[Ffmpeg-devel-irc] ffmpeg.log.20130406

burek burek021 at gmail.com
Sun Apr 7 02:05:02 CEST 2013


[00:09] <brimestone> hey guys, i have this file ( http://pastebin.com/uSTKiLw2 ) its generated by a ( 5DMK3 ) with avc1/H.264 (high) with 23.98 psf as FPS.. i need to add telecine to 30003/1001 or 29.97 But, i can only apply it if im sending it to a ProRes codec which is intruducing a color shift. can anyone help?
[00:15] <durandal_1707> brimestone: what you want to solve? color shift?
[00:17] <Costin> hey guys, i've solved it
[00:17] <Costin> i believe the key parameter was -probesize 32
[00:17] <Costin> that makes the encoding start very quickly
[00:17] <brimestone> yes.. the color shift..
[00:18] <brimestone> its some combination of contrast and brightness
[00:19] <durandal_1707> brimestone: perhaps prores encoder expect different color space, so for this try colormatrix filter
[00:20] <brimestone> i did add -vf colormatrix=bt601:bt709
[00:20] <brimestone> with no difference..
[00:20] <brimestone> optimally, it would be great to copy same codec from the input file..
[00:21] <brimestone> i tried -vcodec copy hehe but then filters will not apply, especially telecine
[00:22] <durandal_1707> brimestone: something like that is impossible right now
[00:22] <brimestone> something like what?
[00:23] <durandal_1707> brimestone: combination of copy and filter to produce same quality
[00:23] <durandal_1707> brimestone: have you tried swapping args for colormatrix?
[00:24] <brimestone> ohh yeah.. i sure would love to figure out how to keep color space
[00:24] <durandal_1707> or adding colormatrix fitler _before_ telecine?
[00:24] <brimestone> -vf colormatrix=bt709:bt601??
[00:24] <brimestone> testingf
[00:25] <brimestone> oh.. how can i add multiple filter arg in a single -vf switch?
[00:27] <brimestone> hmmm looks ok
[00:28] <brimestone> http://pastebin.com/gz6Nmxxr
[00:30] <durandal_1707> brimestone: that is not complete uncut console output, btw did you get it working or?
[00:31] <durandal_1707> also you dont need to put multiple -vf you can use "," and single -vf
[00:31] <brimestone> thats the enter cli arg...
[00:32] <durandal_1707> but it contains 0 useful info for me
[00:32] <brimestone> ohh wait
[00:35] <brimestone> http://pastebin.com/rAGgNk3J
[00:37] <durandal_1707> brimestone: why you are upconverting & transcoding?
[00:37] <durandal_1707> conversion from h264 to prores is pointless
[00:37] <brimestone> ops!?
[00:38] <brimestone> because i can't sticl the telecine on h264
[00:38] <durandal_1707> even so h264 use yuvj420p and prores hd pixel format
[00:38] <durandal_1707> burek: sticl is what?
[00:38] <brimestone> I'm completely open ears right now ( well technically Eyes )
[00:39] <durandal_1707> brimestone: what is sticl?
[00:39] <brimestone> more technical reason why i picked prores was to get it into Avid "easier"
[00:39] <brimestone> sorry, i meant Stick...
[00:40] <durandal_1707> but why to do that at first place with altread commpressed stream?
[00:40] <durandal_1707> *already
[00:40] <brimestone> i got faster than realtime trancoding ProRes files into Avid DNxHD ( in software ) compared to H264
[00:41] <durandal_1707> i'm really confused now
[00:41] <brimestone> :)
[00:41] <brimestone> let me paint the entire picture,
[00:42] <brimestone> if you have time
[00:43] <brimestone> Footages gets to my door in H264 form, from DSLR type camera.. then i have to use these footage in Avid, Resolve, After Effects? etc ( excluding Final Cut! )
[00:43] <brimestone> so what we, do is to prep these footage. Add timecode, Serialize the filename for tracking and archival
[00:44] <brimestone> then get it into a "workable" format for the editors to work with.. ( we use DNxHD )
[00:44] <durandal_1707> you tried colormatrix before and after telecine filter?
