[Ffmpeg-devel-irc] ffmpeg.log.20121217

[burek]

[00:37] <nafiux_> Hi, I have problems with ffmpeg: ffmpeg[12810]: segfault at 35dd000 ip 00007f90fd04ed6a sp 00007fffe43995e8 error 6 in libc-2.12.so[7f90fcfc1000+189000]
[00:56] <maujhsn> Thanks to ubuntu linux programs like "ffmpeg", "kdenlive", "Audacity" & "lib-visual" your videos can be as imaginative as this: https://www.youtube.com/watch?v=9c6AV1tElU4 Enjoy it!
[00:59] <juanmabc> nope, techno jazz is not my thing ;
[00:59] <juanmabc> P
[01:02] <nafiux_> Hi, I have problems with ffmpeg: ffmpeg[12810]: segfault at 35dd000 ip 00007f90fd04ed6a sp 00007fffe43995e8 error 6 in libc-2.12.so[7f90fcfc1000+189000]
[01:03] <nafiux_> Can you help me?
[01:05] <nafiux_> solved :)
[01:17] <djszapi_> juanmabc: hah!
[01:39] <LooX> hi
[01:41] <LooX> i am trying this command: ffmpeg1 -i file.mkv -map 0:0 -map 0:5 -scodec mov_text -vcodec copy -acodec copy file.mov
[01:42] <LooX> but get  segmentation fault (core dumped)
[01:42] <nafiux_> LooX
[01:43] <nafiux_> LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH
[01:43] <nafiux_> export LD_LIBRARY_PATH
[01:43] <nafiux_> and next try your example
[01:44] <LooX> same
[01:44] <LooX> echo $LD_LIBRARY_PATH  .... /usr/local/lib:
[01:45] <nafiux_> Ok, type: tail /var/log/messages
[01:46] <LooX> kernel: pid 3682 (ffmpeg1), uid 1001: exited on signal 11 (core dumped)
[01:50] <LooX> problem is with subtitiles.. it copies ok audio/video
[02:11] <burek> is google down for you guys too, or it works normally?
[02:14] <burek> nafiux_, next time provide a pastebin of output log, not just one line of error, because it's not very useful to us (if you want us to help you resolve issues)
[02:14] <nafiux_> Ok burek!
[02:18] <LooX> i already pasted the relevant parts.. but here it is anyways: http://pastebin.com/zMUQy07j
[02:18] <djszapi_> would it be a hard job to implement the lagarith encoder?
[02:20] <burek> LooX, could you try this http://ffmpeg.org/bugreports.html
[02:22] <burek> djszapi_, are you sure it's not already implemented?
[02:24] <djszapi_> burek: yes
[02:24] <djszapi_> only the decoding.
[02:24] <djszapi_> http://ffmpeg.zeranoe.com/forum/viewtopic.php?f=7&t=649
[02:24] <burek> oh ok
[02:24] <burek> well, then wait for some time
[02:24] <burek> someone will come up with the patch
[02:25] <djszapi_> kinda cannot.
[02:25] <djszapi_> we need such a feature for our product.
[02:25] <djszapi_> that is why I am trying to ask how complex such a patch would be. :)
[02:26] <djszapi_> people have been already waiting for a long time. :-)
[02:27] <burek> if you "need" it, you can always try to hire an ffmpeg developer
[02:27] <burek> to implement it for you
[02:27] <djszapi_> without budget?
[02:27] <burek> nobody is willing to spend their time to help your product for free
[02:27] <djszapi_> no one asked ...
[02:27] <djszapi_> so let me ask the same question again the third time:
[02:28] <djszapi_> how complex would such a feature be to implement?
[02:29] <burek> i dont know, nor do i care that much tbh :)
[02:29] <LooX> dunno if this helps.. i dont have a debug build: http://pastebin.com/VTHW1A38
[02:29] <djszapi_> and actually many people spent their time to help others' product.
[02:30] <burek> LooX, could you provide a sample media
[02:30] <burek> so that developers can reproduce the issue?
[02:31] <burek> djszapi_, yes, but when they find some free time and will to do it
[02:31] <burek> which again means, you have to wait
[02:31] <burek> or rent-a-coder
[02:31] <burek> :)
[02:32] <foonix> or use ffv1/ffvhuff ;)
[02:32] <LooX> yes..
[02:32] <LooX> now i have to cut the file
[02:32] <djszapi_> burek: not really.
[02:32] <djszapi_> I also implemented stuff when others requested, several times.
[02:33] <burek> LooX, ffmpeg -i file -map 0 -c copy -t 60 output
[02:33] <djszapi_> it is a usual daily situation for open source projects actually ...
[02:33] <burek> djszapi_ I believe you but still :)
[02:33] <djszapi_> so "nobody" is a heft exaggeration.
[02:34] <djszapi_> hefty*
[02:34] <burek> djszapi_ dont blame me for pointing out the obvious
[02:34] <djszapi_> obvious wrong, I do blame.
[02:34] <burek> if I were wrong, you wouldn't be here asking the same question for the 3rd time, would you? :)
[02:35] <djszapi_> I think you are just trolling.
[02:35] <djszapi_> the question asked three times is entirely different kind.
