files for this?
02:30:59 Quit _3dsv (Ping timeout: 240 seconds)
02:32:44efqwI was not aware of the existence of a test audio file.
02:33:05efqwI just picked a random flac album I had
02:34:31Stanley00I also used my own flac album
02:34:36_bilgus__they are under codec tests but it gives a good base line for testing BB and doing codec bench too
02:35:00_bilgus__assuming your player handles them all without crashing
02:35:40efqwI'll do the codec bench after the bettery bench is done.
02:36:15_bilgus__its probably jaw dropping fast on that processor
02:36:52efqwyeah, that's what I'd expect tbh :P
02:38:35efqwsomething like 5000%+ perhaps
02:41:56***Saving seen data "./dancer.seen"
02:42:05 Join _3dsv [0] (
03:18:03 Join petur [0] (~petur@
03:18:03 Quit petur (Changing host)
03:18:03 Join petur [0] (~petur@rockbox/developer/petur)
03:48:37 Quit Rower (Read error: Connection reset by peer)
03:49:01 Join Rower [0] (
04:41:59***Saving seen data "./dancer.seen"
04:57:40 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
05:19:01Stanley00I'm working on the poweroff issue on m5
05:19:28Stanley00 <= not sure if we can use this and call to it to shutdown axp after issue the poweroff command
05:24:54 Join prof_wolfff [0] (
05:28:58Stanley00yup, it works on m5... if you all agree on this, we can just have something like `system("poweroff; sleep 10; ioctl /dev/axp173 magic_number");`
05:52:50 Quit prof_wolfff (Ping timeout: 256 seconds)
06:25:37 Join Rower- [0] (~Rower@
06:27:58 Quit Rower (Ping timeout: 256 seconds)
06:31:15wodzspeachy: do you by any chance know how asound.conf values corelate with alsa device string? I mean I am looking way to avoid dynamicaly crafting entry in asound.conf file for bt devices
06:42:00***Saving seen data "./dancer.seen"
07:01:25 Join Rower [0] (
07:05:08 Quit Rower- (Ping timeout: 272 seconds)
07:09:09speachywodz: type:deviceid[,subdevice]
07:09:41Stanley00efqw: <= my current work-around for poweroff issue
07:10:22wodzspeachy: how to pass device specific params?
07:11:10speachy'hw:0' is usually all that's needed for native hw but you're locked into hw capabilities only, 'plughw:0' is the wrapper that will do conversions/etc for you if the hw doesn't do what you want
07:12:09speachywhat sort of params? (I don't think you specify that in the devicestring, that's either in asound.conf or you specify the options when doing your hw/swparams setup api calls...)
07:12:26Stanley00speachy: I'm having weird issue with M5 volume button. it won't take effect until I press pause and the play the current song. Do you have any idea what may causing this?
07:12:52speachy(it's been well over a decade since I've done anything substantial with alsa..)
07:13:20wodzspeachy: For each bt device hiby creates entry in asound.conf like this pcm.80_C7_55_63_59_D6{
07:13:20wodz type bluetooth
07:13:20wodz bdaddr 80:C7:55:63:59:D6
07:13:20DBUGEnqueued KICK wodz
07:13:20wodz profile a2dp
07:13:52wodzthen you can reference this as alsa device '80_C7_55_63_59_D6'
07:14:15speachyif you do 'aplay -l' it will show you the perspective from alsa's api
07:14:22wodzI would prefer to avoid writing entries in aplay.conf
07:14:47wodz**** List of PLAYBACK Hardware Devices ****
07:14:47wodzcard 0: k1 [k1], device 0: k1-cs42l51 cs42l51-hifi-0 []
07:14:48wodz Subdevices: 1/1
07:14:48***Alert Mode level 1
07:14:48wodz Subdevice #0: subdevice #0
07:14:53wodznot really helpful
07:14:53speachyI don't know if that's possiblem. :/
07:17:56wodzbluealsa can do that
07:18:08wodzthe string is like this bluealsa:DEV=00:0B:D5:F5:F8:B9,PROFILE=a2dp
07:18:50speachyyou beat me to it. :)
07:18:52wodzsince bluealsa is based on pcm plugin from bluez4 I thought it would be possible here as well
07:19:59speachyI'd bet bluealsa's using a /etc/asound.conf that's been parameterized.
07:20:41speachysearch for '@args'
07:23:43wodzok, now that starts to make sense
07:24:03 Join cockroach [0] (~blattodea@pdpc/supporter/active/cockroach)
07:24:20wodzwhen parametrization was introduced? Maybe we can do the same
07:24:49***Alert Mode OFF
07:24:56speachyI'd be shocked if it wasn't supported on those hiby targets.