[00:45] <brimestone> now, as much as we smack the "directors" around, most of the time, they can't get the correct parameters for the video..
[00:45] <brimestone> we asked them footage in 29.97 they shoot in 23.98 so we have to add telecine before we make them DNxHD
[00:45] <durandal_1707> thing is first video h264 is reported as yuvj (this is extended jpeg crap)
[00:45] <brimestone> :)
[00:46] <brimestone> thanks to Canon and the Hype of "DSLR" Filmaking
[00:46] <durandal_1707> and prores accepts yuv only
[00:46] <brimestone> so my problem is yuvj420p to yuv422p?
[00:46] <durandal_1707> telecine, dunno for it as it is in another lib
[00:47] <brimestone> it does a good job. ( looking forward to get it in ffmpeg native lib )
[00:47] <durandal_1707> yuv422p10 use more bits than yuvj420p
[00:47] <durandal_1707> per pixel
[00:47] <brimestone> yes.. so what it does is filling up the "empty" space with prolly noise
[00:48] <durandal_1707> dunno, its 12 bit to 20 bit (different bitdepth and subsampling)
[00:48] <brimestone> Avid does a good color management from yuvj420p to DNxHD ( tried it ) only issue is that i can't add pulldown to it..
[00:49] <brimestone> i want the file to have a 29.97 fps before avid see's it ( Workflow )
[00:59] <durandal_1707> brimestone: so h264 is in bt601 ?
[01:00] <brimestone> that was a guess
[01:01] <brimestone> also.. is there another encoder for h264? mine always defaults to Stream #0:0 -> #0:0 (h264 -> libx264)
[01:02] <durandal_1707> that is only one, why?
[01:02] <brimestone> looking at ffmpeg -format  i see DE h264            raw H.264 video
[01:02] <brimestone> which i assume to be "Decoder/Encoder"..
[01:03] <brimestone> well.. when i open the file converted via FFmpeg  it physically say libx264 instead of H264
[01:04] <brimestone> not saying that libx264 is of no quality but for me, it avoids "question" as to why is not different now
[01:04] <durandal_1707> brimestone: -format is for demuxer/muxer
[01:04] <durandal_1707> -codecs is for decoder/encoders
[01:04] <brimestone> ohhh never use -codecs before
[01:05] <durandal_1707> brimestone: you need to conver to bt601 and not from
[01:05] <brimestone> wait.. i just testing one without color matrix
[01:06] <durandal_1707> for name of source i dunno, try bt709 or fcc or smpte240m
[01:31] <brimestone> only thing that stays the same is the change in color =)
[01:32] <durandal_1707> brimestone: does color shit happens without telecine filter?
[01:33] <brimestone> yes
[01:34] <brimestone> but if i use -vcodec copy -acodec copy then no shift is introduce
[01:35] <durandal_1707> brimestone: because you do no conversion to prores
[01:35] <durandal_1707> its not first time i hear that colorspace issue with prores
[01:35] <brimestone> even with H264 to livx264
[01:35] <durandal_1707> that is same
[01:35] <brimestone> well color still change
[01:36] <durandal_1707> than you need to add colormatrix
[01:36] <brimestone> hmmm
[01:36] <brimestone> :)
[01:36] <durandal_1707> what prores encoder you use?
[01:36] <durandal_1707> (there are two)
[01:36] <brimestone> standard
[01:37] <durandal_1707> try prores_kostya
[01:37] <brimestone> -vf colormatrix=bt601:bt709,'mp=telecine,setpts=N/(30000/1001*TB)
[01:37] <brimestone> sorry wronte paste
[01:40] <llogan> refer to -codecs or -encoders
[01:41] <brimestone> do i need to change the colormatrix?
