[00:00] <deivor> I'm currently trying to record from a camera and microphone (linux) both devices have different "open" times, which means that there's a "buffer" of frames from the video at the beginning [00:00] <deivor> I've set the FPS to 15, but there's a "buffer" of images, which means that the output is at, say, 50 fps for 2 seconds, 40fps for 2 seconds ... and gets closer and closer to the 15fps over time [00:01] <deivor> I'm not really sure how to explain [00:02] <voip__> actualy am not tryng to record file i need stream to 2 IP's, i see strange bitrate=N/A in ststistics [00:02] <deivor> My output is something like this : http://pastebin.com/P1FFwi5U [00:03] <deivor> yes voip__ but the idea is that you rule out problems [00:03] <deivor> if you output to one file and it works, then your command is "OK" [00:03] <deivor> then you test with 2 files, to test if the output command is ok [00:03] <deivor> and if that works, the rtmp servers are wrong [00:03] <voip__> right [00:04] <deivor> also, (not necessary) compile ffmpeg with fdk_aac support (you get better sound with lower cpu usage) [00:09] <voip__> deivor, actualy i am using static compiled ffmpegs [00:11] <deivor> It's not necessary to do it, and most likely a pain to compile if you are on windows [00:11] <deivor> (I don't know if you are on windows or not) [00:11] <voip__> aa m on linux [00:11] <deivor> k [00:12] <voip__> i tryed to write to files instead of IP's. Recorded files are ok on vlc, but strange ststistics is same frame= 1648 fps= 25 q=-1.0 Lsize=N/A time=00:01:06.15 bitrate=N/A dup=11 drop=0 [00:12] <voip__> btrate n/a [00:13] <voip__> actualy i need stream to 2 wowza servers, audio on servers is good, but no video [00:16] <deivor> What happens if you stream to only 1 wowza server? [00:27] <voip__> deivor, with 1 server -f flv rtmp://ServerA or ServerB everything is OK. Problem comes when I trying to send 2 servers like: -f tee -map 0:v -map 0:a "[f=flv]rtmp://ServerA|[f=flv]rtmp://SeverB" [00:28] <deivor> what happens if you try with a piped process ? [00:29] <deivor> https://trac.ffmpeg.org/wiki/Creating%20multiple%20outputs [00:31] <voip__> is it will double cpu ussage ? [00:35] <deivor> nope [00:35] <deivor> it will add a little bit of overhead as you have two ffmpeg processes but it's almost nothing [00:38] <voip__> :) [00:39] <voip__> can you please help me covert piped command: ffmpeg/ffmpeg -i http://Source -preset fast -acodec aac -strict experimental -ar 44100 -ac 2 -b:a 96k -vcodec libx264 -b:v 900k -f tee -map 0:v -map 0:a "[f=flv]rtmp://ServerA|[f=flv]rtmp://ServerB" [00:40] <voip__> I just tryed but recieve eror -c copy is not recognized [00:46] <deivor> ffmpeg/ffmpeg -i http://Source -preset fast -acodec aac -strict experimental -ar 44100 -ac 2 -b:a 96k -vcodec libx264 -b:v 900k -f mpegts | ffmpeg/ffmpeg -f fmpegts -i - -c copy -f flv "rtmp://ServerA" -c copy -f flv "rtmp://ServerB" [00:47] <voip__> thank you ! [00:50] <deivor> does it work now? [00:52] <deivor> I'll be back in 45 minutes or so [00:52] <voip__> no :) At least one output file must be specified [00:52] <voip__> np [00:54] <deivor> oh [00:54] <deivor> ffmpeg/ffmpeg -i http://Source -preset fast -acodec aac -strict experimental -ar 44100 -ac 2 -b:a 96k -vcodec libx264 -b:v 900k -f mpegts - | ffmpeg/ffmpeg -f fmpegts -i - -c copy -f flv "rtmp://ServerA" -c copy -f flv "rtmp://ServerB" [00:54] <deivor> dash at the end of the first process [00:54] <deivor> now i'm gone [00:54] <voip__> i removed dash :) [00:54] <voip__> ok np we will talk tmororow :) [01:13] <voip__> guys, what mean : Using AVStream.codec.time_base as a timebase hint to the muxer is deprecated. Set AVStream.time_base instead. [01:13] <voip__> ? [01:32] <deivor> I'm back [01:33] <deivor> change