07:25:15speachyI _think_ they have alsa 1.0.26? circa 2013-ish.
07:25:27speachyparameters are probably at least a decade older
07:26:45wodzgood so it is a matter of crafting parametrized config
07:26:46speachyso we can add a generic parameerized bluetooth entry at the end of /etc/asound.conf when we patch in the bootloadr, or of hiby_player mucks with it (ie strips it out) we can do it the startup script that launches the loader
07:27:43 Quit Stanley00 ()
07:27:44speachybtw, tools/ is what's used to mangle the update.upt images these days.
07:29:45wodzlooks like changing type from bluealsa to bluetooth should do the trick
07:30:42wodzI mean pcm part
07:33:16 Join MrZeus_ [0] (~MrZeus@
07:34:18speachyStanley00, does the volume display change when you hit the buttons?
07:35:40speachyAlso, you need to run battery_bench until the device shuts down. the last time I did a bench run on an uncalibrated device, it was under 10% for over five hours.
07:49:28wodzspeachy: Sometimes I can get aplay to work with bt, but mostly I get this
07:49:30wodzPlaying WAVE 'test_file.wav' : Signed 16 bit Little Endian, Rate 44100 Hz, Stereo
07:49:30wodzALSA lib audio/pcm_bluetooth.c:2056:(audioservice_recv) Too short (0 bytes) IPC packet from bluetoothd
07:49:30wodzaplay: set_params:1166: Unable to install hw params:
07:49:30wodzFORMAT: S16_LE
07:49:33wodzSUBFORMAT: STD
07:49:35wodzSAMPLE_BITS: 16
07:49:37wodzFRAME_BITS: 32
07:49:39wodzCHANNELS: 2
07:49:41wodzRATE: 44100
07:49:43wodzPERIOD_TIME: (46439 46440)
07:49:45wodzPERIOD_SIZE: 2048
07:49:47wodzPERIOD_BYTES: 8192
07:49:49wodzPERIODS: 3
07:49:51wodzBUFFER_TIME: (139319 139320)
07:49:53wodzBUFFER_SIZE: 6144
07:49:55wodzBUFFER_BYTES: 24576
07:49:57wodzTICK_TIME: [0 0]
07:49:59wodzand bluetoothd dies
07:50:03wodzany idea?
08:36:29speachynot particularly. nothing of use in the bluetoothd logs?
08:42:04***Saving seen data "./dancer.seen"
08:43:21speachymaybe it wasn't actively paired at the itme?
08:44:10wodzbt-device -i reported paired and connected
08:46:22wodzBattery is dying in my device. This complicates everything further
08:47:45 Join prof_wolfff [0] (
09:10:18 Join t0mato2 [0] (~t0mato@
09:10:50 Quit t0mato (Ping timeout: 256 seconds)
09:10:50 Nick t0mato2 is now known as t0mato (~t0mato@
09:19:51 Join massiveH [0] (
09:30:16 Join Stanley00 [0] (~stanley00@unaffiliated/stanley00)
09:30:41Stanley00speachy: yes, the volume icon changed as I pressed the volume buttons
09:31:10Stanley00I also tried change volume from sound setting but it's still won't effect until I paused the current playing
09:34:56 Quit Stanley00 (Read error: Connection reset by peer)
09:36:34efqwit's been more than 7 hours and the m3k's battery is still at 3.863v
09:36:59efqwI think it's pretty safe to assume this thing can do more than 14hrs
09:39:41 Join Stanley00 [0] (0135c695@unaffiliated/stanley00)
09:41:55Stanley00since the poweroff is working on my m5 now, I guess I can do the battery_benchmark this weekend
09:42:49Stanley00also, where can I find the bluetooth code for hosted targets? I grep through the rockbox master but I can't find any
09:43:18Stanley00and what can I expected from current bluetooth support for hosted targets?
09:46:02 Quit Stanley00 (Remote host closed the connection)
09:48:22 Join Stanley00 [0] (0135c695@unaffiliated/stanley00)
09:55:30 Quit pamaury (Ping timeout: 272 seconds)
09:57:35 Quit Stanley00 (Remote host closed the connection)
09:59:58efqwthere is no BT code *at all* yet
10:07:22 Join Stanley00 [0] (0135c695@unaffiliated/stanley00)
10:08:05Stanley00efqw: oops, I thought I heard we have somewhat bluetooth for hosted here :(
10:08:25Stanley00anyway, did you check my work-around for m3k poweroff issue?
