[Ffmpeg-devel-irc] ffmpeg.log.20130123

[burek]

[00:04] <dionys> thanks heaven - it works with the latest ffmpeg on my mac :-)
[00:13] <dionys> in order to edit and cut some 1440x1080 MXF files in blender for youtube, I like to format them down to 1280x720. What codec can you recommend? mp4, or avi?
[00:17] <teratorn> dionys: those aren't codecs :)
[00:21] <dionys> sure, so mpeg4 or h264. My guess is to convert into something raw, then edit it in blender and render it to avi
[00:26] <suzaru> i noticed joining crf encoded mp4 files results in vfr result. even when using "-r #" . any suggestions for maintaining cfr output?
[00:41] <llogan> emc: 2004. absolutely ancient.
[00:42] <llogan> just tell us what you're trying to do and we'll give you a ffmpeg command that isn't 9 years old
[00:42] <emc> llogan :) & hehe, i've been playing around with different settings.
[00:43] <emc> i managed to find something that worked. but thanks for your help.
[00:43] <llogan> i recommend using a recent ffmpeg if you have more to encode
[00:44] <emc> but maybe somebody of you knows that: i have a screencast that is recorded at 1280x960 & and i need to downsize it to 1024x760. what's the best way to do this to keep the text still sharp?
[00:45] <emc> llogan: great, thanks for that!
[00:45] <llogan> emc: you can use the scale filter -vf scale=1024:-1
[00:45] <llogan> and what output format are you wanting?
[00:47] <emc> llogan, i want to play it on an ipad.
[00:47] <emc> so probably h.264?
[00:48] <emc> i will try the scale filter, thanks for the hint. much appreciated!
[00:48] <llogan> damn, that's not what i wanted...
[00:48] <llogan> emc: https://ffmpeg.org/trac/ffmpeg/wiki/x264EncodingGuide
[00:49] <emc> cool, thanks llogan!
[00:49] <emc> any recommendation what's the best format for a screencast?
[00:49] <emc> to avoid compression artifacts etc?
[00:49] <llogan> H.264 will be find for ipad.
[00:49] <llogan> *fine
[00:49] <llogan> use CRF as shown in guide
[00:50] <llogan> i'm not sure what profile ipad can use
[00:50] <emc> ok, i will try, thanks a lot!
[00:51] <llogan> high profile apparently...you'll be fine
[00:52] <llogan> wow the Gimps rotate is incredibly slow
[01:03] <funyun> hi. does anyone know how to make ffmpeg out a log file while encoding?
[01:03] <funyun> output*
[01:07] <fatpony> even though i specify -r 23.976 when encoding x264, mediainfo says output mkv has a variable bitrate, is this normal?
[01:07] <fatpony> s/bitrate/frame rate/
[01:11] <funyun> fatpony: -r is framerate. -b is bit rate
[01:11] <fatpony> funyun: i know, 01:04 < fatpony> s/bitrate/frame rate/
[01:15] <llogan> fatpony: i trust ffmpeg more than mediainfo
[01:16] <llogan> and you should use -r 24000/1001 instead of -r 23.976
[01:16] <fatpony> yeah i trust ffmpeg too but it seems strange still... i guess i'll force the bitrate in mkvmerge
[01:28] <suzaru> i get vfr too with joining mp4s that are encoded as crf, used r parameter
[01:29] <suzaru> avidemux has the same problem in 2.6.x. even though content is really crf. i don't think mediainfo is showing it wrong
[01:30] <suzaru> i dont know how to force cbr in the container instead of vfr which ffmpeg introduces
[01:30] <suzaru> cfr*
[01:43] <emc> does anybody know what "illegal instruction: 4" could mean? http://pastebin.com/Mc3BWLCv
[01:44] <sacarasc> What does `file ffmpeg` say?
[01:44] <emc> sacarasc: ffmpeg: Mach-O 64-bit executable x86_64
[01:45] <emc> sacarasc: i guess it's the wrong binary?