[02:35] <burek> ok, then just ignore me
[02:35] <burek> :)
[02:42] <LooX> ok.. have a 5mb file that crashes ffmpeg .. with 1 of the text streams
[02:42] <LooX> the other 2 work fine
[03:57] <Cpt-Oblivious> Can I use ffmpeg to stream a lot of .mkv files? Say i got 1000 .mkv files on my hard disk, can i use ffmpeg to stream those? Say, i point it to the file on the harddisk with the request, can ffmpeg then start streaming that?
[03:59] <sacarasc> ffserver might be able to, but I am unsure.
[11:01] <lundtor> I have a 1080x1080 avi file. I need to add black borders on each side (420px) to make it 1920x1080. i have looked at the pad option, but i fail to configure it (i have used -vf 420:1080:0:0:black) to add the left border
[11:02] <lundtor> i get this error: Input area 0:0:1080:1080 not within the padded area 0:0:420:1080 or zero-sized
[11:04] <lundtor> http://pastebin.com/9jnf9Gfs
[11:14] <lundtor> found the solution: -vf pad=1920:1080:420:0:black
[11:55] <Zandman26> Hi all, can someone tell me if its possible to split my mic sound from my monitor sound during recording in ffmpeg, want to do this as I want to be able to fix eventual sound issues in post. Like screen audio drowning out mic or vice versa
[12:31] <Zandman26> My current setup can be found on: http://pastebin.com/cY7L66rU . So can I rewrite this in someway to split the audio from mic and monitor?
[12:36] <klaxa> Zandman26: wat
[12:36] <klaxa> define mic and monitor
[12:37] <klaxa> also, if you aren't recording lossless, don't use -preset ultrafast, if you are editing the resulting file afterwards don't upscale while recording
[12:38] <klaxa> also you would get one mkv with audio and video, you can't put pcm into mp3 (i think?) and if you are editing afterwards you can easily extract the audio track from the matroska file
[12:39] <Zandman26> Mic is microphone/boost channel on my AlsaMixer and monitor is PCM output
[12:40] <klaxa> you only have one audio input in that line
[12:40] <klaxa> and i think that's the mic
[12:40] <klaxa> i don't know how you get a monitor on some output in alsa :X
[12:40] <Zandman26> well it records both monitor sound and my mic with that setup
[12:41] <klaxa> in that case i think it's impossible to split it, because you only have one audio-stream
[12:41] <klaxa> are you sure your mic isn't recording somethink like your speakers or anything?
[12:42] <Zandman26> however I cant seam to split them, so I can edit sound levels of the PCM and Mic seperatly in post
[12:42] <klaxa> <klaxa> in that case i think it's impossible to split it, because you only have one audio-stream
[12:42] <klaxa> you have only one audio input stream (-f alsa -i hw:0,0)
[12:42] <klaxa> therefore you can't split it easily
[12:42] <Zandman26> :/
[12:42] <klaxa> did you extract the audiostream and open it in an audio editing tool? (i assume you have audacity?)
[12:43] <Zandman26> Yes but I get both PCM and Mic sound in the same channel
[12:44] <klaxa> in that case you can't split it, are you using headphones while recording?
[12:44] <klaxa> maybe your alsa setup is not optimal
[12:44] <Zandman26> Yes
[12:44] <klaxa> hmm i never managed to get alsa monitors, lol
[12:45] <Diogo> hi this is possible to stream cdg file to rtmp server??
[12:45] <klaxa> IF it doesn't add too much overhead you *might* want to try pulse, its abstraction makes it easier to handle these things
[12:45] <Zandman26> Im sure it is aint:) But this was the first thing that worked to both caputre PCM and Mic for me
[12:47] <klaxa> Diogo: if cdg is cdgraphics then i guess it should be possible
[12:51] <Diogo> i'm using this code..
[12:51] <Diogo> code: /servers/ffmpeg/bin/ffmpeg -re -i 1.cdg -i 1.mp3 -vcodec libx264  -acodec libmp3lame  -f flv "rtmp://RTMP_SERVER"
[12:52] <cbsrobot> Zandman26: isnt 0:0 the master channel ?
[12:52] <Diogo> but i can play the stream ..in the encoding...appear Multiple frames in a packet from stream 0
[12:52] <Diogo> [libmp3lame @ 0xa91580] Que input is backward in time
[12:52] <cbsrobot> open alsamixer and see on what input you have sound ...
[12:55] <Yulth> I folks!
[12:55] <Zandman26> in alsa mixer I have enabled capture on my soundcard and told it to record PCM + Mic + Mic Boost
[12:55] <Diogo> klaxa: something wrong in the command?
[12:55] <Diogo> thanks..
[12:56] <klaxa> Diogo i don't see anything actually, try using -acodec copy instead maybe? i mean it's already mp3, no need to reencode really
[12:58] <Yulth> I'm having quality issues transcoding from mp3 to HE-AAC: for basses, 40kbps HE-AAC sounds better than 64kbps HE-AAC, althoug the rest of middle and higher frequencies sounds better at 64kbps HE-AAC. I'm using libfdk_aac, of course. Any idea about how to fix this strange problem?
[13:00] <Zandman26> Tried captureing pulse, it tells me that it should start recording in stream 0:0 but then nothing happens and no errors:/ strange
[13:01] <klaxa> did you run it like this? ffmpeg -f alsa -i pulse ?