10:08:41efqwit's still doing the battery bench :P
10:09:35Stanley00I see
10:10:08Stanley00FiiO stated that m5 can last for 10 hours, so I could be happy with around 9 hours for rockbox
10:10:40Stanley00consider I didn't tune and turn off unneeded chipset/peripherals yet
10:11:29Stanley00funny thing is my laptop sometimes can see my M5 via bluetooth while I boot rockbox
10:24:39efqwfor me a dedicated DAP needs a fairly good battery life that lasts through multiple listening sessions without needing to recharge
10:28:53 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
10:29:12 Quit massiveH (Quit: Leaving)
10:30:54 Quit Stanley00 (Ping timeout: 245 seconds)
10:42:07***Saving seen data "./dancer.seen"
10:57:48 Quit Oksana (Ping timeout: 260 seconds)
11:07:20 Join Oksana [0] (~Wikiwide@Maemo/community/ex-council/Wikiwide)
11:16:20speachyStanley00: the m5 is a modification of the existing m3k/fiio drivers, correct?
11:24:24 Quit petur (Quit: Connection reset by beer)
11:25:05 Join Stanley00 [0] (0135c695@unaffiliated/stanley00)
11:25:25Stanley00sorry, I got very unstable network here
11:27:56 Quit Stanley00 (Remote host closed the connection)
11:29:08 Join Stanley00 [0] (0135c695@unaffiliated/stanley00)
11:29:47Stanley00if you're asking about the volume buttons issue. All I changed that can related to sound is replace '/dev/ak4376' to /dev/ak4377 on m5
11:30:53Stanley00one side note is it used to worked before when I edit m3k code directly, but then I pull the master branch, and create separated files for m5, the issue appears
11:47:09 Quit Stanley00 (Ping timeout: 245 seconds)
11:49:37 Join Stanley00 [0] (0135c695@unaffiliated/stanley00)
11:50:05Stanley00speachy: I think I found it, <= this muted supposed to be initialized by 0 instead?
11:50:52speachymute state wouldn't affect the inability to adjust the volume when playing (assuming you could hear anything to start with)
11:52:04Stanley00but later, at line 129, it won't enter the code to apply volume,
11:54:01speachyadd audiohw_mute(false); to the end of audiohw_preinit()
11:54:50Stanley00it make sense. So, we use value of -1 for another special meaning of muted state then?
11:55:21speachythats to make sure we explicitly call audiohw_mute()
11:55:35speachyinstead of relying on something happening implicitly.
11:56:20speachynot entirely clear it'll do any good on this platform
11:56:42speachysince there's no dedicated mute control on the audio path
11:57:09speachyoh! everyone, gerrit's going offline in a few minutes.
11:57:21speachyanongit and gitweb will continue fine
11:59:14 Quit Stanley00 (Ping timeout: 245 seconds)
12:01:23speachyand it's now offline.
12:02:45 Quit MrZeus_ (Ping timeout: 240 seconds)
12:03:31 Join Stanley00 [0] (0135c695@unaffiliated/stanley00)
12:09:17efqw`up 10:07, 0 users, load average: 2.08, 2.04, 2.05`
12:09:34efqwbattery still at 3.825v
12:09:48efqwdon't tell me this is gonna last 20h, lol
12:14:24fs-bluebot_Build Server message: New build round started. Revision 97b8692, 293 builds, 9 clients.
12:15:25speachyok, gerrit is back up.
12:16:19 Join MrZeus_ [0] (~MrZeus@
12:18:24 Quit Stanley00 (Ping timeout: 245 seconds)
12:24:35speachyStanley00: I committed that change I suggested. tbh I'm not sure how you got _any_ sound out at all without it.
12:26:13speachynever say never! :D
12:26:26speachyaro you playing that flac file, or something else?
12:29:40fs-bluebot_Build Server message: Build round completed after 915 seconds.
12:29:49fs-bluebot_Build Server message: Revision 97b8692 result: All green
12:30:31 Join Stanley00 [0] (0135c695@unaffiliated/stanley00)
12:32:33Stanley00my collections is mostly flac, with a few mp3 and ogg
12:32:58Stanley00there's maybe some code path relate to play/pause logic that effect the volume, I guess
12:36:42 Quit Stanley00 (Quit: Ping timeout (120 seconds))
12:42:08***Saving seen data "./dancer.seen"
12:43:18speachydo you have current code otherwise?
