[Ffmpeg-devel-irc] ffmpeg.log.20150925

[burek]

[00:10:43 CEST] <C0nundrum> is there a windos compiled version of ffmpeg that supports screen capture on nvidia cards ?
[00:57:36 CEST] <C0nundrum> anyone here
[00:57:52 CEST] <C0nundrum> How come this room is always so dead
[00:58:31 CEST] <c_14> Because nobody understood your question?
[00:58:48 CEST] <c_14> Is "windos" supposed to mean "Windows" or is it an evil amalgamation of Windows and DOS?
[00:58:57 CEST] <c_14> Also, why wouldn't gdigrapb work on nvidia cards?
[00:58:59 CEST] <c_14> *gdigrab
[01:00:04 CEST] <C0nundrum> what i'm trying to do is
[01:00:14 CEST] <C0nundrum> ffmpeg -f dshow -i video="screen-capture-recorder" -ac 6 -strict -2 -c:a aac -c:v libx264 -preset ultrafast "rtmp://192.168.1.145/live1/test"
[01:00:51 CEST] <C0nundrum> But i get  Unable to find a suitable output format for 'rtmp://192.168.1.145/live1/test'
[01:00:51 CEST] <C0nundrum> rtmp://192.168.1.145/live1/test: Invalid argument
[01:01:05 CEST] <c_14> -f flv rtmp://
[01:01:23 CEST] <c_14> Why did you assume that had something to do with either the graphics card or screen capturing?
[01:01:54 CEST] <C0nundrum> and if i specify that way i get
[01:02:32 CEST] <C0nundrum> http://i.imgur.com/PKyIAHu.png
[01:02:56 CEST] <C0nundrum> well c_14 i know new vidia cards have a api for low latency capture
[01:03:32 CEST] <C0nundrum> but anyway desided to got with screeen capture recorder and i get the above error
[01:03:38 CEST] <C0nundrum> go with*
[01:04:01 CEST] <c_14> You're not encoding fast enough. (probably)
[01:04:04 CEST] <c_14> So the input buffer is too full.
[01:06:35 CEST] <C0nundrum> What could do to prevent that ?
[01:06:42 CEST] <C0nundrum> could i*
[01:07:47 CEST] <C0nundrum> it happens even if i set framerate to 1
[01:08:12 CEST] <c_14> How's the cpu usage?
[01:09:51 CEST] <C0nundrum> while incrementing last messeage repeated the process uses 4-7%
[01:09:56 CEST] <C0nundrum> using the following command
[01:10:04 CEST] <C0nundrum> ffmpeg -f dshow -i video="screen-capture-recorder" -framerate 1 -ac 6 -strict -2 -c:a aac -c:v libx264 -preset ultrafast -f flv "rtmp://192.168.1.145/live1/test"
[01:10:33 CEST] <c_14> -framerate has to go before -i
[01:12:08 CEST] <C0nundrum> moved it an ffmpeg ran for prob 3 secodns and gave this error http://i.imgur.com/URlUGdV.png
[01:14:10 CEST] <C0nundrum> if i do a fileoutput instead however it works
[01:14:23 CEST] <c_14> Oh, then your network is just too slow.
[01:14:47 CEST] <C0nundrum> the rtmp server is on the same machine
[01:14:57 CEST] <C0nundrum> how can it be to low at 1fps...
[01:16:27 CEST] <C0nundrum> could it be that for some reason it isn't sending to the rtmp server ?
[01:17:09 CEST] <C0nundrum> ah fixed it
[01:19:32 CEST] <C0nundrum> well kinda it streams now but there is a large delay
[01:19:54 CEST] <c_14> How large is large?
[01:22:42 CEST] <C0nundrum> ffmpeg cpu usage is around 14% Stream has a delay of 12 seconds and goes grey sometimes. is there anway to improve performance, my pc should be able to do better than that
[01:23:21 CEST] <C0nundrum> the the logs are like this http://i.imgur.com/TQP2mvL.png
[01:23:40 CEST] <C0nundrum> with frametime set tp 20 fps i would like to have 30fps or evenr 60fps with less than a second delay
[01:24:25 CEST] <c_14> Have you tried using gdigrab instead of screen-capture-recorder?
[01:25:08 CEST] <C0nundrum> well i tried ffplay with the same input and there is almost no delay so i would asume it is in the encoding process but i will try streaming gid now
[01:31:34 CEST] <C0nundrum> gdigrab seems to stream better but it still has a playback delay of 10 seconds. Could this be because of ffmpeg buffer ?