[01:42] <durandal_1707> brimestone: for h264 to h264 probably not
[01:42] <brimestone> yeah.. so thats why im doing
[01:44] <brimestone> http://pastebin.com/rrk26tZx
[01:46] <brimestone> can i show you guys the color diff?
[01:47] <brimestone> https://philocast.lieberman.com/diff.png
[01:47] <brimestone> Right is transcoded via FFmpeg and left if the original from camera
[01:50] <brimestone> then using the same files imported into Avid i guess these color shifts.. https://philocast.lieberman.com/diff2.png
[01:55] <sacarasc> brimestone: That means it can encode the container, not the stream.
[01:56] Action: brimestone grasping to understand that sentence? 
[01:58] <mark4o> brimestone: try tagging the h.264 with the color standard you are using when you encode it: -x264opts colorprim=bt709:transfer=bt709:colormatrix=bt709  for example (adjust to whatever standards you are using for color primaries, color transfer, and color matrix if not bt.709 HD standard)
[01:59] <mark4o> some programs look at those settings but others ignore them
[02:00] <brimestone> testing
[02:06] <brimestone> how do i use those tags from a file? say from ffprobe?
[02:11] <mark4o> brimestone: I don't think ffprobe uses them.  x264 can write them and some programs (including apple quicktime) pay attention to them when it reads it and converts it to the color profile of the display.  I don't actually know if avid uses them.
[02:11] <brimestone> i think they do..
[02:11] <brimestone> i want to copy the flags from the original file and apply it to my trancoded file
[02:12] <mark4o> I don't see it in the ffprobe output, not sure there is a way to get it from ffprobe, although mediainfo shows them
[02:13] <brimestone> thanks i see it
[02:14] <sacarasc> -format is all containers.
[02:15] <brimestone> now i can copy everything on the originatin file
[02:49] <sacarasc> brimestone: What says libx264?
[02:49] <sacarasc> And where?
[02:52] <brimestone> sorry.. i got distracted.
[02:58] <brimestone> here we go.. http://pastebin.com/d3zF9eye
[02:59] <brimestone> sorry for the delay
[03:59] <relaxed> burek: Add that fflogger is a bot, because people are increasingly responding to it. :)
[04:01] <mark4o> maybe change its name to ffbot?
[04:56] <gp> Is there anywhere I can read about how to use the idet filter?
[05:04] <relaxed> gp: did you look at "man ffmpeg-filters"
[05:05] <gp> I have.  I mean more about how I interpret it
[05:10] <relaxed> gp: from outputting to -f null I see "[Parsed_idet_0 @ 0x2795e60] Multi frame detection: TFF:1943 BFF:4 Progressive:0 Undetermined:0" which probably means it's top field first.
[05:11] <gp> That makes sense.  That video is pretty straight forward
[05:12] <gp> I've been seeing a lot of Undetermined
[05:13] <gp> I will play around with it some more and see if I can come up with some more specific questions
[05:22] <gp> relaxed: Thanks for pointing that out.  I think I was over complicating it.  It actually seems pretty straightforward.  I think I just had a bad test file last time I looked into it or something
[05:26] <gp> Any idea how many frames I would need to inspect to get a pretty good guess?
[05:36] <gp> So if I am trying to determine if I should enable yadif or not using idet, would I just want to add TFF and BFF and compare them against Progressive?  If the sum of TFF and BFF is larger than Progressive I would deinterlace?
[05:54] <relaxed> gp: on progressive content I get "TFF:0 BFF:0 Progressive:198 Undetermined:47"
[05:55] <relaxed> to answer your question, I'd say yes
[05:56] <gp> relaxed: much appreciated
[09:37] <post_janitor> hey guys.. how do i encode to h264 High at L5.1?
[12:36] <mastro> I want to convert a VOB file into an MP4 with hard-coded subtitles (and choosing a channel)... googling I found some command but I'm not satisfied of commands given without explanation.