13:13:47 Join Stanley00 [0] (0135c695@unaffiliated/stanley00)
13:14:39Stanley00I was trying latest code from master before bed, but it looks like you forgot to update muted variable in audiohw_mute =]]
13:15:45speachy...oh, lovely.
13:16:11speachythat's what happens when I can't test the code I'm changing
13:16:43fs-bluebot_Build Server message: New build round started. Revision 02c5dd3, 293 builds, 9 clients.
13:20:07_bilgus__and what happens when I _CAN_ test the code I'm changing :p
13:20:31Stanley00it's time for bed, good night everyone
13:27:35Stanley00okay, now's bed time storry: I'm testing the test_files earlier, and it seems that ape_c5000 is killing my m5... cpu top at 97% and the sound is lagging... =]]
13:28:07_bilgus____builtin you mentioned xworld does it use xlcd or are the unrelated?
13:28:20 Part Stanley00
13:30:55speachyStanley000, APE @c5K is ... quite intensive.
13:30:59fs-bluebot_Build Server message: Build round completed after 856 seconds.
13:31:01fs-bluebot_Build Server message: Revision 02c5dd3 result: All green
13:31:02fs-bluebot_Build Server message: New build round started. Revision 1e12990, 293 builds, 9 clients.
13:32:34_bilgus__thats why I mentioned 'if the player can play them'
13:33:28_bilgus__I tell you what though I'm still tired of that song that most of the test files use
13:35:59speachymost (all?) of our targets can't handle ape @c4K either.
13:44:18fs-bluebot_Build Server message: Build round completed after 798 seconds.
13:44:20fs-bluebot_Build Server message: Revision 1e12990 result: All green
14:02:25 Join lebellium [0] (
14:21:09 Quit pamaury (Ping timeout: 265 seconds)
14:41:42 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
14:42:12***Saving seen data "./dancer.seen"
15:00:46_bilgus__well I tried and failed to make xlcd stride agnostic, oh well I left the pieces in there for anyone who wants to expand it
15:01:22_bilgus__getting tired of this patch, get it working then expand from there
15:06:22_bilgus__OK g#2811 SHOULD be ready!
15:06:24fs-bluebot_Gerrit review #2811 at : LCD core move buf ptr and address look up function viewport struct by William Wilgus
15:06:38speachy25th time's the charm, eh?
15:06:55_bilgus__look at the advanced keylock patch I think I had 50
15:07:55_bilgus__I do try to tend to push my days work up there so I can read through it on breaks
15:09:18_bilgus__there is still 2bit corruption on non-native strides but I want to get all these other targets right that can be a different patch since nothings using it ATM
15:10:41speachyit works!
15:10:51speachyokay, this is... funny
15:11:07speachyI hit play on the UIsim on my workstation, and music started playing out of my laptop.
15:11:46_bilgus__I didn't do that!
15:12:21_bilgus__ g#1417 I guess it was only 40 patch sets
15:12:23fs-bluebot_Gerrit review #1417 at : Selective Backlight/Advanced Softlock - Selective actions based on context by William Wilgus
15:12:36speachyI must have paired them via bluetooth at some point because nothing else makes any sense
15:13:18_bilgus__adv keylock was actually much more difficult but this one is far more annoying and sprawled out
15:14:45_bilgus__I'll commit this by the weekend assuming no more flaws pop up
15:15:51speachycan't say I've run into anything new, but it does fix two bugs I've personally tripped over.
15:16:17_bilgus__I think the other thing I want to do is to wire the dirty vp stuff a little deeper in the core to the point that it will not refresh a vieport that is already clean
15:16:59_bilgus__bugs you found prior to this rewrite? NO way :p
15:17:55speachyof course not, rockbox's code is pristine and perfect, as was handed down by the Swedish Demigods
15:18:27_bilgus__If I had to guess one of the bugs was the wps thing and a stride issue in plugins?
15:19:03speachydisplay&|viewport corruption in multiple plugins and that wps thing.
15:19:27speachypossible this likely fixes some of the crash-on-exit plugin issues seen on the hosted targets
15:20:22_bilgus__well it does force vp init now on plugin start and cleans it up on exit
15:21:03_bilgus__before it was just kinda luck of the draw
15:21:19CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
15:21:19*speachy nods.