[01:45] <sacarasc> That would be my first guess...
[01:46] <emc> sacarasc: great, thanks& will look for another one
[01:49] <dunpeal> Hey. Anyone knows what this error message means?  "decode_band_types: Input buffer exhausted before END element found"
[02:52] <spillere> is there how to convert a rmvb into mkv?
[02:53] <relaxed> did you try ffmpeg -i blah.rmvb -c copy output.mkv
[03:04] <spillere> will try
[03:04] <spillere> thanks
[05:21] <Demon_Fox> What is a way to reduce quality without setting a bitrate?
[05:22] <Demon_Fox> I know it starts with a q
[05:22] <Demon_Fox> quantilizer?
[05:22] <Demon_Fox> quant
[05:27] <Demon_Fox> I know xvid had an option to do variable bitrate encoding more directly, but what was it?
[06:31] <relaxed> Demon_Fox: maybe -q:v 2 (it works with ffmpeg's mpeg4 encoder)
[06:38] <Mista_D> fresh centos install, on ffmpeg compile I get "/usr/bin/ld: cannot find -lc"... any ideas?
[06:40] <Mista_D> my mistake.. "yum install glibc-static" worked.
[07:18] <praveenmarkandu> hi is there anywwhere i can read up about the -threads command in ffmpeg
[07:18] <praveenmarkandu> the man doesnt seem to say much
[07:18] <praveenmarkandu> like what does thread 0 do?
[07:19] <praveenmarkandu> or if i have 4 transcodes going at one time, how can i make each one go to its separate thread
[07:21] <relaxed> praveenmarkandu: use -threads 1 for each
[07:21] <praveenmarkandu> relaxed, it will then automatically assign a free thread to each one ?
[07:21] <relaxed> 0 is auto when using libx264 (the default)
[07:22] <praveenmarkandu> oh okay
[08:51] <praveenmarkandu> can anyone give some basic system specs to handle 4 transcodes?
[08:51] <praveenmarkandu> will an i7 sandy bridge suffice?
[08:53] <spaam> it depends on what you are going to encode.
[08:55] <praveenmarkandu> spaam, h264
[09:00] <praveenmarkandu> im actually down encoding like 10mbps files to 1.5mbps, 1.2mbps, 800kbps and 500kbps files
[09:00] <praveenmarkandu> and i want to do it in one shot
[09:00] <praveenmarkandu> looking into also xeon processors
[09:01] <praveenmarkandu> company buying a hp blade server for dedicated transcoding
[10:45] <plavcik> quit
[11:51] <kode54> is it possible to do frame doubled yadif?
[11:52] <kode54> since I thought yadif was capable of motion estimating two frames out of two fields
[11:57] <untaken> with 2 pass encoding, what does the pass 1 do exactly? As it creates no file other than a log? I notice a lot of command say this rawvideo -y /dev/null... so where does the first pass actually do. I understand the second pass, as it creates the file
[12:03] <JEEB> it goes through the clip to see its complexity in one way or another, also possibly decides on the positioning of IDR frames so it isn't done in the second pass to make it faster
[12:04] <JEEB> bit rate based encoding pretty much needs this to be able to know how to evenly distribute bits for a leveled'ish "quality" while hitting a certain average bit rate goal
[12:05] <kode54> JEEB: I take it you don't know the answer to my question
[12:05] <untaken> JEEB: thanks for that, very interesting
[12:05] <JEEB> kode54, yadif had a mode for one picture from one field if you look at the filter's documentation :P
[12:05] <JEEB> it's not a really great bobber tho
[12:06] <JEEB> untaken, do note that only modes that need to hit a certain average rate need something like this, as the rate control needs to be limited
[12:06] <JEEB> stuff like x264's crf rate control mode just set a value to what you specify and encode with it through the clip
[12:07] <JEEB> (unlimited rate control with rate factor set to X)
[12:11] <JEEB> with x264 you can have a long as hell lookahead though, so with slower settings the true need for a second pass gets lower, but naturally 2pass > lookahead in most cases
[12:30] <erle-> what can be the problem if atempo filter just does nothing?