[12:36] <mastro> For instance: ffmpeg -i input.vob -acodec libfaac -ac 2 -ab 128k -vcodec libx264 -vpre fast -crf 20 -threads 0 output.mp4
[12:38] <mastro> I understand it is setting audio codec faac, channel 2, audio bitrate 128k, video codec x264, -vpre fast dunno what's that for, -crf 20 neither, -threads ? ....
[12:38] <mastro> But I'd like to be able to choose those value on my own, and I probably need to analyze the source vob file to avoid produce a too-high quality or too-low one
[12:39] <mastro> what do you usually do to choose those parameters?
[12:46] <relaxed> -threads 0 tells libx264 to use all your cpu cores for encoding.
[12:47] <relaxed> you can't hardsub vobsubs as far as I know.
[12:47] <relaxed> mastro: ^
[12:47] <LithosLaptop> I always shiver when I see libfaac
[12:48] <relaxed> put a sweater on
[12:48] <LithosLaptop> :)
[12:48] <mastro> relaxed, thanks
[12:49] <relaxed> can't using ffmpeg, I meat
[12:49] <mastro> LithosLaptop, I've not signed anything forcing me to use faac, that just what I found googling :)
[12:49] <relaxed> meant*
[12:49] <LithosLaptop> hehe ok
[12:49] <relaxed> libfdk-aac is recommended
[12:50] <LithosLaptop> yeah
[12:50] <mastro> relaxed, what should I use? mencoder?
[12:50] <mastro> LithosLaptop, why is that? I mean, why one should prefer libfdk-aac to libfaac ?
[12:51] <mastro> LithosLaptop, I assume faac is "free" and fdk-aac isn't
[12:51] <relaxed> libfdk-aac produces better quality and has more features.
[12:51] <LithosLaptop> mastro: libmp3lame which is MP3 produces better quality then libfaac which is AAC
[12:52] <LithosLaptop> but generally AAC should be better quality than MP3
[12:52] <mastro> relaxed, can I use ffmpeg to build an mkv containing both audio channel and the subtitles?
[12:52] <relaxed> neither libfdk-aac or libfaac are gpl compatible.
[12:52] <LithosLaptop> libfaac isn't transparent relaxed: for met at 128Kbit/s
[12:52] <LithosLaptop> oops
[12:52] <LithosLaptop> I accidently put your name in there
[12:52] <relaxed> mastro: yes
[12:53] <relaxed> ffmpeg can only hardsub ass/srt as far as I know.
[12:53] <mastro> relaxed, that would be an option.. I'll try to google something.. thanks for the suggestion about libfdk-aac I'll look for it
[12:53] <mastro> LithosLaptop, ^
[12:53] <mastro> what about the -vpre fast ?
[12:53] <mastro> and -crf 20 ?
[12:54] <Hans_Henrik> i can't find "-level" argument in docs.. but some example for "ipad format conversion" use -level 30       .. any1 know what it's for?
[12:54] <Hans_Henrik> full is ffmpeg -i "%%A" -acodec aac -ac 2 -strict experimental -ab 160k -s 1920x1080 -vcodec libx264 -preset slow -profile:v baseline -level 30 -maxrate 10000000 -bufsize 10000000 -b 1200k -f mp4 -threads 0 "%%A.ipad.mp4"
[12:54] <mastro> oh.. vpre = preset
[12:55] <relaxed> Hans_Henrik: it limits the h264 features used
[12:55] <mastro> crf I found it in the man but do not understand what's for
[12:55] <relaxed> certain hardware only supports specific levels
[12:55] <relaxed> constant rate factor
[12:56] <relaxed> https://ffmpeg.org/trac/ffmpeg/wiki/x264EncodingGuide
[12:56] <mastro> can I list presets / read their documentation?
[13:00] <mastro> relaxed, that's a very useful link
[13:01] <mastro> relaxed, is there something like that for MKV container and fdk-aac encoder ?