15:51:27 Quit MrZeus_ (Ping timeout: 260 seconds)
15:58:47speachygrr, gitlab oauth integration into our gerrit instance is still shot
16:04:42 Join MrZeus_ [0] (~MrZeus@
16:06:27_bilgus__reviewing the framebuffer patch I see a lot of code for clipping that could probably be moved to a clip func ptr within the viewport
16:07:47 Quit fs-bluebot_ (Ping timeout: 260 seconds)
16:07:48_bilgus__same ida as the fb address function pointer, I mean it could be moved there as well but I wouldn't want the overhead on every pixel
16:07:54 Join fs-bluebot [0] (
16:10:28speachyI like the sound of that.
16:10:56_bilgus__hmm I really didn't consider the implications on the bootloaders I wonder how much of this code they pull in
16:12:23speachymost of 'em don't display anything at all.
16:13:17speachyso I expect minimal effect on bl code size.
16:14:11_bilgus__heeh I already got a error :p
16:14:42_bilgus__time to upgrade my toolchain
16:14:43speachymight have IRAM size problems.
16:15:09_bilgus__do I just rebuild off the new
16:16:30 Quit pamaury (Ping timeout: 258 seconds)
16:16:43speachyI'd recommend moving the old ones out of the way rather than installing over them, but otherwise it's just run and wait a while.
16:20:45 Quit Rower (Ping timeout: 240 seconds)
16:21:22_bilgus__ok when I get back I shall continue this adventure
16:21:42speachyonly (native) arm and m68k changed
16:28:30_bilgus__was about to ask
16:38:20braewoods_bilgus__: well i know the h120 bootloader does display stuff to remote LCD
16:38:24braewoodsand the native one
16:38:53braewoodsbut just text
16:40:24_bilgus__as longas it doesn't do scrolling should be ok and I believe I pulled scrolling from all bootloaders at some point
16:42:14 Quit Oksana (Ping timeout: 256 seconds)
16:42:16***Saving seen data "./dancer.seen"
16:55:15 Join Oksana [0] (~Wikiwide@Maemo/community/ex-council/Wikiwide)
17:28:32 Quit lebellium (Quit: Leaving)
18:42:18***Saving seen data "./dancer.seen"
19:19:57 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
19:19:57 Join fs-bluebot_ [0] (
19:22:23 Quit fs-bluebot (Ping timeout: 260 seconds)
19:23:14 Quit bluebrother^ (Ping timeout: 265 seconds)
19:35:07 Join Alexander [0] (
19:35:32 Nick Alexander is now known as Guest40192 (
19:36:13Guest40192Hey there! So i encountered an "error/bug" of some sort. Should i expose it here, or where should i head to?
19:36:59 Nick Guest40192 is now known as zazuradia (
19:38:26zazuradiaI mean, my player model is iPod 6G 160GB and i installed Rockbox less than a month ago.
19:43:04 Quit MrZeus_ (Ping timeout: 256 seconds)
19:51:00speachyfirst, are you using the 3.15 release or a dev/nightly build?
19:52:41zazuradiait says 3.15 in the Info section. and i don't use to install nightly nor dev builds.
19:55:01speachyplease see if the issue is stull present in a recent nightly or dev build, and if it still is, then it's worth talking about
19:55:58speachyFWIW I'm not aware of any reason to not be using a dev/nightly build
19:57:05zazuradiai avoid them in general because of potential stability problems.
19:57:57braewoodsi think that's more true for recent ports
19:57:58zazuradiathe "bug" is just that the player refuses to play any file beyond 14 hours. (ie: audiobooks, which is the main reason i use this ipod)
19:59:12speachyyeah, please make sure it's still present first
19:59:23speachyI cant' say I've personally tried anything over about 9h.
19:59:26zazuradiai just say it plain and simple because it may have to do with some kind of filesystem or hardware limitation already known. i've googled like an hour and didn't find anything related to rockbox that could be of use.
20:00:33braewoodszazuradia: is this a single file that's 14 hours+ or a bunch of files adding up to 14 hours+?
20:01:58zazuradiasingle files only. i've aded many days worth of music to the database and it plays as expected. but when it comes to a file longer than 14h it just skips it and starts playing the next one.
20:03:41braewoodsencoded as mp3?
20:04:16braewoodsthis may be an area that isn't well tested since i can't say i know many people that encode single files that long
20:04:57zazuradiampeg-4, 134kbps vbr,
20:05:03braewoodsso AAC?
20:05:58braewoodseither way first thing to try is a development build
20:05:59zazuradiathe file properties say mpeg-4, but i encoded them with the "send itunes as audio track" tool on mac.
20:06:13braewoodsi see.
20:07:27zazuradiai just installed the lates dev build i found, rebooted, tried to play and it skips every track over 14h
20:07:48braewoodsi wonder if it's something to do with RAM or so...