your parameters as the error tells you to [01:52] <jellosea> hello, how do i get CBR for vp8? ive tried two pass but it doesnt give me CBR [01:57] <llogan> jellosea: why do you want CBR? [01:57] <llogan> or maybe i should ask why do you want vp8? [01:58] <deivor> if you can go vp9 / x264 [01:58] <llogan> vp9 is too slow to encode for sane people [02:01] <jellosea> no have to use vp8 [02:01] <jellosea> for my application [02:01] <jellosea> llogan: why do i need CBR? because i need it [02:01] <jellosea> for my application [02:01] <jellosea> i need to create a file with a specifc exact bitrate [02:05] <llogan> jellosea: then you'll have to read this http://www.webmproject.org/docs/encoder-parameters/ [02:06] <llogan> map the options to the libvpx avoptions by looking at 'ffmpeg -h encoder=libvpx' [02:06] <llogan> also https://trac.ffmpeg.org/wiki/Encode/VP8#ConstantBitrate [02:06] <jellosea> i did [02:06] <jellosea> the shit dont work [02:06] <jellosea> i read all dat shit [02:07] <deivor> Is there a specific signal to use for killing ffmpeg? So it closes the current stream correctly [02:07] <deivor> I'm running pkill atm [02:10] <llogan> jellosea: if you think it's a problem specific to libvpx then you'll have to ask in #vp8 or their mailing list webm-discuss [02:11] <jellosea> well theres no --end-usage in ffmpeg [02:11] <jellosea> but its in the webm encoder [02:11] <jellosea> and end-usage in webm says its for cbr [02:11] <jellosea> however ffmpegs instructions for cbr doesnt work [02:13] <jellosea> and im using 2 pass on top of this shit [02:13] <jellosea> how is it giving me really off bitrates with two pass [02:15] <llogan> the wiki article is from other users. it may not be correct. [02:15] <jellosea> whats -g for ffmpeg? [02:17] <llogan> group of pictures [02:17] <jellosea> a few other people seem to have this problem [02:17] <jellosea> but there seems to be no real answer [02:17] <llogan> does it work the way you want using vpxenc? [02:20] <jellosea> what is 3pass [02:20] <jellosea> i havent used vpxenc [02:20] <jellosea> will a 3rd pass help me [02:20] <llogan> 3 pass is an alias for "a waste of time" [02:20] <jellosea> ok [02:21] <jellosea> does the K have to be captialized [02:21] <llogan> if vpxenc doesn't create what you want then the problem is probably not with ffmpeg [02:21] <jellosea> -bitrate:v 250K [02:21] <llogan> no, don't capitalize [02:21] <jellosea> ok i didnt [02:21] <llogan> -b:v 250k [02:21] <jellosea> yea [02:22] <jellosea> well with -loglevel debug it seems that [02:23] <jellosea> libpvx seems to be receiveng cbr mode [02:23] <jellosea> but this bitch wont output cbr [02:25] <deivor> llogan, anything on 'correctly' killing ffmpeg? (for example, ffmpeg will produce a 'corrupt' file if killed when encoding mp4 ) [02:28] <llogan> deivor: how are you currently doing it that makes the corrupt output? [02:29] <deivor> It doesn't, yet (I'm making a web-remote for live-streaming) [02:29] <deivor> I'm just running pkill, but I don't know if it'll mess up the RTMP stream [02:31] <llogan> i can't duplicate the issue using pkill if outputting mp4 [02:35] <relaxed> deivor: try, killall -INT ffmpeg [02:38] <deivor> llogan: If you start transcoding certain codecs and ctrl-c, you'll get a 'broken video' (easy to fix), whereas if you hit 'q' the video stops at the same place but works [02:38] <deivor> that's what I mean [02:38] <deivor> relaxed: thanks I'll try it [02:39] <llogan> unable to duplicate [02:41] <relaxed> llogan: I think it happens when the output container is mp4 [02:42] <llogan> that's what i tried [02:44] <relaxed> yeah, pkill works here too [03:24] <nezZario> Hi folks [03:29] <deivor> I'm trying to get www-data to execute ffmpeg as another user but won't give me access to alsa.. is