[13:01] <mastro> relaxed, afk 1hour (I'll surely read what you reply anyway, thanks)
[13:12] <Vardan> hi all
[13:12] <Vardan> people I have compiled ffmpeg for android, but when I'm trying to use avcodec_find_encoder(CODEC_ID_MJPEG) it return 0
[13:15] <durandal_1707> Vardan: what version of libavcodec?
[13:15] <Vardan> 0.8.6
[13:16] <durandal_1707> that is very old and not libavcodec version....
[13:16] <Vardan> new version not compiling with Android NDK
[13:17] <durandal_1707> avcodec_find_encoder_by_name("mjpeg") does not work?
[13:35] <Vardan> ok, now it's find CODEC_ID_MJPEG encoder but couldn't open :(
[13:38] <durandal_1707> Vardan: what new version and from where you tried when commpiling it?
[13:39] <Vardan> not a new version the same 0.8.14 version
[13:39] <durandal_1707> Vardan: you said that new version does not compile
[13:39] <Vardan> downloaded from ffmpeg.org
[13:39] <durandal_1707> trying to get help for 0.8.14 is harder because its very old
[13:41] <Vardan> ok, let me try to compile latest
[13:41] <durandal_1707> but anyway that error should not happen
[13:48] <misterli> Hello. Ive a question about ffmpeg. Im using PHP to convert Videos from 3gp to mp4. i've about 200 converts within 30 minutes and in most cases it works well. Anyway, i some cases ffmpeg stuck and shoot up the CPU L oad to 100%. After that all other ffmpeg instances will try to get 100% CPU Load. Yesterday i'd see that a FFMPEG Process took more then 2hours and didnt do any work (just 100%CPULoad).
[13:48] <misterli> Is there any way to set a timeout for 1 instance?
[13:53] <Vardan> when I'm trying to compile latest version got this message: "{standard input}:9962: Error: selected processor does not support ARM mode `ubfx r7,r2,#0,#4'"  ... make: *** [libavcodec/aacdec.o] Error 1
[13:58] <durandal_1707> Vardan: the ony way this can get fixed is if you report that bug to bug tracker
[13:59] <durandal_1707> misterli: there is no enought info provided by you to provide and useful help
[13:59] <durandal_1707> misterli: what ffmpeg version?
[14:10] <misterli> its ffmpeg version 0.7.15,
[14:11] <durandal_1707> misterli: very old (contains only security fixes by now)
[14:11] <durandal_1707> misterli: also did you explore if it hang on certain file
[14:12] <durandal_1707> otherwise timeout and such are configured at OS level
[14:12] <misterli> There are too many file to explore the file which causes the stucking
[14:12] <durandal_1707> and i dont see how good would be to implement it inside ffmpeg
[14:12] <durandal_1707> misterli: than if stucking was ever caused by bug in code it would not get fixed
[14:24] <mastro> relaxed, thanks anyway :)
[14:34] <mastro> when using ffmpeg -map to build a matroska. Is there a way to specify the name of an audio channel?
[14:35] <durandal_1707> audio channel?
[14:38] <mastro> durandal_1707, yes
[14:38] <mastro> durandal_1707, like: channel 1: en, channel 2: it, channel 3: jp etc...
[14:42] <durandal_1707> that is more like stream
[14:43] <durandal_1707> and language for each stream
[14:44] <mastro> durandal_1707, -alang en / -alang it / -alang ja
[14:45] <durandal_1707> it matroska track language
[14:46] <durandal_1707> mastro: you set "language" metadata for such stream/track
[14:46] <mastro> in a syntax like the one described here: http://ffmpeg.org/trac/ffmpeg/wiki/How%20to%20use%20-map%20option   -c:v , -c:a:1 etc... how would I set -alang? -alang:a:0 it ?
[14:47] <durandal_1707> alang is nowhere in help output? from where you get alang?