[13:03] <Zandman26> ffmpeg -f x11grab -s 720x470 -r 20 -i :0.0 -f alsa -ac 2 -i pulse -ab 192k -ar 44100 -c:v libx264 -preset ultrafast -s 1280x800 -c:a pcm_s16le -threads 2 minecraftCap.mkv aud01.mp3
[13:05] <Diogo> not working ...^Came= 2182 fps= 16 q=28.0 size=    2643kB time=00:02:05.72 bitrate= 172.2kbits/s dup=0 drop=8602
[13:06] <Diogo> drop=8602  ... :(
[13:08] <Zandman26> Here you have both my current setup for recording and the result from consolehttp://pastebin.com/wvvXs6Ez
[13:19] <klaxa> Zandman26 that one looks like it's missing a lot
[13:22] <Zandman26> klaxa: Maybe:P I have tried to understand ffmpeg during the last 2 weeks but still very new at it. The config of this installation is of YAST and packman repos so I hope it had a better clue of what it was doing then when I was trying to configure and build it my self:)
[13:36] <urand0m> Just got a 64GB SSD for my Samsung S3. What would be a good a decent resolution for that 4.8" screen?
[13:37] <urand0m> blah ... mean a SDXC cards
[13:37] <klaxa> how about the S3's display resolutionß
[13:37] <klaxa> ?
[13:37] <urand0m> Its crazy high.
[13:37] <urand0m> more then I need to say the least
[13:38] <urand0m> 1,280 x 720 pixels
[13:39] <klaxa> well... 720p it is?
[13:39] <urand0m> Would be nice to maybe get the movies down to lets say 250 ?
[13:39] <urand0m> MB that is
[13:40] <urand0m> yeah but I dont think a I will be able to see that kind of detail. 720p on a 4.8" screen?
[13:41] <klaxa> it will be super sharp, you'll love it
[13:41] <klaxa> if you appreciate high quality as much as i do
[13:47] <onryo> I can play around with it a bit. I wrote a program that does a 2 pass compression. Its asks that the total size of the video/sound file should be when done. Then asks if you want really nice sound or not.
[13:48] <onryo> Then it executes this line ffmpeg -i $title -an -vcodec libx264 -pass 1 -preset veryslow -threads 0 -b $videoBR -x264opts frameref=15:fast_pskip=0 -f rawvideo -y /dev/null && ffmpeg -i $title -acodec libfaac -b:a $audioBR -ac 2 -vcodec libx264 -pass 2 -preset veryslow -threads 0 -b $videoBR -x264opts frameref=15:fast_pskip=0 ${title%.*}.$totSizeMB.mkv
[13:48] <Yulth> hello! Any ideas about my last question? :)
[13:48] <Yulth> "I'm having quality issues transcoding from mp3 to HE-AAC: for basses, 40kbps HE-AAC sounds better than 64kbps HE-AAC, althoug the rest of middle and higher frequencies sounds better at 64kbps HE-AAC. I'm using libfdk_aac, of course. Any idea about how to fix this strange problem?"
[13:51] <onryo> Yulth I would help you out of I could but I don't know dick about sound =/
[13:53] <pierre-olivierro> Hi all, I got some questions about x264 encoding, threw ffmpeg -tune
[13:55] <Yulth> onryo: :)
[13:56] <pierre-olivierro> its about the use of av_packet_new_side_data stuff
[13:59] <onryo> pierre-olivierro You might have to ask this unruly mob to stop fighting and make a single file line to help you out :)
[13:59] <onryo> Or just ask and see if somebody helps you out =P
[14:01] <pierre-olivierro> onryo: in fact i konw how to use these functions, but in some case the output x264 stream is corrupted...
[14:02] <onryo> what does it say?
[14:02] <onryo> Or do ..
[14:03] <onryo> Is my little program worth its weight in electrons? Thinking about the last line there. http://pastebin.com/yTr4js1X
[14:04] <onryo> lol, this place is a real party!
[14:05] <pierre-olivierro> onryo:  it says : sps_id out of range, missing picture, non existing PPS ref ...
[14:06] <pierre-olivierro> onryo: but i notice that it seems to be related to the extradata size...
[14:07] <onryo> If I bump into stuff like that I normally try mkvmerge just to see if there is some weird stream or something.
[14:09] <onryo> Can you do some thing like -vcodec copy with -sn -an just to make sure it really is the the video?
[14:10] <onryo> no streams or audio that is.
[14:10] <burek> S3 has got 300 dpi ?
[14:10] <pierre-olivierro> yes, just a minute please...
[14:11] <burek> wtf.. almost as a printer
[14:12] <onryo> burek I guess it would be about 306 ppi -- 720 x 1280 pixels, 4.8 inche
[14:13] <zmbmartin> How can I create a video that has one image as the background then a number of transparent pngs overlayed with a .wav file as audio?
[14:13] <zmbmartin> I somewhat got it working with -loop but then it never stops until I hit q
[14:13] <onryo> burek yeah its crazy high. So I am kind of wondering what a good dpi would be. 75 dpi?
[14:13] <burek> 96.. 100
[14:13] <onryo> I guess like an old CRT?
[14:13] <onryo> kk
[14:13] <onryo> good to know
[14:14] <zmbmartin> I have been unsuccessful getting the audio in though
[14:15] <onryo> zmbmartin what container type are you using?