[12:30] <erle-> i use 2.0 btw, a speedup factor in range
[12:35] <erle-> it just converts and does everything i tell it just tempo is ignored
[12:35] <erle-> audio bitrate, sample rate etc is converted as i wanted it
[12:41] <kode54> ah yeah
[12:41] <kode54> seems he was using yadif=1
[12:41] <kode54> but the issue may be that his videos are not being detected as the correct field order
[12:41] <kode54> which didn't seem to be an issue for two fields become one frame
[12:42] <erle-> i was using
[12:42] <erle-> -filter:a 'atempo=2.0'
[12:42] <erle-> only mp3 audio
[12:42] <erle-> no video
[12:45] <erle-> i would like to speed up the mp3
[12:45] <erle-> it is a talk from a very slow professor
[12:45] <erle-> i would like to keep the tone
[12:45] <erle-> any suggestions what filters are available?
[12:46] <erle-> speedup to 1.5x 1.7x and 2.0x
[12:46] <durandal_1707> filter works fine here, you still did not show your ffmpeg command output
[12:46] <durandal_1707> without it i cant help you
[12:47] <durandal_1707> ffmpeg -filters shows all available filters for that version/build
[13:01] <erle-> durandal_1707, that filter doesnt seem to be available
[13:02] <erle-> now i made a audacity script
[13:02] <erle-> thank you anyway
[16:06] <Guest14279> can any one help me with ffprobe options -show_frames ?
[16:06] <Guest14279> ffprobe -show_frames  *.mp4
[16:06] <Guest14279> it shows unrecognized option
[16:06] <Guest14279> ?
[16:07] <Mavrik> just how old is your ffprobe?
[16:07] <Guest14279> installed few min back
[16:08] <Guest14279> currently using linux mint os
[16:08] <klaxa> sounds like you installed avconv :V
[16:08] <Guest14279> can u please tell me what it is ? avconv :v ?
[16:08] <Mavrik> yep
[16:09] <klaxa> avconv is an ffmpeg fork
[16:10] <Guest14279> avprobe version 0.8.4-6:0.8.4-0ubuntu0.12.10.1, Copyright (c) 2007-2012 the Libav developers
[16:11] <Guest14279> this is shown inthe sarting line of the out put
[16:11] <klaxa> ask in #libav or get ffmpeg (static build or compile from source)
[16:12] <klaxa> with ffprobe 1.0.1 i can confirm it to work
[16:13] <Guest14279> thank u
[16:37] <guest7> Hi everyone. I'm developing a multithreaded (pthreads) piece of software that uses ffmpeg (libav{format,codec} to be more precise). If avcodec_open2() has been called once, then used with avcodec_encode_video() for a while and then closed using avcodec_close() and I try to call avcodec_open2() again with the same encoder codec, I get a segfault. I lock around it as described in the documentation so there is no concurrent access. Any i
[16:42] <durandal_1707> guest7: and?
[16:46] <guest7> side question: avcodec_close() doesn't do anything to the codec associated with the context, does it? I'm using the same encoder codec in several different codec contexts concurrently.
[16:48] <durandal_1707> avcodec_close free all buffers codec may allocated
[16:48] <durandal_1707> so if you still use same context it will obviously crash
[16:49] <guest7> not the same context, a different context with the same codec as the former.
[16:51] <guest7> I mean, avcodec_find_encoder() just looks for the codec and returns a pointer to it so I figured it doesn't matter whether I call it seperately for every context or just use the same codec for all the contexts.
[16:51] <guest7> durandal_1707: Is that correct?
[16:56] <durandal_1707> guest7: there is documentation
[16:56] <Umeaboy> Hi!
[16:56] <Umeaboy> Would it be possible to download Premium-content with ffmpeg by somehow using the userinfo? I mean like this example: http://pastebin.com/U6q6pssd
[16:57] <guest7> durandal_1707: Are you referring to anything specific regarding my question?