20:08:16braewoodszazuradia: one idea, see if encoding one of those files temporarily as something else like Opus would change anything
20:08:57braewoodsit may help narrow it down to codec specific issue vs general decoding issue
20:09:15braewoodsthough it may simply be running into some kind of limiter on length
20:12:22 Join Stanley00 [0] (~Stanley00@unaffiliated/stanley00)
20:12:54braewoodszazuradia: i assume you haven't modified yours with a higher capacity drive or so?
20:13:06braewoodsRB has known issues if used with iflash
20:13:46zazuradianope. only a sticker for the screen. everything else is out of the box. not even the hard drive or the battery
20:14:34Stanley00FYI: bbout ape file last night, m5 can handle c4000 with 60% cpu usage
20:16:02zazuradiabraewoods: i could try to reencode the file i have, but i can't output any other format in the mac's built-in tool
20:16:29braewoodswasn't saying you should... i've never used ipods myself.
20:16:35braewoodsi mean
20:16:38braewoodswith their tool
20:16:55braewoodsyou'd probably have to install third party tools to do it
20:18:07zazuradiaalso worth noting, i started using rockbox on this "newer" ipod because the (stock firm) 5G gave me problems keeping track on files over 14h too.
20:19:22zazuradiai mean, the files played well, but when i tried to ff or rw when reached that limit, if i tried to do it it would freeze, start over or skip the track.
20:19:39braewoodsjust how long are your audio books?
20:19:51braewoodsmost things i know of would split them across multiple files
20:21:40zazuradiafor example, the longest one is IT (S. King) with almost 23 hours.
20:22:29zazuradiai "make" them myself and leave them single filed for as long as they would go. i have no way of knowing how long will it end up being.
20:23:22braewoodswell, transcoding them as a more compact format like Opus may help narrow it down
20:23:24Stanley00Hmm... 14h is like 50.000 seconds. Maybe there's limit at 65536 or something like that?
20:24:10braewoodsit would be closer to 18h if that was the case
20:25:20Stanley00zazuradia: how did you come up with 14h limit, do you have other file that last around 10h?
20:25:51Stanley00May be it's 32768 limit for signed type
20:27:51braewoods32767 actually
20:28:27zazuradiawell, not seconds, but since this is a device intended to play music, maybe it calculates the time in minutes, and it doesn't have the capacity to track minute quantities over 4 digits, but 999 minutes is like 16h, and that doesn't match either.
20:31:16zazuradia(i'm trying to reencode the file to ogg, btw. but it's a huge file and it's taking some time)
20:31:48 Quit Acou_Bass (Ping timeout: 256 seconds)
20:42:20***Saving seen data "./dancer.seen"
20:44:55 Join Acou_Bass [0] (
20:48:32 Quit Stanley00 ()
20:52:41speachyzazuradia: oh, it's mp4? yeah, that's definitely a problem for larger files.
20:53:38braewoodsspeachy: why?
20:53:44braewoodsram requirements?
20:54:05zazuradiai have a file that is 14h exactly and another one 13 and a half or so. one of them doesn't play and the other does perfectly
20:54:19speachythe container is more oriented towards streaming.
20:54:54braewoodsis the same true for opus then? it was originally conceived for realtime audio streams
20:55:17fs-bluebot_ rockbox hangs on certain m4a files (bugs, unconfirmed)
20:55:25speachyI reported this bug over two years ago. :D
20:56:44speachybasically parsing the stream is memory intensive.
20:57:04braewoodsmy first hunch was some kind of RAM limiter
20:57:19speachyIIRC on less-ram-blessed targets the failure point happens a lot earlier (~6-7h, I think)
20:59:31zazuradiai could give you the gdrive link for the 14h file, if you want to try anything.
20:59:52braewoodsi'll give it a try
21:00:17zazuradia(it's in spanish, and pretty fast, btw)
21:01:26braewoodsmost of my rockboxs have 32MB
21:03:04braewoodszazuradia: for reference the upper limit is 2TB for storage media for RB, if that's even possible to achieve
21:03:12braewoodsdepends on the port as well
21:03:34braewoodsbut 2TB is the limiter inherited from the disk label most use
21:03:53braewoodsthough i have seen a few targets that use no partition table
21:05:08zazuradiamy ipod are windows formated. do you think this would change if i mac partitioned them?