[01:31:50 CEST] <c_14> ffmpeg + x264 + network probably
[01:53:23 CEST] <C0nundrum> alright i managed to reduce the delay to 1 second
[01:53:34 CEST] <C0nundrum> ffmpeg -f gdigrab -i desktop -f flv rtmp://192.168.1.145/live1/test2"
[01:53:42 CEST] <C0nundrum> is there anwya to get it down durther ?
[01:58:19 CEST] <c_14> Under a second? Have you tried -tune zerolatency?
[01:59:07 CEST] <DHE> that's x264 only which he doens't specify a codec
[01:59:27 CEST] <c_14> Doesn't flv default to x264?
[01:59:39 CEST] <c_14> Ah, no it does not.
[01:59:46 CEST] <C0nundrum> http://i.imgur.com/fsACD86.png
[02:00:37 CEST] <c_14> -fflags +nobuffer maybe?
[02:01:03 CEST] <c_14> Wait, no that's an avioflag
[02:01:11 CEST] <c_14> No, it is an fflag
[02:01:23 CEST] <c_14> This section needs some decent indentation or something.
[02:06:12 CEST] <C0nundrum> also i'm trying to set framerate for gdidraw but it won't let me do anything over 14 fps ?
[02:06:53 CEST] <C0nundrum> even if i do ffmpeg -f gdigrab -framerate 60 -i desktop -f flv rtmp://192.168.1.145/live1/test2"
[02:06:58 CEST] <C0nundrum> it says it recording at 13 fps
[02:07:35 CEST] <c_14> Where's it say that?
[02:08:09 CEST] <C0nundrum> http://i.imgur.com/G9Amrzi.png
[02:08:57 CEST] <c_14> That just shows you how fast it's encoding.
[02:09:11 CEST] <c_14> Your system can't encode using that encoder at more than 13ish fps
[02:09:40 CEST] <C0nundrum> what ecnoder is it defaulting to ? i thought it was just sending it raw ?
[02:09:53 CEST] <c_14> flv1
[02:10:33 CEST] <c_14> If you want it raw, use -c:v rawvideo (though I don't know if that fits in flv)
[02:10:51 CEST] <C0nundrum> Hm well which encoder whould i try out. with this encoder i get pretty good results. what seems like less than 500ms but i want more than 13 fps :/
[02:12:22 CEST] <onyx_> hello
[02:12:42 CEST] <onyx_> Im trying to use nginx rtmp with ffmpeg
[02:12:53 CEST] <c_14> maybe something i-frame only?
[02:13:08 CEST] <onyx_> when I tried before it says libfdk_aac error or something
[02:13:19 CEST] <onyx_> is there a way to use another audio other than libdfk?
[02:13:25 CEST] <c_14> -c:a aac -strict -2
[02:13:35 CEST] <onyx_> would that work on iphone?
[02:13:50 CEST] <c_14> It produces standards compliant AAC just like every other aac encoder.
[02:14:14 CEST] <waressearcher2> is aac codec better than libmp3lame ?
[02:14:43 CEST] <waressearcher2> I know its good for youtube uploads to have audio track in "aac" format but other than that, why use "aac" ?
[02:15:06 CEST] <c_14> In git ffmpeg, probably. Might soon even be about as good or better than libfdk_aac (don't quote me on that).
[02:16:58 CEST] <waressearcher2> libfdk_aac is the best implementation ?
[02:17:15 CEST] <c_14> Currently, yes.
[02:17:49 CEST] <waressearcher2> they have dfferent licenses ? thats why they can't be merged ?
[02:18:06 CEST] <c_14> The libfdk_aac license is incompatible with GPL
[02:18:12 CEST] <onyx_> like this? http://pastebin.com/BCMedBvF
[02:18:25 CEST] <c_14> onyx_: yep
[02:19:10 CEST] <onyx_> c_14: do you see anything on there that can be improved?
[02:19:58 CEST] <c_14> Do you need the -g 50 for something?
[02:20:10 CEST] <onyx_> I have no idea what that is
[02:20:34 CEST] <c_14> get rid of it
[02:20:41 CEST] <onyx_> ok
[02:21:21 CEST] <onyx_> what does that -g 50 do?
[02:21:46 CEST] <c_14> Sets the gop size to 50.
[02:22:06 CEST] <c_14> No point changing it from the default if you don't need it any different.
[02:22:43 CEST] <onyx_> basically what Im trying to do is restream a live feed from an IP camera to rtmp and hls
[02:23:00 CEST] <onyx_> rtmp for desktop browsers and hls for mobile
[02:24:04 CEST] <waressearcher2> will these options "-c:v h264 -vb 2000k -preset ultrafast" generate good quality or should I use another "preset" ?