this ffmpeg's fault? [03:30] <deivor> running as user "curso" I get my script running wonderfully [03:30] <deivor> running as www-data with correct sudo permissions: sudo -u curso script [03:30] <deivor> ALSA lib audio/pcm_bluetooth.c:1614:(audioservice_expect) BT_OPEN failed : Invalid argument(22) [03:30] <deivor> [alsa @ 0x1669510] cannot set parameters (Invalid argument) [03:30] <deivor> btheadset: Input/output error [03:30] <deivor> Should these permissions be handled by ffmpeg or not? (If not, I'll have to check alsa/bluetoothd) [03:35] <deivor> nevermind, I had a stray screen blocking access [04:24] <nezZario> Guys [04:24] <nezZario> How can I convert .mod to .flv or .mp4? [04:32] <llogan> nezZario: ffmpeg -i input.mod output.mp4 [05:09] <jellosea> does the order of arguments matter in ffmpeg [05:19] <llogan> jellosea: yes http://ffmpeg.org/ffmpeg.html#Synopsis [05:22] <nezZario> hmm [05:22] <jellosea> are things like sn and an input file options [05:22] <nezZario> Well let me ask this, .. what would be a good 'something' (heck, I'd go for a book if it's <$20) to teach me about codecs/containers/video/things I'll encounter in ffmpeg? [05:23] <jellosea> lol [05:23] <nezZario> I'm pretty lost.. I did some converting from mod to flv (no idea what codecs?) .. when I re-play it in flv, the quality is pretty decent in the flv but the video is literally a little smaller even though VLC player says they're both 720x480 [05:23] <jellosea> mod? do you mean mov [05:24] <nezZario> no, the file is ".mod" .. "JVC .mod format" or something [05:24] <jellosea> llogan: is -passlogfile video_name.log file? [05:24] <jellosea> llogan: fine rather [05:24] <jellosea> is that how i use a unique two pass log file name [05:27] <llogan> nezZario: https://trac.ffmpeg.org/wiki/Encode/H.264 [05:28] <llogan> jellosea: -sn and -an are output options. see the documentation. it will tell you if the option is global, input and/or output. [05:29] <llogan> http://ffmpeg.org/ffmpeg.html [05:29] <jellosea> oh ok [05:29] <nezZario> llogan: any idea how I check if my ffmpeg is compiled with h264 support? ( I don't think it is, everytime I tried it said invalid encoder: ... or something along those lines) [05:30] <jellosea> so what are generic options [05:30] <jellosea> just type ffmpeg [05:30] <llogan> nezZario: run "ffmpeg" the console output. look for "--enable-libx264" [05:30] <jellosea> llogan: is generic global [05:30] <llogan> i don't know of an option named generic [05:31] <jellosea> no i mean in the documentation [05:31] <jellosea> for example -loglevel [05:31] <jellosea> it doesnt say if its global input or output [05:33] <llogan> i guess they don't all say, but i believe it is considered a global option [05:36] <jellosea> ok [05:44] <jellosea> llogan: what should i set -bufsize if i want strict CBR [05:55] <relaxed> why do you need/want cbr? [05:57] <jellosea> not relevant [06:00] <relaxed> jellosea: https://trac.ffmpeg.org/wiki/Encode/H.264 [06:00] <jellosea> im not even using h264 [06:00] <jellosea> also that shit dont work [06:03] <jellosea> https://gist.github.com/anonymous/a6378ef3d7e4f8baf552 [06:05] <relaxed> please reread fflogger's message [06:11] <jellosea> https://gist.github.com/anonymous/38596fe82b0518589471 [06:13] <relaxed> and you're issue is you think it's not CBR? [06:25] <deivor> Hi, anyone up? [06:29] <jellosea> relaxed yes it is not CBR [06:29] <jellosea> i ask for the bitrate to be 250k [06:30] <jellosea> and the output is like 500k [06:38] <deivor> I have a camera that provides constant FPS, but ffmpeg is giving me 1 fps extra, which makes the video drift from the audio over time [06:38] <deivor> I have ffmpeg -framerate 20 -r 20 -i - [...] -vsync drop -r 20 -async 1 OUTPUT [06:39] <deivor> that's the relevant