[14:47] <durandal_1707> mastro: you set track language, in same way as you set any other metadata
[14:47] <mastro> durandal_1707, googling
[14:48] <mastro> durandal_1707, ok, how do I set a metadata?
[14:48] <durandal_1707> mastro: there is documentation, i'm not paid to give you exact answer because i have other things to do
[14:49] <mastro> durandal_1707, sorry I didn't know about metadata found it now, thanks
[14:50] <durandal_1707> http://ffmpeg.org/ffmpeg.html
[18:30] <Shingetsu> hello. I'm trying to direct x11 output (x11grab) into a v4l2 device (v4l2loopback). " sudo ffmpeg -f x11grab -i :0.0 -s 600x400 -f v4l2 - > /dev/video1 " gives out "[NULL @ 0x1cd53e0] Requested output format 'v4l2' is not a suitable output format". if I use null as a format, it's not readable as a result
[18:31] <Shingetsu> can anyone please direct me into the proper way of doing this?
[18:32] <ubitux> there is no v4l2 output support yet
[18:32] <ubitux> patch welcome
[18:33] <Shingetsu> @ubitux is there any feeder capable of v4l2 output that I could look into for making the patch?
[18:34] <ubitux> gstreamer has an output support
[18:34] <ubitux> otherwise, just look at the v4l2 api documentation
[18:34] <ubitux> i think v4l2loopback also has an example for pushing stuff
[18:34] <Shingetsu> is v4l2 output planned in the future?
[18:35] <ubitux> i don't think anyone is working on it currently
[18:35] <ubitux> but it might be wise to announce you're working on it on the devel mailing list
[18:35] <ubitux> maybe ask __gb__ on #ffmpeg-devel too
[18:36] <Shingetsu> I'm not entirely sure I will~ I have a lot of projects going on already
[18:36] <Shingetsu> I'll look and see if any quick solution is possible, if not I'll try and take a look (in which case I'll talk to __gb__)
[18:39] <Shingetsu> alright, thanks~
[19:17] <keyzs> sirs what seems to be the error here? http://pastebin.mozilla.org/2279240
[19:20] <ubitux> you should upgrade.
[19:20] <keyzs> the ffmpeg?
[19:20] <ubitux> yes
[19:20] <keyzs> ok
[19:21] <ubitux> compressed swf support was added months ago
[19:21] <ubitux> we are in ffmpeg 1.2
[19:21] <ubitux> you are using ffmpeg 0.5..
[19:21] <keyzs> let me upgrade this then
[19:21] <ubitux> keyzs: note that debian doesn't distribute ffmpeg anymore
[19:21] <ubitux> and the fork they distribute, even if bleeding edge, will not have compressed swf support
[19:22] <keyzs> ubitux, but are the commands correct?
[19:22] <ubitux> should be ok, somehow
[19:22] <ubitux> too much stuff changed anyway
[19:22] <ubitux> you are using ffmpeg 0.5, i don't think you realize how old it is
[19:23] <ubitux> and how much changes happened since then :p
[19:23] <keyzs> i installed via apt-get
[19:23] <ubitux> 0.5 was released in 2009
[19:23] <ubitux> it's 4 years old now
[19:23] <ubitux> basically you're doing archeology right now
[19:26] <keyzs> http://ffmpeg.gusari.org/static/32bit/ffmpeg.static.32bit.2013-02-01.tar.gz is this the right package for bt5 r3?
[19:34] <ubitux> i have no idea what bt5 r3 is
[19:36] <dot> hello i need help. I have files like this image_01210.png up to image_0748.png i would like to creat a movie: ffmpeg -f image2 -r 25 -i image_%4d.png -vcodec mpeg4 -qscale 0 video.avi but it says Could find no file with with path 'image_%4d.png' and index in the range 0-4 can someone help me please
[19:37] <durandal_1707> dot: there is option to set start number in recent ffmpeg version
[19:37] <durandal_1707> -start_number 748
[19:37] <dot> aha
[19:38] <dot> durandal_1707: so i just add -start_number 748
[19:38] <durandal_1707> dot: yes, after -f image2
[19:39] <dot> durandal_1707: thank you it works
[19:47] <dot> durandal_1707: is it possible to delete video on one side
[19:47] <dot> ?