[16:57] <durandal_1707> guest7: that particular function accpets codec is as arguments and return encoder struct with calls to init/close/encode....
[16:58] <durandal_1707> and it always return same pointer to struct if given same argument
[16:58] <durandal_1707> so you do not need to call that function multiple times to get same enocder
[16:58] <guest7> durandal_1707: Right, so basically what I've been saying is correct? There exists only one encoder codec object for every codec.
[16:59] <guest7> durandal_1707: All right, so that should be fine then, thanks.
[16:59] <durandal_1707> guest7: there is multiple encoders possible per codec
[16:59] <durandal_1707> if you want specific one there is another function which searches by name
[17:00] <durandal_1707> guest7: AVCodec *avcodec_find_encoder_by_name(const char *name);
[17:00] <guest7> durandal_1707: Yeah I know that one, thanks for the hint though.
[17:01] <guest7> durandal_1707: Anyway, my original question is the bigger problem, do you have any idea what could be happening in avcodec_open2() that causes a segfault? I'm pretty sure I clean up properly before opening the codec in a different context.
[17:02] <durandal_1707> guest7: what version are you using?
[17:02] <durandal_1707> that function is not thread safe, mentioned in documentation
[17:03] <guest7> durandal_1707: The latest version available on Ubuntu, libavcodec53
[17:03] <guest7> durandal_1707: I'm locking around the function.
[17:04] <durandal_1707> you probably should always call alloc_context
[17:04] <guest7> durandal_1707: libavcodec 53.35.0 (major/minor/micro)
[17:04] <guest7> durandal_1707: I'm doing that too.
[17:04] <durandal_1707> that is very old
[17:05] <durandal_1707> you call alloc_context before every call to open2?
[17:07] <guest7> durandal_1707: No, I call avcodec_alloc_context3(), then avcodec_open2(), then I use avcodec_encode_video() for a while, then I call avcodec_close() and av_free() on the context again. This may repeat a number of times.
[17:07] <guest7> The encoder codec stays the same across the different contexts though, it is opened only once using avcodec_find_encoder() at the beginning of the program.
[17:08] <Umeaboy> I've perfomed this command to download from SVT Play before: ffmpeg -i "http://svtklipp-f.akamaihd.net/i/fritt/20130121/K-20130121-222819-4387/GULDBAGGEN-2013-54ac85ab5be910b4_,900,348,564,.mp4.csmil/index_0_av.m3u8?null=&e=b2a7ace1bf661c32&id=" -acodec copy -vcodec copy -absf aac_adtstoasc "Ett pucko i Phuket.mp4"
[17:09] <durandal_1707> guest7: you must call alloc always before open
[17:09] <Umeaboy> Would it be possible to download the content even thou it's for Premium-users?
[17:10] <guest7> durandal_1707: That's what I'm doing already.
[17:10] <durandal_1707> guest7: i'm lost, you said "No" first time
[17:11] <quechec> hi; can I say something like `ffmpeg -i INPUT -c:a vorbis -c:b 256k -vn OUTPUT.ogg' and tell ffmpeg to set the output bitrate lower, respectively, depending on whether it makes sense regarding the input quality?
[17:11] <guest7> I said: "I call avcodec_alloc_context3(), then avcodec_open2(), then I use avcodec_encode_video() for a while, then I call avcodec_close() and av_free() on the context again."
[17:11] <guest7> durandal_1707: Right, that should have been a yes. Sorry to confuse you.
[17:14] <durandal_1707> and it crashes on next open?
[17:16] <guest7> durandal_1707: Yes. Although sometimes only the third or fourth time or so, which makes me think I could be using it wrong. But so far you would say everything look correct?
[17:16] <durandal_1707> nope, i did not even looked at code ....
[17:18] <durandal_1707> i mean every open2 shuld come after alloc context
[17:19] <msmithng> I posted this a couple days ago, but didn't see any reply... I'll ask again and then shoot it to the list
[17:19] <msmithng> is there anyway for me to cross reference a Lavf version to a git branch revision?