part, at least [06:40] <jellosea> so relaxed you have anythign to say [06:56] <relaxed> jellosea: how did you verify it's not CBR? [06:59] <jellosea> relaxed: its not hitting my target bitrate [06:59] <jellosea> how do i get it to hit my target bitrate [07:00] <jellosea> Bit rate : 435 Kbps [07:00] <jellosea> in mediainfo [07:00] <relaxed> did you calculate this from the ouput's size or are you just guessing? [07:00] <jellosea> for video [07:00] <jellosea> mediainfo [07:00] <deivor> ffprobe ? [07:00] <relaxed> mediainfo is wrong quite a bit [07:02] <jellosea> its not wrong [07:02] <jellosea> based on the filesize [07:02] <jellosea> and based on ffmpegs output [07:03] <jellosea> so you have no info [07:08] <deivor> Does your output have audio? Are you sure you are comparing audio+video bitrates (and you have set them correctly on the command) ? [07:12] <jellosea> yes [07:12] <jellosea> yes i am [07:12] <jellosea> jesus [07:12] <jellosea> no one is giving me any help [07:12] <jellosea> fucking stupid ass advice [07:13] <relaxed> what's the output's size in MB? [07:13] <jellosea> 14.6 MB [07:13] <jellosea> this is retarded [07:13] <jellosea> all you are gonna ask me is if ithe bitrate is right [07:13] <jellosea> instead of finding out why the bitrate is off [07:14] <jellosea> if you dont know why the bitrate will be off then just say so [07:14] <jellosea> instead of trying to find out that the bitrate is actually right [07:18] <relaxed> we're not paid to sit here and provide you with real-time support, so why don't you develop some patience [07:29] <jellosea> why are you wasting my time if you dont know what the problem is [07:30] <jellosea> just say you dont know instead of [07:30] <jellosea> paste the log [07:30] <jellosea> paste this [07:30] <jellosea> why do you want cbr [07:30] <jellosea> why do you want to encode video [07:30] <jellosea> maybe you should read a book [07:30] <jellosea> like wtf [07:31] <deivor> relaxed: do you know if I can override the image PTS with constant PTS? I'm getting just a little bit over the fps I should from the camera, and over time it goes out of sync [07:34] <relaxed> deivor: pastebin the command and output so i can see what's going on. [07:37] <deivor> http://pastebin.com/mMXy30dr [07:38] <deivor> the input is supposed to be 20fps, it hovers at about 20.3 (at the moment I've the command with -r 20.3) but I'm going to use in sessions about 2 hours long, so a small drift is huge [07:40] <deivor> So I'm thinking dropping the PTS and forcing it to spit 20 frames per second with correct PTS should do it (if fps > 20, drop, if fps < 20 duplicate) [07:43] <relaxed> instead of -r 20 use -framerate 25 [07:44] <relaxed> remove the follow -r -vsync and -async [07:44] <relaxed> following* [07:44] <deivor> framerate 25 for output or input? [07:45] <relaxed> -framerate applies to the input [07:45] <deivor> my input is fixed at "20"fps, why should I use -framerate 25? (isn't framerate used to 'ask' for a certain framerate?) [07:46] <relaxed> I'm going by "Stream #0:0: Video: h264 (High), yuv420p, 1280x720, 25 tbr, 1200k tbn, 50 tbc" [07:47] <relaxed> but first off, start with a slimmed down command and it encode it to a local file [07:48] <deivor> The drift problem is from the camera [07:48] <relaxed> ok [07:48] <deivor> the 25 tbr is because I didn't let ffmpeg analyze the stream (analyzduration and probesize too small) to reduce latency [07:49] <relaxed> I see [07:52] <relaxed> then "ffmpeg -framerate 20 -i input -r 20 ..." should drop or dup frames to maintain sync. [07:54] <relaxed> play with -framerate's value to see if it helps [08:23] <jellosea> how can i add two vf filters [08:24] <relaxed> https://www.ffmpeg.org/ffmpeg-filters.html#Description [08:25] <jellosea> thanks [08:39] <jellosea> if i do -vf scale=-1:480; ass=sub.ass it doesnt work [08:39] <jellosea> something with the  i think [08:40] <jellosea> if i do -vf ass=sub.ass it works, or -vf scale=-1:480 it works too [08:40] <jellosea> combining it with quotes and semicolon doesnt seem to work [08:41] <relaxed> -vf scale=-1:480,ass=sub.ass [08:43] <jellosea> works thanks [08:45] <relaxed> if the subs look odd you might have to reverse the filters [09:06] <svvitch> what format to use for png image, as -f param ? [09:08] <svvitch> when adding overlay [09:13] <relaxed> svvitch: you shouldn't need to use -f unless you're piping it in. [09:14] <relaxed> ffmpeg -i video -i image.png -filter_complex overlay output [09:29] <svvitch> relaxed: the problem is that png file is without extension and ffmpeg does not understand what kind of file is [09:36] <relaxed> try, ffmpeg -i video -f image2 image ... [09:36] <relaxed> er, ffmpeg -i video -f image2 -i image [10:42] <alexvf> hi, i need help with custom I/O to read a network stream (2.2.2) [10:42] <alexvf> i had a blocking implementation of the read callback that worked right, but i had to changed to non-blocking because that thread needs to attend some user input [10:43] <alexvf> now, i'm returning AVERROR(EAGAIN) when i can't read from network [10:43] <alexvf> and i'm seeing noise in the decoded images [10:44] <alexvf> is even possible in the current ffmpeg state to have a non blocking custom I/O for network streams? [11:17] <ahessler> how much ram is needed for 2350x1080 ? [11:17] <ahessler> 1GB is not enough it seems [11:22] <ahessler> oh, 2G was enough [11:26] <alexvf> please, someone with some knowledge on AVIOContext can tell me if custom I/O supports AVERROR(EAGAIN)? It seems that av_read_frame retursn incomplete frame when the read callback returns that value [11:40] <akusingla> hello!! I am new to ffmpeg , I wanted to know if this api could be used for clipping videos on server side? [11:50] <LucasG> Hi guys, anyone can share ffserver for a while? I have Connection reset by peer when i`m trying sendstream via http from ffmpeg to my ffserver. Complete log: http://pastebin.com/D9LhWrUc [12:41] <line0> is there any way to tell ffmpeg to not do any character encoding conversions with -sub_charenc? [12:42] <line0> (when extracting subtitles) [12:43] <line0> like, just extract bytes and let something else take care of the rest? [14:00] <relaxed> line0: did you see `ffmpeg -h full|less +/sub_charenc_mode` [14:03] <line0> ah [14:03] <line0> thanks [16:33] <djook> hye huys [16:33] <djook> I have a little question, not really code related [16:34] <djook> I'd like to know if there is a specific country of origin fro the project [16:34] <djook> as vlc is French, mplayer Hungarian and cyrus British.... [16:36] <djook> I found the dns to be Name Server:NS.ARTHIST.MTA.HU [16:36] <djook> Name Server:THOT.BANKI.HU [16:36] <djook> Name Server:NS3.MPLAYERHQ.EU [16:36] <djook> so Hungarian as well ? [16:40] <klaxa|work> afaik the original author of ffmpeg was fabrice bellard [16:40] <iive> ffmpeg is hosted on same server as mplayer, that's why it is hosted in hungary [16:42] <djook> iive, thanks to let me know ... [16:42] <djook> I'll have a closer look at Bellard "origins" [16:42] <iive> it started on sourceforge, then moved to arpi's mphq in hungary [16:43] <iive> then a new machine that KotH set up in Switzerland. [16:44] <iive> after the take over, arpi gave a new hosting machine. [16:44] <djook> In the french wikipedia article about Bellard, He is said to be the first ffmpeg coder/dev [16:45] <djook> iive, I remember that some servers were compromised circa 2010/11 [16:45] <djook> ffmpeg compilation is one of my favorite computer game [16:46] <djook> I got myself an xserver ppc64 just for