21:05:35braewoodsand RB would no longer work
21:05:46zazuradiaoops. XDD
21:06:23_bilgus__braewoods has it try opus or flac
21:06:31speachy(or even mp3)
21:06:57_bilgus__14 hours is excessive
21:06:59braewoodsnot yet. i'm about to myself but it takes time
21:07:25speachyIt's not clear yet if SDUC (ie >2TB cards) required changes to the SD command protocol.
21:08:45braewoodsi'll see what happens once this is done
21:10:53zazuradiaomg. i changed the format (from whatever itunes uses, mp4, aac...) to ogg and it seems to work. :)
21:11:15_bilgus__:((( [ERR] Packed data (135305 bytes) doesn't fit in the firmware (134008 bytes)
21:11:20zazuradiai mean, i recoded the complete file. it took a little while. ^_^U
21:11:24_bilgus__I KNEW IT!
21:11:44_bilgus__!k is a lot of code
21:11:55speachy_bilgus__: what target?
21:12:10_bilgus__clip+ its the one thast closest to the line
21:12:21_bilgus__multiboot was within 30 bytes
21:12:36_bilgus__30 BYTES! lol
21:15:04braewoodsopus is encoding at 48x
21:15:21braewoodsfirst hour is nearly done
21:15:47zazuradiado you know of any offline converter?
21:15:51braewoodszazuradia: realistically your best option is opus. it's fairly CPU friendly and is very compact in terms of audio.
21:16:03braewoodszazuradia: uh...
21:16:12braewoodszazuradia: i'm just using a command line approach
21:16:26braewoodsfaad Memorias\ de\ idhun\ 2\ Triada.m4a -o /dev/stdout | opusenc - Memorias\ de\ idhun\ 2\ Triada.opus
21:16:29braewoodsfor my test
21:16:51braewoodsit will probably lose metadata
21:17:01braewoodsa more professional tool can probably help there
21:17:21braewoodsi'll see if this works
21:17:24braewoodslet me finish encoding first
21:17:31braewoodsat 3 hour mark
21:17:53braewoodsthis is with my quad core ivy bridge laptop
21:18:56braewoods17% faad, 100% opusenc lol
21:19:09_bilgus__personally flac is better optimized and faster encoding too
21:19:26speachybut a real waste of disk space for a 14h audio book.
21:19:27braewoodsbut it's lossless so more space consumption
21:19:31zazuradiahmm. command line isn't friendly enough. i would need some kind of front-end so i could batch reencode files.
21:19:51_bilgus__zazuradia, linux?
21:19:52braewoodsright now i'm just trying to re-encode this one file to see how it compares.
21:19:57braewoodshe's using Mac
21:20:06braewoodsi just assumed lol
21:20:20braewoodsat 5 hour mark
21:20:51braewoodsthis'll take around 20 minutes
21:20:55braewoodsat 48x
21:21:01zazuradianah, i use mac only for that tool, but i bootcamped my laptop quite long ago. i don't like mac, but it's the only thing i know of that has this handy tool.
21:21:17_bilgus__windows how about audacity it has a batch convert mode
21:21:29braewoodsdoes that include metadata translation?
21:21:43braewoodsto the extent that is possible
21:21:52_bilgus__doubtful if you want that the music brainz might help after the fact
21:22:27braewoodshm AAC is fairing better than I expected
21:22:39braewoodsi guess opus isn't able to dip the bitrate as much
21:22:49_bilgus__pretty sure it includes scripting support so not too hard to get a transfer method going
21:23:28braewoodsinterestingly i think opus will end up around 20% smaller
21:23:39braewoods8 hour mark
21:24:22zazuradia50% smaller in my case
21:24:31braewoodsfor ogg?
21:24:36braewoodsit probably dropped the bitrate
21:24:45zazuradiaaround 800mb to around 400
21:24:48braewoodsopus is around the same bitrate
21:25:01zazuradianah, from 134 vbr to 128. i don't need more than that.
21:25:06braewoodsi see.
21:25:12braewoodsmaybe ogg is better sometimes.
21:25:22braewoodsi thought opus was always the king lol
21:25:51zazuradiawell, since it's not music i'm working with, i'm okay with pretty low quality, if this means i'm gonna save on disk space.
21:26:27braewoodsi've mainly used opus to avoid having tradeoffs in ripped CD tracks
21:26:32braewoodsthey're much smaller than mp3s
21:26:58braewoods3.3 hours to go :x
21:27:45zazuradiamany years ago i would care about codec compatibility, but it's nothing to worry about today. rockbox and vlc got me covered.
21:29:29speachyok...32-bit to 38-bit addressing. But the old address field was only 32-bit.