[14:16] <burek> Yulth, did you try neroaacenc
[14:16] <burek> or libaacplus
[14:16] <zmbmartin> onryo: Not sure what you mean. I am new to ffmpeg here is the command that I have sort of working but just keeps going until I hit q > 'ffmpeg -loop 1 -i original.jpg -r 5 -i image_%d.png -filter_complex overlay -shortest testvid.mp4'
[14:16] <burek> libfdk_aac is kinda new and i personally did not test it that much, but i did test libaacplus a lot and i can tell it's superb
[14:17] <pierre-olivierro> onryo: could you please copy pase a command line example to do what you want ?
[14:17] <onryo> Sure I wrote this http://pastebin.com/yTr4js1X
[14:17] <onryo> but I am not sure what rez I should use =)
[14:18] <onryo> but you more or less gave me a magic and sensible number with 100 dpi
[14:18] <burek> onryo, are you targeting file size or qualitu
[14:18] <burek> quality*
[14:18] <onryo> well both =)
[14:19] <burek> why are you using -b in 2-pass encoding
[14:19] <onryo> dang I did ????
[14:19] <burek> wait, let me check other examples on google
[14:19] <onryo> sure
[14:20] <onryo> ffmpeg -i $title -an -vcodec libx264 -pass 1 -preset veryslow -threads 0 -b $videoBR -x264opts frameref=15:fast_pskip=0 -f rawvideo -y /dev/null
[14:20] <onryo> then this ffmpeg -i $title -acodec libfaac -b:a $audioBR -ac 2 -vcodec libx264 -pass 2 -preset veryslow -threads 0 -b $videoBR -x264opts frameref=15:fast_pskip=0 ${title%.*}.$totSizeMB.mkv
[14:20] <burek> oh, you're missing 'k'
[14:20] <burek> or not :D
[14:21] <onryo> but the beginning of the program asks what size do you want the end product to be. Does all the "math" for you. with bit rates and junk
[14:21] <burek> ok ok, so what's the issue then
[14:21] <burek> cmd looks fine
[14:22] <onryo> well that is a good start =)
[14:22] <burek> (also check http://ffmpeg.org/trac/ffmpeg/wiki/x264EncodingGuide )
[14:22] <onryo> Well now I am going to just have it ask. What sizy is your screen. How many dpi do you want =)
[14:23] <zmbmartin> onryo: Any ideas how I can make this work > 'ffmpeg -loop 1 -i original.jpg -r 5 -i image_%d.png -filter_complex overlay -shortest testvid.mp4'
[14:23] <zmbmartin> onryo: that just loops until I quit and no audio file
[14:23] <pierre-olivierro> onryo: i got a lot of errors...
[14:23] <burek> zmbmartin where did you specify audio input?
[14:23] <onryo> burek when you said dpi it became a lot easer to foresee what the outcome would be =)
[14:24] <burek> :)
[14:24] <zmbmartin> burek: I didn't because all my attempts errored. It is a wav file.
[14:24] <burek> zmbmartin http://ffmpeg.org/trac/ffmpeg/wiki/Create%20a%20video%20slideshow%20from%20images
[14:24] <pierre-olivierro> you can see the result off ffprobe here : http://pastebin.com/8aWmFxgg
[14:25] <burek> ffmpeg -loop 1 -f image2 -r 5 -i image_%d.png -i audio.wav -c:v libx264 -c:a aac -strict experimental -shortest testvid.mp4
[14:26] <burek> pierre-olivierro those are warnings only
[14:26] <pierre-olivierro> onryo: do you know if there is a size limit or another thing to respect in extradata pascket stuffs ?
[14:26] <burek> because no key frame was detected and delta frames do not have a reference to calc their data on
[14:26] <burek> as soon as first key frame gets in, warning will stop
[14:28] <pierre-olivierro> Yes, but how can i manage this kind of error in my code ?
[14:28] <burek> it's not an error
[14:28] <burek> it's just warning
[14:28] <burek> ignore/drop those packets
[14:28] <burek> until you get a keyframe
[14:28] <zmbmartin> burek: Thanks but that gives me this error -> Error while opening encoder for output stream #0:1 - maybe incorrect parameters such as bit_rate, rate, width or height
[14:29] <burek> or use some "default" (custom-generated) frame (blank frame) as a key frame, to calc those delta frame
[14:29] <zmbmartin> burek: with this also Too many bits per frame requested
[14:29] <pierre-olivierro> I think the decoder is like lost because of extradata, and i notice thath when i use mod 16 packets error are very few...
[14:29] <burek> !pm zmbmartin
[14:29] <burek> :)
[14:30] <pierre-olivierro> I don't use ffmpeg command line, but directly the API
[14:31] <zmbmartin> burek: sorry -> http://pastie.org/5542715
[14:31] <burek> onryo, just calc the bitrate by b = size_of_the_output_file / duration_of_the_video_in_seconds
[14:31] <burek> of course take care of K M G
[14:32] <onryo> doing it as we speak =)
[14:33] <burek> for example, if you want the output file size of 1 MB and the video is 10 seconds long, then b = 1 MB / 10 = (approx) 0.1 * 8 Mbps = 800k
[14:33] <burek> byte = 8 bits :)
[14:34] <burek> although, to be perfectly correct, you should convert 1 MB to 1024*1024*1024 bits
[14:34] <burek> but, it's close enough :)
[14:34] <onryo> Yup dats wat I be doing =) Just dropped in a new menu what asks if you want to enter dpi or rez
[14:34] <onryo> burek so the way I calculated the size is not really right? I was not 100% sure about that.