ffmpeg tests [16:46] <djook> I played a bit with a ps3 in 2011 but du to the 256M of ram [16:46] <djook> the OOM occured very fast [16:46] <iive> hehe, that's great, do you run fate on it? [16:47] <djook> fate ? [16:47] <djook> is it an OS ? [16:47] <iive> it is automated test environment. [16:47] <iive> http://fate.ffmpeg.org/ [16:47] <djook> ok [16:48] <djook> I ll have a look at that [16:49] <djook> For the moment I only encode libx264 \/libfaac [16:50] <iive> these automated tests allow finding out quickly if some change breaks some specific architectures. [16:50] <iive> e.g. when some change is not endian safe. [18:21] <astr> how to sync audio with video? [18:26] <voip_> hi guys [18:26] <voip_> pls help me with 1 stream processing and multiple IP outputs. In ffmpeg status i have someting like frame= 412 fps= 26 q=-1.0 Lsize=N/A time=00:00:16.74 bitrate=N/A dup=19 drop=0. My output hase audio, but no video. [18:26] <voip_> <voip__> http://pastebin.com/wByE8CB8 [18:39] <khaosmarmoset> hey. [18:39] <khaosmarmoset> so regarding my webcam recording project. [18:39] <khaosmarmoset> I can now set the bit rate and the i-frame period to suit youtube and everything seems great [18:40] <khaosmarmoset> the only weird thing is that I can't seem to affect the fps of the webcam myself very much. [18:40] <khaosmarmoset> or at all, really. [18:40] <khaosmarmoset> sometimes when I start the process I get (after stabilisation) 25fps. [18:40] <khaosmarmoset> usually it stabilises to about 15 fps. [18:41] <c_14> Can you/are you using -r as an input option? [18:41] <khaosmarmoset> the bitrate doesn't seem to have very much to do with it. I can set it higher or lower, but usually I drift back into 15 fps. [18:42] <khaosmarmoset> I am using -r, but again it doesn't seem to be doing anything. [18:42] <khaosmarmoset> v4l2-ctl's --set-parm also doesn't seem to be doing very much. [18:42] <khaosmarmoset> ditto with my uvch264ctl's frame interval setting. [18:44] <khaosmarmoset> so basically I have three knobs to twiddle, but none of them work. [18:45] <khaosmarmoset> I don't really mind all that much, because 15fps is enough for me. 24 would be better. 30 would be great [18:45] <c_14> You can get ffmpeg to drop/duplicate frames, but if your webcam isn't being reasonable there isn't much you can do on that end. [18:46] <khaosmarmoset> when using -acodec copy -vcodec copy, then is the fps I see when recording the actual measured fps from the webcam? [18:47] <c_14> The fps= you see when recording is the recording fps, if you can record in realtime, then yes it is. (assuming real-time source) [18:47] <khaosmarmoset> right, ok [18:47] <khaosmarmoset> damn, gotta run out for a bit before the shops close. brb. [18:54] <astr> khaosmarmoset, how d you keep audio and video in syic? [18:54] <astr> sync [19:10] <Lokie> hello, I am trying to change a mkv to a mp4. -codec copy works fine for the video & audio streams but the subtitles (.ass) don't get copied. Shouldn't -codec copy copy all the streams? [19:11] <rcombs> ASS subtitles aren't supported in MP4 [19:11] <Lokie> oh thought it was srt [19:11] <rcombs> neither are those [19:12] <Lokie> oh [19:13] <Lokie> mp4 supports only hardsubs? [19:13] <khaosmarmoset> astr: erh. I'm not sure. [19:13] <rcombs> I think you could convert to mov_text [19:13] <astr> khaosmarmoset, ok [19:14] <khaosmarmoset> astr: which is to say, I just run ffmpeg -f alsa -i hw:1 -f v4l2 -r 24 -vcodec h264 -i /dev/video0 -vcodec copy -acodec aac -strict -2 -f flv [19:14] <khaosmarmoset> astr: if you have any pointers, please let me know! [19:14] <rcombs> khaosmarmoset: 0x3A28213A [19:14] <astr> when 2 "-vcodec" ? [19:14] <astr> when=why [19:15] Action: khaosmarmoset faults rcombs [19:16] Action: rcombs SIGSEGVs [19:16] <khaosmarmoset> astr: the first one chooses the codec for the v4l2 input. the second one says to just move it along [19:16] <khaosmarmoset> (I think. I'm not 100% sure) [19:18] <khaosmarmoset> astr: yeagh, without the first -vcodec h264 ffmpeg resets it to raw YUYV and then the frame rate drops to 5 fps [19:19] <voip_> trate doesn't seem to have very much to do with it. I can set it higher or lower, but usually I drift [19:19] <astr> what about -input_format see ffmpeg -h full |less press "/" and paste "V4L2 indev AVOptions"and enter to search [19:19] <voip_> http://pastebin.com/jePTMDbc [19:19] <astr> don't know only just leaned of it [19:19] <astr> "first -vcodec h264 ffmpeg resets it to raw YUYV" doesn't make sense to me [19:20] <khaosmarmoset> I mean if I run it without the -vcodec h264, then I see that ffmpeg has chosen YUYV as the pixel format according to v4l2-ctl --all [19:21] <khaosmarmoset> this is basically raw uncompressed frames afaict, and the hw only supports 5 fps for that case [19:21] <astr> ok [19:22] <kepstin-laptop> oh, that's an input option on the v4l device, right. [19:23] <Lokie> rcombs I get Stream #0:2 -> #0:2 (ssa -> mov_text) [19:23] <Lokie> and subtitle:10kB but I don't see any subs [19:23] <Lokie> used: -i source.mkv -i source/para04.ass -c copy -c:s mov_text target [19:26] <rcombs> your player might not support it? [19:26] <rcombs> mov_text is rather obscure [19:26] <Lokie> using ff to play it so maybe [19:26] <Lokie> FFmpeg does not (yet) support [19:26] <Lokie> setting all properties for mov_text correctly. mar 2013 [19:28] <voip_> please help http://pastebin.com/jePTMDbc [19:30] <llogan> voip_: loglevel debug is often more annoying that helpful [19:30] <llogan> why are you piping from ffmpeg to ffmpeg? [19:33] <khaosmarmoset> how can I have the output go nowhere? [19:33] <llogan> -f null - [19:33] <llogan> http://ffmpeg.org/ffmpeg-formats.html#null [19:34] <khaosmarmoset> thanks [19:34] <khaosmarmoset> so that has nothing to do with it. [19:35] <khaosmarmoset> I wanted to eliminate the -f flc as something that might be contributing to my problem [19:35] <khaosmarmoset> -f flv even [19:37] <voip_> llogan, i neeed [19:37] <voip_> multiple outputs [19:37] <voip_> https://trac.ffmpeg.org/wiki/Creating%20multiple%20outputs#no1, adm my command is ffmpeg -i input1 -i input2 -acodec & -vcodec & -f mpegts - | \ [19:37] <voip_> ffmpeg -f mpegts -i - \ [19:37] <voip_> -c copy output1 \ [19:37] <voip_> -c copy output2 \ [19:37] <voip_> -c copy output3 \ [19:38] <voip_> i dont know why in debug null, but in command i typed right output adress [19:39] <llogan> why not use tee muxer? [19:40] <voip_> i tryed with muxer, dosent work for me [19:41] <voip_> ffmpeg/ffmpeg -i http://Source -preset fast -acodec aac -strict experimental -ar 44100 -ac 2 -b:a 96k -vcodec libx264 -b:v 900k -f tee -map 0:v -map 0:a "[f=flv]rtmp://ServerA|[f=flv]rtmp://ServerB" [19:42] <llogan> did you try the -flags +global_header suggestion? [19:42] <voip_> no, can you help me, write right command ? [19:43] <llogan> i mentioned it yesterday, but you didn't respond [19:43] <voip_> i dont know how use -flags +global_header [19:43] <llogan> just add it as an ouptut option [19:44] <voip_> like: ffmpeg/ffmpeg -i http://Source -preset fast -acodec aac -strict experimental -ar 44100 -ac 2 -b:a 96k -vcodec libx264 -b:v 900k -f tee -map 0:v -map 0:a -flags +global_header "[f=flv]rtmp://ServerA|[f=flv]rtmp://ServerB" ? [19:45] <llogan> sure, try that [19:45] <voip_> lets me chek [19:46] <Lokie> hmm if I were to burn the subs on the mp4 is there a flag that will used the same code / quality from the source? [19:46] <Lokie> -d [19:46] <Lokie> codec* [19:46] <llogan> not with a lossy encoder [19:46] <Lokie> source is a h264 mkv [19:47] <Lokie> not sure that I understood your comment though [19:47] <Lokie> ffprobe gives: Stream #0:0: Video: h264 (High), yuv420p, 1280x720 [SAR 1:1 DAR 16:9], 23.81 fps, 23.81 tbr, 1k tbn, 47.95 tbc [19:47] <llogan> you can't get "same quality" if you use an encoder that makes a lossy output [19:48] <llogan> you can get something the looks lossless or nearly so, but technically it is not [19:51] <voip_> llogan, brilliant, it works! but durig encode i see bitrate=N/A why ? [19:52] <llogan> i don't know [19:53] <voip_> thank you ! [19:56] <voip_> guys, during encoding frame= 77 fps= 33 q=24.0 size= , what is meaning variable q=24.0 ? [20:02] <khaosmarmoset> oh [20:02] <khaosmarmoset> wow. [20:03] <khaosmarmoset> when I turn the lights on I get 30 fps just like momma promised [20:03] <khaosmarmoset> woot! [20:03] <kepstin-laptop> yeah, most webcams use slower shutter speeds in the dark, limiting fps. [20:04] <khaosmarmoset> right. turns out I didn't know that! [20:04] <khaosmarmoset> thanks webcam. you were saving me from meself. [20:04] <khaosmarmoset> :-D [20:10] <voip_> guys, during encoding i see ststus: frame= 77 fps= 33 q=24.0 size=.., what is meaning variable q=24.0 ? [20:11] <JEEB> the value is rather useless in modern formats, but it is supposed to note some average quantizer or so [20:11] <JEEB> it really is useless and in general tells you absolutely nothing [20:12] <voip_> oh, thanks [20:15] <voip_> one more question, can you sgggset any special parameters, for taking streams from far away servers (too many hops to server, pinging time too high) [21:46] <voip_> [aac @ 0x317c7e0] Input buffer exhausted before END element found [21:46] <voip_> Error while decoding stream #0:0: Invalid data found when processing input [21:46] <voip_> [h264 @ 0x38b40e0] negative number of zero coeffs at 29 18 [21:46] <voip_> [h264 @ 0x38b40e0] error while decoding MB 29 18 [22:04] <voip_> llogan, do you know how to fix problem with : [22:04] <voip_> [h264 @ 0x38b40e0] negative number of zero coeffs at 29 18 [22:04] <voip_> [h264 @ 0x38b40e0] error while decoding MB 29 18 [22:04] <voip_> [h264 @ 0x38b40e0] concealing 812 DC, 812 AC, 812 MV errors in I frame [22:05] <llogan> voip_: no. and when you have a problem you should always include your command and the console output (using a pastebin service) [22:07] <voip_> llogan, sorry http://pastebin.com/PdmRAcBn [22:44] <voip_> please help http://pastebin.com/PdmRAcBn [23:08] <henriqueleng> I'm recording my desktop but it is using a lot of my CPU, can I speed up this process, maybe with post video processing? [23:11] <henriqueleng> pastie.org/9684356 [23:11] <henriqueleng> [23:12] <llogan> henriqueleng: where's the rest? [23:12] <henriqueleng> llogan: rest of what? [23:12] <llogan> show your EXACT ffmpeg command and the COMPLETE console output [23:13] <llogan> that seems to get ignored more than it doesn't [23:13] <llogan> why? [23:16] <henriqueleng> llogan: http://pastebin.com/ybk4ALuC [23:22] <llogan> henriqueleng: why use the 1.2 release branch instead of git master or most recent release branch? use -framerate instead of -r for x11grab. changing "-b:v 10k" to "-qp 0" may be slightly faster (and it seems to be ignoring your -b:v anyway) [23:24] <llogan> ...also it was set to 10k. no wonder it ignored. [23:26] <llogan> oh, it ignored because you tried to set -b:v and -preset as input options [23:26] <llogan> move preset after "-i :0.0+0.0" [23:33] <henriqueleng> llogan: Qhat is -qp 0 for? [23:34] <llogan> lossless output. it may encode "faster". then you can re-encode it at your leisure. [23:35] <henriqueleng> llogan: And should I maintain the -preset ultrafast? [23:35] <llogan> yes [23:35] <henriqueleng> llogan: Right, thanks for helping [00:00] --- Thu Oct 30 2014