[19:47] <dot> i mean cut out video
[19:47] <dot> i expressed myself strange. to shorten width so i cut some part off
[19:48] <dot> trim video width ::)
[19:53] <dot> aha i found it crop
[19:53] <dot> :)
[19:56] <keyzs> ubitux, i upgraded and says that compressed format of the swf is not supported http://pastebin.mozilla.org/2279328, can anything be done?
[19:56] <ubitux> 0.10 is very old.
[19:56] <ubitux> upgrade to 1.2 or git/master
[19:57] <keyzs> guess this ppa aint correct way
[19:57] <keyzs> i should install a static build?
[19:57] <ubitux> no need to install
[19:57] <ubitux> you can use it out of the box
[19:59] <keyzs> how is it done
[20:02] <ubitux> ./ffmpeg
[20:04] <keyzs> git clone git://source.ffmpeg.org/ffmpeg.git ffmpeg then ./ffmpeg
[20:29] <ubitux> keyzs: no, if you don't use the static build, ./configure --enable-<whatever-you-need> && make
[20:29] <ubitux> then ./ffmpeg
[20:29] <ubitux> see the wiki for help ^
[20:43] <keyzs> ubitux got this http://pastebin.mozilla.org/2279416
[20:44] <ubitux> keyzs: can i see the sample?
[20:44] <ubitux> [swf @ 0xb4ef5c0] Could not find codec parameters for stream 0 (Video: vp6f): unspecified size
[20:44] <keyzs> sure
[20:44] <ubitux> this seems to be the problem
[20:46] <keyzs> its the first time i touch ffmpeg and i dont know much on video
[20:48] <ubitux> might be a bug or something unsupported
[20:49] <ubitux> i'm looking at it
[20:50] <ubitux> it expects 5 unknown tags
[20:52] <ubitux> keyzs: is the file supported somewhere?
[20:52] <ubitux> i'm not able to play it in my browser with the flash player
[20:53] <keyzs> this file is from a cd i have i´ve been for hours tryong to convert it
[20:53] <keyzs> *trying
[20:54] <keyzs> all softwares were failing so i decided to try on linux with ffmpeg
[20:54] <ubitux> well, are you even able to play it somewhere?
[20:54] <ubitux> the official flash player doesn't seem to do anything
[20:55] <keyzs> ubitux im gonna copy the file again
[20:57] <keyzs> it played on my browser in less than a second
[20:57] <keyzs> like 0.5seg
[20:57] <ubitux> strange, doesn't work here
[20:57] <keyzs> let me grab bigger file
[20:57] <keyzs> on vlc doesnt go
[20:57] <ubitux> fc5df85c93135a7eaa5deef7d332c484  samples/colunaVertebral_grandedorsal_decubitoventral2.swf
[20:58] <ubitux> opening this with firefox or chromium doesn't give me anything
[20:59] <keyzs> give me just a minute i? copying more swf files
[21:00] <keyzs> ubitux now i get why conversion wasn´t going good
[21:00] <keyzs> this cd has a protection maybe it interferes on the file
[21:00] <ubitux> :)
[21:02] <keyzs> ubitux you gave me awesome help
[21:02] <keyzs> i was beeing dumb
[21:02] <keyzs> and didn´t realize this
[21:03] <ubitux> are you able to get it working?
[21:04] <keyzs> via browser it flashes a image
[21:04] <keyzs> swf doesnt play on vlc here
[21:04] <keyzs> none of them
[21:12] <keyzs> ubitux, shouldn´t this play?
[21:13] <ubitux> "this"?