[14:35] <burek> zmbmartin: [libx264 @ 0x7fe849818200] width not divisible by 2 (1131x707)
[14:35] <burek> crop your image first
[14:36] <burek> onryo, what is the duration of your video and what output file size do you want
[14:36] <onryo> The program only asks what size should the file be when its done. It goes in and figures out the duration for you =)
[14:37] <burek> onryo, so what help do you need then if program does it all? :)
[14:37] <onryo> I did it like this to get it in sec
[14:37] <onryo> movieTimeSec=$(awk '/Duration/ {gsub(/,/, "", $2); split($2, a, /:/); print 3600*a[1] + 60*a[2] + a[3]}' $ffmpegResize)
[14:37] <zmbmartin> burek: When I try with a proper sized image -> http://pastie.org/5542738
[14:38] <Macey> Hi All, I'm trying to transcode a http mp3 stream using  my own custom app. I can connect to the stream and get info on it.. How do i get the actual bytes out of the stream to pass into an encoder?
[14:38] <burek> zmbmartin: ffmpeg -loop 1 -f image2 -r 5 -i original.jpg -i sound.wav -c:v libx264 -c:a aac -ab 128k -ar 44100 -ac 1 -strict experimental -shortest testvid.mp4
[14:39] <burek> Macey, is that really an ffmpeg question?
[14:39] <Mavrik> Macey, what's your goal? transcode mp3 into something else?
[14:40] <burek> zmbmartin, btw your input audio is Mono, that's why there is -ac 1
[14:40] <burek> if you have a stereo audio, change it to -ac 2
[14:40] <Macey> i'm using the libaries so i believe so if i've asked in the wrong channel i apologise
[14:41] <burek> Macey, i mean, you said you are using your own custom app, so i figured you are doing everything yourself, hence my question
[14:41] <zmbmartin> burek: cool thanks, how do I control the loop to stop at say end of input images I give
[14:41] <Mavrik> Macey, the channel is right, your question just isn't clear enough.
[14:41] <zmbmartin> burek: also how do I control the FPS?
[14:41] <Mavrik> Macey, what's your input, do you need to do demuxing, what's "encoder" etc.
[14:41] <burek> zmbmartin, you already did that with -shortest
[14:42] <burek> ffmpeg -loop 1 -f image2 -r 5 -i original.jpg -i sound.wav -c:v libx264 -crf 23 -r 30 -c:a aac -ab 128k -ar 44100 -ac 1 -strict experimental -shortest testvid.mp4
[14:42] <burek> crf controls the output quality (the lower crf, the better quality)
[14:42] <burek> and -r controls output frame rate
[14:43] <zmbmartin> Well it should only be about 4 seconds but it is 16 seconds
[14:43] <burek> (first -r controls at what speed the input images will be read, the 2nd -r controls the output frame rate)
[14:43] <burek> pastebin?
[14:44] <Macey> Mavrik, ok. I am opening a http stream using avformat_open_input(&pFormatCtx, streamurl, NULL, NULL); I then go on to find a decoder from the decoder ( AVCodec ) how do i get bytes from the stream?
[14:44] <Mavrik> Macey, ok... you need to 1.) demux the stream first (to get audio from mp3 container) and then decode it (to get raw audio)
[14:45] <Mavrik> then you pass raw audio to encoder
[14:45] <onryo> burek the help I was asking for was that would be a sane rez to use on a tiny 4.8" screen. Guess 100 dpi is a good number. Also if that last 2 pass cmd in my program was optimal. No biggie really.
[14:45] <Mavrik> Macey, go check "doc/examples/demux.c" and "doc/examples/decode.c" I think
[14:45] <Mavrik> Macey, those will show you what you have to do
[14:46] <Mavrik> Macey, basically you call "av_read_frame" on input format to get packets
[14:46] <Mavrik> and pass them to avcodec_decode_audio3 to decode them
[14:46] <burek> onryo, since your phone uses 720p (@ 300 dpi) and if you want to use 100 dpi, then try with 427x240
[14:47] <onryo> I think that is the magic number for my phone.
[14:47] <burek> or 640x360
[14:47] <burek> (half the size)
[14:47] <burek> or quarter
[14:48] <burek> try and see what looks better :)
[14:48] <onryo> Should be OK with 100 if you think about it most pictures on the net are 74-76 ....
[14:48] <onryo> I would think
[14:49] <burek> ok, but find some standard size
[14:50] <burek> http://en.wikipedia.org/wiki/Display_resolution
[14:50] <Macey> thanks Mavrik, Great Help!
[14:50] <onryo> lol, I was reading that when you pasted it =P
[14:50] <Macey> sadly gotta do some proper work now though! :)
[14:51] <onryo> Oh man Sweden sucks nuts! 2:51pm and the sun is almost down ... seriously? Shit like that reminds me what I code all day =)
[14:51] <burek> it's because of the incoming apocalypse ^^
[14:51] <Macey> hehe
[14:52] <zmbmartin> burek: Here is what I am using for the command from what you gave me -> http://pastie.org/5542786
[14:52] <burek> anyway onryo, look at that image that shows many resolutions and follow the diagonal that says 16:9
[14:52] <burek> and find something around the hight of 240
[14:52] <zmbmartin> burek: Thanks by the way, but I have 119 images so at 24 FPS shouldn't that be just under 5 seconds?