[17:19] <guest7> durandal_1707: I'm talking about the general sequence of function calls and allocation of objects.
[17:19] <guest7> durandal_1707: I do call alloc_context3 before every open2.
[17:19] <durandal_1707> good
[17:20] <durandal_1707> msmithng: can you elaborate? I do not know what you really mean.
[17:21] <durandal_1707> guest7: well. problem may be buggy version you are using....
[17:22] <guest7> durandal_1707: valgrind tells me there's an invalid write of size 4 in avcodec_open2(). I'll try to debug that, thanks for your help and time, I appreciate it!
[17:23] <msmithng> durandal_1707: sure... let's say I'm seeing Lavf56.blah.foo in my output... can I cross reference that to a particular ffmpeg version, N-41668-g564bb24 for example
[17:25] <msmithng> I'm sorry if this doesn't make sense. I'm speaking directly to lavf53.6.0
[17:26] <durandal_1707> msmithng: lavf*.*.* is version number which does not map 1 to 1 with ffmpeg-version N-.... ....
[17:26] <msmithng> ah, fair enough
[17:26] <durandal_1707> there may be multiple versions N-.... with sampe lavfX.Y.Z number
[17:26] <msmithng> that makes sense
[17:27] <msmithng> I'll see if I can get the actual ffmpeg-version N-... from the client
[17:27] <durandal_1707> i think last git commit is not exported in lib versions
[17:27] <msmithng> anyway to confirm that?
[17:28] <msmithng> if not, no worries
[17:28] <durandal_1707> so you are only left with relase builds which have  known git version....
[17:32] <antonello> hallo , I am doing an audio transcode , Can help me ?
[17:34] <msmithng> Ah, ok. This is coming back to me. I've been removed from the community for a few months on other projects. Thanks for the reminder!
[17:34] <durandal_1707> antonello: yes, can you explain what are you doing?
[17:36] <guest7> durandal_1707: FYI: I just grabbed the debug symbols and found out that the segfault happens in avcodec_open2() -> encode_init() -> MPV_encode_init() -> ff_convert_matrix(). I just loaded the source files and looked at it but I don't know what's going on. Might try a more recent version first.
[17:38] <antonello> i want create I audio transcoding , in this moment i want only obtain a raw file from aac and encoding in aac , after i want obtain a pcm.My actual problem is ....my output file is only noise
[17:38] <antonello> Can i post my code ?
[17:42] <durandal_1707> antonello: sure, use pastebin or similar
[17:45] <antonello> ok this is my code http://pastebin.com/QuEw2ZrF
[17:49] <durandal_1707> antonello: you dont set sample_fmt for decoder/encoder
[17:50] <antonello> yes it is AV_SAMPLE_FMT_S16P
[17:50] <durandal_1707> you set request_sample_fmt instead if decoder support that sample fmt it will use one, otherwise you need to check what sample format encoder/decoder is using
[17:50] <durandal_1707> antonello: decoder/encoder set sample_fmt value
[17:50] <durandal_1707> not you
[17:51] <durandal_1707> also sample fmt tells how raw audio data you get from decoder or feed to encoder looks like
[17:52] <durandal_1707> data you give to encoder must have same sample_fmt encoder expect
[17:52] <durandal_1707> encoder/decoder do not make any conversion between them
[17:53] <durandal_1707> so if decoder outputs SAMPLE_FMT_S16 but encoder wants SAMPLE_FMT_S16P and you give encoder exact data from decoder you will get mostly noise
[17:54] <antonello> if i don't set in decoder the sample_fmt  it use the AV_SAMPLE_FMT_FLTP
[17:54] <durandal_1707> what decoder?
[17:54] <antonello> but this format is unsupportet by encoder
[17:55] <durandal_1707> swr_convert from libswresample converts between different SAMPLE_FMTs
[17:56] <durandal_1707> antonello: sample_fmt should not be modified from user
[18:01] <durandal_1707> does somebody know important reasons of current dsputil split surge?