21:31:47speachywhich implies new CMDs were defined.
21:32:34__builtin_bilgus__: xworld and xlcd are unrelated
21:32:53braewoodszazuradia: 13% smaller
21:33:02__builtinxworld's name was inspired by xrick, which was sitting in gerrit around the same time
21:33:10zazuradiasame settings, i guess?
21:33:24_bilgus__ok I was just making sure, I had started converting xlcd to arbitrary strides but it kept causing issues
21:33:41speachyunless someone happens to have a copy of the SD 7.0/7.1 specs lying around....
21:33:55braewoodszazuradia: -shrug- But it works.
21:34:02braewoodsso whatever works i guess lol
21:34:05braewoodsanything but AAC
21:34:07_bilgus__also the patch is up on gerrit if you want to check out the changes that apply to your babies I'd highly appreciate it
21:34:33__builtinalright, I'll take a look
21:34:42__builtin"my babies" :)
21:35:52_bilgus__for the most part I only gave them their own FB I didn't do anything to optimize beyond using rb->screens[MS].currentvp for stuff that needs performance it a lot
21:36:06zazuradiaoh yeah. when i started using this tool i used to end up with files above 9GB, for some reason. the tool saves a huge file to the desktop, THEN it recodes it through itunes and saves it in it's directory system. i learned this second file sounds about the same and it's like 1/10 the size.
21:36:48_bilgus__'it a lot' yeah no clue either..
21:37:22__builtindoesn't look like a huge change to my plugins at least
21:37:33__builtinif it compiles I've got no issue with it
21:40:31speachySo the question is.. did they bump up the block size from 512B to 32K, or did they define a new cmd to allow specifying the extra 6 bits?
21:47:26 Quit cockroach (Quit: leaving)
21:51:44zazuradiaokay guys, it's pretty late here and i kind of need a little bit of sleep.
21:52:16zazuradiaTHANKS A LOT for the help solving my problem. i can't thank you enough.
21:52:19zazuradiasee ya!
21:53:27zazuradiaoh btw, i started a thread on the forums before anyone here responded, so i marked it as [SOLVED] and replied with the answer and the steps for the workaround. :)
21:54:14_bilgus____builtin, thanks
22:05:23 Quit zazuradia (Remote host closed the connection)
22:16:38 Quit prof_wolfff (Ping timeout: 256 seconds)
22:42:21***Saving seen data "./dancer.seen"
22:53:19_bilgus__speachy WELL it turns out that the bootloader doesn't fit @ head either [ERR] Packed data (135305 bytes) doesn't fit in the firmware (134008 bytes)
22:53:45_bilgus__let me double check that I'm not doing something stupid
22:57:05_bilgus__so yeah my patch doesn't increase the size at all but now gonna have to find 1K somewhere
22:57:59_bilgus__speachy does the buildfarm create bootloader builds anymore?
22:59:03_bilgus__AHH I bet the build but don't go through the mkamsboot stuff
23:20:04speachyyeah, they just genetate the binaries.
23:20:14speachyas in whatever 'make' generates
23:21:46speachywonder where teh size bump came from. (git-bisect?)
23:21:54speachy(new toolchain?)
23:23:03_bilgus__maybe and that might be why I was sticking to the old one originally but can't say for sure
23:24:13_bilgus__I do know buffers in bss are inconsequential (~20 bytes each) because of the compression
23:24:55_bilgus__so this is going to be a stictly code game I already did this once to get MB inthere so the easy stuff is gone
23:25:19_bilgus__rtl language support can probably go
23:25:54_bilgus__diacritics can probably get condensed to a single codepage or so
23:27:32_bilgus__if we make the boot logo the size of the screen (if its not already) we can get rid of the lcd buffer probably not too much there on the clip+ though
23:28:24_bilgus__oh yeah 20 bytes!
23:35:32_bilgus__wonder what buflib is used for in the bootloader
23:44:52 Quit [7] (Disconnected by services)
23:44:58 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
23:45:31_bilgus__well printf is using 2k wonder if unneeded things are being exposed again
23:48:59_bilgus__maybe it is time to put in a tiny limited printf with the bootloaders
23:56:41_bilgus__looks pretty standard bootloaders use printf snprintf springt {vsnprintf}
23:56:59_bilgus__and I see %x %d %ld %s %02d 0x%02x %.4s
23:57:26_bilgus__so standard types and still need precision and padding
23:58:20_bilgus__wonder what our old printf looks like from before this expansion by JhMikes

Previous day | Next day