[14:53] <zmbmartin> I it ending up just under 7 seconds
[14:53] <burek> zmbmartin i need a full output log
[14:53] <burek> also, you can't use 2 -i
[14:53] <burek> also, you can't use 2 -i options if you expect them to be joined somehow
[14:53] <burek> it won't work that way
[14:54] <burek> that's why there is -f image2 and % in the filenames, to specify multiple files
[14:54] <burek> it's not the same thing as specifying -i file1 -i file2
[14:54] <zmbmartin> burek: I want to set the original.jpg as the background for the video
[14:55] <burek> then take a look at filters
[14:55] <zmbmartin> burek: Here is the output -> http://pastie.org/5542794
[14:55] <burek> and try to figure out how to create a filter chain to do that
[14:55] <onryo> burek I put on suntan lotion on 12-12-12 and drank beer up on my roof with a friend. Was hoping a nuke or something just to kill up the boredom a bit. We got gypped! No epic apocalypse!
[14:56] <zmbmartin> burek: hmm& what I did is working it just seems a few seconds longer then I thought.
[14:56] <zmbmartin> burek: Thanks again for all your help
[14:56] <burek> zmbmartin how many images you have
[14:57] <zmbmartin> burek: 119
[14:57] <burek> that's weird
[14:58] <burek> 119/5 = 23.8
[14:58] <burek> are they numbered sequentially?
[14:58] <zmbmartin> yeah from image_0 to image_118
[14:58] <burek> onryo, do the same at 21-12-12 :)
[14:59] <onryo> sad thing is Ill prolly still be coding then too =P
[14:59] <burek> zmbmartin, it seems it only took 32/33 images
[14:59] <burek> can you do ls *.png
[14:59] <burek> and check there are no missing images
[15:01] <zmbmartin> burek: Yup ls *.png shows all 119
[15:02] <burek> that's strange.. is your output showing all 119 of them?
[15:02] <burek> can you do a playback?
[15:03] <zmbmartin> burek: Yeah output seems fine. All the images show in order as they should just after it is done there is a couple second more if silence.
[15:03] <zmbmartin> burek: If I take out the image sequence and add just one image it still comes in at 6:50
[15:03] <zmbmartin> Do you think the audio is causing that.
[15:04] <zmbmartin> the audio sound.wav shows as 4 seconds.
[15:04] <burek> can you do ffmpeg -i audio.wav -ar 22050 audio2.wav
[15:04] <burek> and see the duration
[15:05] <zmbmartin> burek: 4.12 -> http://pastie.org/5542834
[15:06] <burek> can you now use that audio2.wav as an audio input
[15:06] <burek> and see if it now works better
[15:07] <zmbmartin> burek: with all the same options?
[15:07] <burek> yes, just use the new audio as input now
[15:08] <zmbmartin> still outputs the same 6.50
[15:08] <burek> you could report it as a bug...
[15:09] <zmbmartin> burek: Alright thanks for all your help I appreciate it.
[15:09] <burek> :beer: :)
[15:19] <onryo> OK so I am guessing that PPI is just  Pythagorean theorem to get the diagonal and then dividing the diagonal but the screen size in inches.
[15:20] <onryo> seems right ... well we will see how this program turns out.
[15:23] <t4nk040> Would some be so kind as to explain why ffmpeg -y -f lavfi -i "aevalsrc=0::s=48000:d=3:n=(48000/25)" output.wav does not produce exactly 3 seconds of audio?
[15:27] <t4nk040> fflogger, sorry. Here is the pastie: http://pastie.org/5542894
[15:28] <onryo> So the ppi for my screen would be 305.959 ppi if its 1280x720 on a 4.8" screen?
[15:28] <onryo> Did it like this
[15:29] <onryo> ((1280²+720²))÷4.8
[15:29] <burek> t4nk040 http://ffmpeg.org/ffmpeg.html
[15:29] <burek> ‘duration, d’
[15:29] <burek> Set the minimum duration of the sourced audio. See the function av_parse_time() for the accepted format. Note that the resulting duration may be greater than the specified duration, as the generated audio is always cut at the end of a complete frame.
[15:30] <t4nk040> Yes, but 1920 divides into 48000 evenly. Shouldn't that be sufficient?
[15:30] <burek> you might want to use 't'
[15:30] <burek> onryo, why are you complicating your life :)
[15:30] <burek> use 640x360
[15:31] <burek> it's a standard half hd
[15:31] <t4nk040> I am using the silence as an input to a concat filter, so I _believe_ I need to limit the duration
[15:31] <burek> and your videos will be compatible with most devices
[15:31] <burek> t4nk040 did you try using 't' instead of 'd'
[15:32] <t4nk040> burek as an input option?
[15:32] <burek> no, wait
[15:32] <burek> Set the number of samples per channel per each output frame, default to 1024.
[15:32] <onryo> Anything I have to do more then 3 times I write a program to do. Just will ask what size is your screen or heck it will look in xorg if its there. Then ask how many dpi do you want and the rest is automagic
[15:32] <burek> that means
[15:32] <burek> your output will be a multiple of 1024 right?