[18:01] <durandal_1707> lol. wrong ch
[18:02] <antonello> ok , I use  swr_convert for convert SAMPLE_FORMAT and after codific the new raw format , but the must I set the coding SAMPLE_FORMAT ?
[18:02] <durandal_1707> you set input and output sample format as argument for swr_convert
[18:03] <durandal_1707> there are examples in doc/example of source code
[18:05] <antonello> I read that, but i found very difficulty combyne the encoding and decoding
[18:05] <durandal_1707> what exactly is difficult?
[18:06] <durandal_1707> you need extra buffer to resample input sample format to output sample format
[18:06] <durandal_1707> and resampled buffer you give to encoder
[18:07] <antonello> I tried but avcodec_encode_audio2 gave me an error, now I think it's for the incopatible raw format
[18:10] <antonello> ora tu pensi che il mio codice possa funzionare?
[18:10] <durandal_1707> antonello: paste error you get
[18:10] <antonello> it's only RUN FAILED exit 1
[18:17] <Rajeeev> with the latest ffmpeg my thoughput goes down is there any known issues?
[18:19] <Rajeeev> if i do 50 simultaneous trans-coding my CPU usage is low but Mbps per CPU percentage is going 30% higher?
[19:11] <Holden> saste, Hello, are you there?
[19:12] <saste> Holden, yes
[19:12] <DelphiWorld> hey everyone
[19:12] <DelphiWorld> burek:
[19:12] <Holden> saste, hi... I tried to debug the problem we were talking about yesterday, I wanted to show you a pastebin, do you have 5 minutes?
[19:12] <DelphiWorld> :P
[19:13] <saste> Holden, show it
[19:13] <DelphiWorld> saste: hey, how do i find the libx264 presets?
[19:14] <Holden> saste, thanks, here it is: http://paste.ubuntu.com/1563787/
[19:14] <saste> DelphiWorld, checing in the docs ;-)
[19:14] <saste> *checking
[19:14] <DelphiWorld> saste: ;)
[19:14] <DelphiWorld> saste: i checked it but didnt find them
[19:15] <saste> DelphiWorld, sorry i have no time, check the wiki as well
[19:15] <saste> -preset or something like that
[19:16] <saste> Holden, no i don't know how to fix it at the moment
[19:16] Action: DelphiWorld push saste ;-)
[19:17] <saste> Holden, but again filing a ticket seems in order, so we don't discard it/you can track it
[19:20] <Holden> saste, ok, I will file a ticket then with all the details, I was hoping I could make it work (seemed easy) but no luck :|
[19:27] <DelphiWorld> saste, the issue is if i use a preset i get: File for preset 'ultrafast' not found
[19:29] <DelphiWorld> saste: log: http://paste.debian.net/227914/
[19:30] <DelphiWorld> saste: configuration: http://paste.debian.net/227915/
[19:32] <saste> DelphiWorld, ffserver -> out of my league
[19:32] <DelphiWorld> saste: same issue if i use ffmpeg... issue is the file not found
[19:32] <saste> pb ffmpeg command and output
[19:33] <DelphiWorld> saste: see... in /usr/local/share/ffmpeg/ i see libvpx-1080p50_60.ffpreset  libvpx-720p50_60.ffpreset  libx264-ipod640.ffpreset
[19:34] <DelphiWorld> saste: what should i pb
[19:34] <saste> DelphiWorld, can you read?
[19:35] <DelphiWorld> saste: i can. but you just done pb, what i  should pb for you
[19:35] <saste> *ffmpeg* command
[19:36] <DelphiWorld> saste: i'm using ffserver directly and ffserver is launching ffmpeg. ffmpeg -i http://localhost:8090/sast http://localhost:8070/sast.ffm
[19:36] <DelphiWorld> first url is source, second is ffserbver to transcode
[19:52] <DelphiWorld> saste: , ffmpeg log: http://paste.debian.net/227931/
[19:54] <DelphiWorld> Error while opening encoder for output stream #0:1 - maybe incorrect parameters
[19:55] <saste> DelphiWorld, broken ffmpeg default settings detected
[19:55] <DelphiWorld> saste: what setting, where?