[21:18] <keyzs> yes the file
[21:22] <ubitux> as i said i'm not even able to play it in my browser
[21:31] <keyzs> yes thats why i asked, shouldn´t it play normally?
[21:31] <keyzs> i have more files here from the cd they do same thing
[21:35] <ubitux> if the file doesn't play with anything, even the official flash player, then something is definitely broken
[21:35] <ubitux> and i can't do much about it in ffmpeg
[21:37] <Demon_Fox> Can video get longer if it is not a standard framerate
[21:37] <Demon_Fox> like 24.22FPS instead of 23.976?
[21:38] <brontosaurusrex> Demon_Fox, shorter in that case
[21:38] <Demon_Fox> I am converting something right now that has a length of 25m13s
[21:39] <Demon_Fox> and right now ffmpeg is encoding 25m49s
[21:39] <Demon_Fox> still going
[21:39] <Demon_Fox> or maybe part of the video was corrupted
[21:40] <Demon_Fox> and ffmpeg is picking up what VLC can't, which would not be the first time
[21:41] <Demon_Fox> wait, vlc is picking it up, but its estimate was wrong
[21:41] <Demon_Fox> wait
[21:41] <Demon_Fox> It still says 25:13
[21:41] <Demon_Fox> It ends at exactly 25:13
[21:57] <brontosaurusrex> whats the most popular aac-enc on linux this days?
[21:58] <brontosaurusrex> i mean in ffmpeg
[21:59] <klaxa> libfdk-aac
[22:07] <brontosaurusrex> ok
[22:12] <brontosaurusrex> Demon_Fox, vlc sucks big time
[22:12] <brontosaurusrex> if you need alternative media info, use mediainfo
[22:12] <Demon_Fox> vlc must be broken on my system
[22:13] <brontosaurusrex> what os?
[22:13] <Demon_Fox> considering that it is counting seconds almost half as fast on one as the other
[22:13] <Demon_Fox> brontosaurusrex, Fedora on this machine
[22:13] <Demon_Fox> Slackware on another
[22:13] <klaxa> someone else once had sync problems with vlc
[22:13] <Demon_Fox> and arch on other
[22:14] <brontosaurusrex> yeah, iam on debian and my solution was apt-get remove --purge vlc
[22:15] <Demon_Fox> The only thing that ffmpeg is doing that is annoying is increasing the framerate to 60FPS instead of using the nonstandard one
[22:15] <Demon_Fox> more like
[22:16] <Demon_Fox> 60000/1001
[22:16] <Demon_Fox> so slightly more than half the frames repeat
[22:16] <brontosaurusrex> cant you force that?
[22:17] <Demon_Fox> I think it is ffmpegs way of compensating for the nonstandard framerate, so it can be compliant to blu ray stuff.
[22:18] <Demon_Fox> that is atleast the thought
[22:18] <Demon_Fox> at least*
[22:19] <brontosaurusrex> i doubt that
[22:20] <Demon_Fox> It might be x264 that is doing it
[22:20] <Demon_Fox> I am using the client directly at it is saying 60000/1001 for the FPS
[22:20] <klaxa> you can always force a framerate
[22:20] <klaxa> ah
[22:20] <brontosaurusrex> using a bluray preset in x264?
[22:21] <Demon_Fox> no
[22:22] <Demon_Fox> I don't think that is a preset but more like a profile that forces compliance by setting limits
[22:28] <Demon_Fox> Unless interlaced material is different
[22:29] <Demon_Fox> which it should not be
[22:32] <Demon_Fox> that might be it
[22:32] <Demon_Fox> if ffmpeg is decoding the video and making it seem interlaced
[22:32] <Demon_Fox> Maybe not though
[22:33] <klaxa> it's interlacing the video?
[22:33] <klaxa> wut?
[22:36] <Demon_Fox> I think I fixed it
[00:00] --- Sun Apr  7 2013


More information about the Ffmpeg-devel-irc mailing list