[15:33] <t4nk040> burek I thought I could use n to create a frame size that matched, in this case 48000.
[15:33] <burek> t4nk040, can you type just ffmpeg -i output.wav
[15:33] <burek> to check the duration
[15:34] <burek> onryo, have fun with it then :)
[15:34] <t4nk040> burek 00:00:03.04
[15:35] <burek> t4nk040 and if you use aevalsrc=0::s=44100:d=3:n=1024
[15:35] <burek> does it produce 3 seconds precisely?
[15:36] <t4nk040> burek unfortunately, no: 00:00:03.02
[15:37] <burek> just a sec
[15:37] <t4nk040> sure (thank you)
[15:47] <burek> no developers online right now.. you'll have to wait a little bit i guess
[15:48] <t4nk040> burek I'll come back a little later and see if someone is available. Thank you.
[15:48] <burek> np :)
[16:33] <meekohi> I've installed ffmpeg on OSX using homebrew, but it doesn't seem to come with the VP8 encoder, which I need. Is there an easy way to add that encoder to ffmpeg at runtime, or does it need to be cooked in at compile time?
[16:38] <burek> meekohi, can you type: ffmpeg -version
[16:38] <burek> and copy it to a pastebin-like site (like www.pastebin.com)
[16:39] <meekohi> burek: http://pastebin.com/JexWdLaL
[16:39] <burek> that doesn't look like ffmpeg..
[16:39] <meekohi> How so?
[16:40] <burek> usually it looks like this
[16:40] <burek> http://pastebin.com/t8bhYGQn
[16:40] <meekohi> http://pastebin.com/VvX73FPq
[16:40] <meekohi> Sorry I thought the rest was excess
[16:42] <burek> you'll need to recompile your ffmpeg...
[16:42] <burek> it wasn't compiled with support for libvpx
[16:43] <burek> can you use libx264 instead
[16:44] <mateo`> meekohi: it might be a good idea to suggest libvpx support in ffmpeg formula (homebrew)
[16:45] <meekohi> burek, mateo`: In this case I need to use libvpx alas. There is an option to support libvpx in homebrew I just didn't realize it wasn't on by default.
[16:45] <meekohi> I'll try reinstalling it.
[16:46] <mateo`> meekohi: i just saw it :)
[16:46] <meekohi> I'm just always afraid to reinstall something when it's mostly working ;)
[16:52] <meekohi> Nice! That was surprisingly painless.
[18:17] <mang0> Hey guys, I'm trying to record my desktop with this: 'ffmpeg -f alsa -ac 2 -i hw:2,0 -f x11grab -r 20 -s 1920x1080 -i :0.0 -acodec pcm_s16le -vcodec libx264 -pix_fmt yuv420p -preset ultrafast -threads 0 output.mkv' but I'm getting trouble with the mic. It works fine to record audio, but mic sound isn't recorded. Any help appreciated. I'm using sound/mic off an external USB soundcard
[18:18] <mang0> Eh, disregard that I've discovered the problem/fix. Just gotta work out how to do it!
[18:30] <dericed> Can ffmpeg adjust DAR while using -codec copy? (going from quicktime to quicktime)
[18:46] <Mavrik> dericed, not really, SAR is flagged within the video stream which is unchanged when doing copy
[19:12] <Jack_D> Please help. I've been trying to convert a video from my S3 to put on a SD DVD. ffmpeg is cutting out the sound starting about 8 minutes in (4th scene/chapter), and cuts back in about 3 minutes later near the start of the next scene.
[19:15] <Jack_D> The only warning I'm getting is : Incompatible sample format 's16' for codec 'ac3', auto-selecting format 'flt'
[19:16] <Jack_D> This is my commandline :  ffmpeg -i 20121201_135852.mp4 -target ntsc-dvd -aspect 16:9 XMasRectl_1w.mpg
[19:19] <Jack_D> Well I also get this message in the stream: Seems stream 0 codec frame rate differs from container frame rate: 180000.00 (180000/1) -> 90000.00 (180000/2)
[19:21] <Jack_D> The original video plays fine in the computer with mplayer, so I don't understand why the problem. I've tried just copying the audio with -acodec copy , which gave me a garbage video out.
[19:22] <Jack_D> Anyone?
[19:23] <Jack_D> Suggestions on where I can get help, if no one is willing to help here?
[19:24] <Jack_D> Suggestion on how to convert using something else?
[19:25] <Jack_D> is there no user support on the user support channel?
[19:26] <Jack_D> Or no one who can help?
[19:26] <sacarasc> Or people are busy with other things.
[19:27] <Jack_D> thanks, For a minute I thought no one could see me.
[19:34] <Jack_D> What is !pb. I'm basically an irc newb.
[19:35] <Jack_D> Like this http://pastie.org/5543840
[19:42] <Jack_D> Did that paste work? Looks like I got knocked out and reconnected. Note the paste is from the currently running command so the final output is not there. I didn't have a complete listing from a previous run.
[19:46] <Mavrik> Jack_D, try specifying "-codec:a mp2"
[19:46] <Jack_D> A further note. I'm doing ffmpeg from the CLI because none of the GUI wrappers (dvdstyler, et al) are working either.
[19:46] <Jack_D> ok Thanks Mavrik!