[19:55] <saste> you should use libx264 presets
[19:55] <DelphiWorld> saste: is it used but inside of FFSERVER
[19:55] <saste> the libx264 wiki encoding guide tells everything you want to know about that
[19:56] <DelphiWorld> saste: so where did i get it from then ?
[19:57] <saste> DelphiWorld, do your own research, i'm not an help desk
[19:58] <saste> about ffserver i can't help
[19:58] <DelphiWorld> thank saste
[20:24] <DelphiWorld> saste: ok, you have ffmpeg compiled and runing?
[20:26] <DelphiWorld> saste: ffprob http://46.165.201.28:8090/kashmir.flv
[21:47] <brontosaurusrex> how would "422 hi-quality photojpeg" command line look like?
[21:56] <llogan> brontosaurusrex: -pix_fmt yuvj422p -q:v 2
[21:56] <llogan> i guess
[21:56] <brontosaurusrex> ok, whats that "2" ?
[21:58] <llogan> qscale range for mpeg*: a linear scale of 1-31 where 31 is lowest quality
[21:58] <brontosaurusrex> ok, but iam getting really low quality with this
[21:59] <llogan> there was a time when qscale didn't work with jpeg, IIRC.
[21:59] <brontosaurusrex> expected bitrate  for 1920x1080 should be around 100 Mbits and up, getting 12 Mbits
[21:59] <Mavrik_> wat
[22:00] <llogan> lol
[22:00] <brontosaurusrex> what?
[22:00] <brontosaurusrex> using almost latest git
[22:01] <llogan> you asked about a "photojpeg" which i assumed a normal still jpeg image.
[22:01] <llogan> then you mention 100Mbits
[22:02] <llogan> -flags +quality-enhance!!!!!!!!!!!
[22:02] <brontosaurusrex> photojpeg in quicktime container
[22:02] <llogan> i don't know what that is
[22:02] <brontosaurusrex> me neither
[22:02] <brontosaurusrex> certain internet service requests that ....
[22:02] <brontosaurusrex> iam guessing
[22:03] <llogan> show output of: "ffmpeg -i input.mov" if you already have one of these to see what the non-apple name is for it
[22:03] <brontosaurusrex> ok
[22:08] <brontosaurusrex> llogan: https://dl.dropbox.com/u/79532365/photosomething.txt
[22:09] <llogan> -c:v mjpeg -pix_fmt yuvj422p -q:v 2
[22:09] <llogan> i don't think i've ever encoded with mjpeg, so i'm just guessing
[22:10] <brontosaurusrex> thats exactly the cli i used, but i guess quality switch does not kick in
[22:10] <llogan> does it actually look shitty?
[22:10] <brontosaurusrex> yes
[22:11] <llogan> maybe someone else has a better idea
[22:11] <brontosaurusrex> another unrelated puzzle is how the hell did i get this 1906x1080 resolution ....
[22:11] <brontosaurusrex> thanks llogan
[22:11] <llogan> you could always declare an arbitrary -b:v, but i'd assume -q:v should have worked.
[22:12] <brontosaurusrex> yeah, but quality based switch would be actually usefull
[22:12] <llogan> yes
[22:13] <brontosaurusrex> well, i guess its late, iam making one silly mistake after another, tommorow
[22:18] <llogan> brontosaurusrex: seemed to work for me
[22:19] <brontosaurusrex> -q:v 2?
[22:19] <llogan> yes
[22:19] <brontosaurusrex> hmmm
[22:19] <brontosaurusrex> ok, will do more testing
[22:20] <llogan> of course i didn't try the output in qiucktime
[22:20] <llogan> maybe it's being retarded like usual
[22:24] <brontosaurusrex> i see
[22:24] <brontosaurusrex> later
[00:00] --- Thu Jan 24 2013