[19:47] <Mavrik> maybe ac3 encoder's having problems, mp2 should be DVD compatible as well
[19:50] <Jack_D> It didn't like that. "Unrecognized option 'codec:a'", I tried using -acodec mp2 instead and it didn't like that either. "[mp2 @ 0xab0620] bitrate 448 is not allowed in mp2"
[19:55] <Jack_D> I added an -ab 192k and it running now.
[19:55] <Jack_D> Thanks Mavrik! Hope it works.
[19:56] <Mavrik> ah, your ffmpeg is probably old enough to not recognise new format
[19:56] <Mavrik> tell us if it works :)
[19:57] <Jack_D> Ah! Hadn't thought about that. If this fails, I'll download the source and build the latest ffmpeg.
[20:41] <Jack_D> Well that didn't fix the sound. Video and sound are still perfect from start to 6:02 and then a crackle and no sound until 8:45. One entire dance routine and the intro before and after. Then it's perfect again for the remaining performances. http://pastebin.com/2PGfbyj6
[20:41] <Jack_D> I'm going to download the latest stable ffmpeg and try again.
[20:59] <Jack_D> Wow a lot configuration options. Just going to try building with the defaults. Not sure that's a good thing. Lots of things look disabled by default. Anyone have a suggested config options list?
[21:01] <t4nk499> Would someone be so kind as to explain why this command does not result in exactly three seconds of silence, in spite of having an apparently appropriate block size? http://pastebin.com/9ertsjjG
[21:06] <beastd> Jack_D: There are some hints in the wiki at http://ffmpeg.org/trac/ffmpeg/wiki/CompilationGuide . When you are only decoding you mostly do not need any external deps. SDL would be needed for ffplay to work though.
[21:07] <Jack_D> Thanks beastd!
[21:16] <Jack_D> beastd, I'm really unschooled in the whole audio-video dept. I'm trying to convert from 1080p video from my Samsung S3 to a DVD, and will probably want to do Xvid or something else in HD that allows doing menus and such. That would mean both decoding and encoding, I presume.
[21:17] <beastd> yes
[21:18] <beastd> But usually you are limited regarding codecs when targeting DVD.
[21:19] <beastd> E.g. normal DVD standard uses MPEG-2 video (there is a bit more restrictions than just that)
[21:24] <beastd> If you have a DVD player that can play e.g. Xvid files you may also use that or maybe even just FFmpeg MPEG-4 encoder (in that case you may have to pretend that the video was encoded by Xvid). But generally the topic with hw playback and mpeg-4 codecs can be quite tricky because hw players can have all kind of limitations.
[21:25] <beastd> so prepare to do some tests before finding something that works for you.
[21:26] <beastd> Jack_D: also other people are really more knowledgable about this than me. i am sticking to software players since 2000... so i do not have much experience with encoding for hardware players.
[21:29] <Jack_D> Thanks! Yes, I've been testing and experimenting for over a week now. I'd laugh, but it's just frustrating. Here, I thought it would be easy to go from 1080p to 480p. What was I thinking?
[21:32] <beastd> Jack_D: It is not unusual for people to underestimate complications (be it accidental or essential) of multimedia (movie formats)
[21:33] <cbsrobot> Jack_D: what device ?
[21:35] <MIG-> Is my understanding correct with regards to ffmpeg command line flags?  ffmpeg helps to aggregate common settings between multiple codecs.
[21:35] <Jack_D> Going from my Samsung S3 to just DVD so I can share with other people. So many different DVD players most likely. But, I'll be picking up a new DVD player that handles xvid. Maybe a bluray. And there's a new thing  possibly a sandisk player soon.
[21:36] <MIG-> Thus, there are some codec settings/parameters that aren't accessible from ffmpeg (I am specifically interested in x264)
[21:36] <MIG-> for example.  How do I specify the keyint parameter via ffmpeg to x264 ?
[22:35] <Snaggle> bleh.  latest openjpeg got rid of the compatibility %p/include/openjpeg.h symlink
[22:36] <gp5st1> hello. is there any way to set the copyright and original bits differently in each frame of an MP3 and would that cause any problems in playback?
[22:37] <Snaggle> oops.  wrong channel. sorry
[22:51] <Jack_D>  warning: array subscript is below array bounds [-Warray-bounds] ! That's a just a warning? Seems it's aught to be an error.
[22:51] <klaxa> why? if it's below array bounds does it matter? all it does is allocate too much memory, nothing to worry about unless it's allocating more memory than ram
[22:52] <klaxa> even then you might have swapping :V
[22:57] <Jack_D> sounds like an explit waiting to happen to me.
[22:57] <Jack_D> exploit
[23:05] <LBo> I installed x264 & ffmpeg from source (git)
[23:05] <LBo> When using the presets with ffmpeg it work
[23:06] <LBo> But when I use "AVPresetVideo medium" with ffserver I get an error: "File for preset 'medium' not found"
[23:06] <LBo> Does ffserver handle the preset differently or something?
[23:48] <Jack_D> xD compiled and installed the latest stable ffmpeg. Still it's killing the audio from 6:02 to 8:49. Why!? There must be a way to fix this. Does it only work right in 32 bit OSes? http://pastebin.com/0jCdYFcF
[23:50] <urand0m> If I want to change the size of a video when doing 2 passes do I need to define the new size in both passes? ie  -s qvga ?
[00:00] --- Tue Dec 18 2012