[02:24:24 CEST] <c_14> Use the slowest preset you can stomach.
[02:27:43 CEST] <onyx_> mmm =( is not playing...
[02:28:55 CEST] <c_14> onyx_: any error?
[02:29:13 CEST] <onyx_> nginx started fine
[02:53:52 CEST] <waressearcher2> c_14: with "ultrafast" it took 11 minutes
[02:55:15 CEST] <waressearcher2> c_14: will I get smaller size with slower preset ? its a record of a desktop, do I need better "preset" ?
[02:57:20 CEST] <c_14> Since you're specifying an average bitrate, no the video won't get smaller.
[02:57:24 CEST] <c_14> The quality will get better.
[02:58:13 CEST] <waressearcher2> how to specify variable bitrate ?
[02:59:25 CEST] <c_14> There's only average bitrate, constant bitrate, and "constant quality" the last one being crf
[03:29:24 CEST] <votz> I want to find the nearest keyframe after N seconds in a video. ffprobe can't seek. How can ffmpeg be used to create a subtrack, centered around N seconds, where the H.264 stream therein is preserved such that ffprobe can analyze it and get the nearest keyframe relative to the original, larger file.
[03:39:43 CEST] <c_14> ffmpeg -ss $((time - 5)) -t 10 -i input -c copy out.mkv
[03:39:53 CEST] <c_14> ffprobe that and take the middle keyframe?
[03:40:33 CEST] <c_14> You can also change the deltas, -c copy means it'll pick the nearest keyframe anyway so a smaller delta should be fair game as well
[03:41:00 CEST] <votz> c_14: I'll try that. Thank you.
[03:41:40 CEST] <votz> c_14: The tricky part is stream copying (-c copy) will adjust the pts and dts values appropriately, thus destroying the time relationship with the original video.
[03:41:45 CEST] <votz> -copyts looks like the key to solving this problem.
[04:50:53 CEST] <waressearcher2> c_14: it took 11 minutes on "ultrafast" preset, now I changed "-preset" from "ultrafast" to "medium" and its 3 hours allready and it still converting
[04:52:28 CEST] <c_14> Yes, each step down makes it take longer for less qualitative gain compared to the previous step.
[04:52:34 CEST] <c_14> Diminishing returns and all taht.
[04:52:36 CEST] <c_14> *that
[05:05:05 CEST] <waressearcher2> actually it took 70 minutes
[05:05:17 CEST] <waressearcher2> but still noticeably longer than 11 minutes
[06:27:14 CEST] <sr> im trying to figure out how to get ffmpeg to use more than a single core.. i tried using the -threads switch but that doesnt help.. google isnt helping either.. any ideas?
[06:38:32 CEST] <na-g> Not all codecs support multi-thread encoding. Which one are you using?
[06:51:26 CEST] <sr> im trying to extract the audio from a video, looks like im trying to go from aac to mp3 i guess
[06:54:28 CEST] <votz> sr: lame is single threaded iirc
[06:57:57 CEST] <sr> ah
[07:47:43 CEST] <koz_> Hi! I normally use this command to convert video: ffmpeg -i "$1" -codec:a opus -b:a 48k -codec:v vp9 -crf 10 -b:v 500k -cpu-used 8 -threads 4 "$2" (with appropriate $1, $2). Is there a way I can modify this command so it only converts a 5-minute segment from 30 minutes into the video?
[08:17:46 CEST] <relaxed> koz_: -t 00:05:00
[08:18:11 CEST] <relaxed> oh, wait
[08:18:50 CEST] <relaxed> that will work
[08:19:19 CEST] <relaxed> -ss 00:30:00 -t 00:05:00
[08:23:39 CEST] <koz_> relaxed: Thanks!
[08:34:20 CEST] <koz_> relaxed: I keep getting this: [aac @ 0x562c173c16e0] element type mismatch 3 != 0
[08:34:20 CEST] <koz_>     Last message repeated 198 times
[08:34:29 CEST] <koz_> Like, over and over and over again.
[08:47:27 CEST] <relaxed> koz_: you can ignore them
[08:48:32 CEST] <koz_> OK.
[08:49:03 CEST] <koz_> Why do they occur?
[09:07:14 CEST] <koz_> relaxed: Why do these messages even come up?
[09:11:24 CEST] <koz_> Also, holy crap do I get shit fps on encoding into VP9...
[09:16:57 CEST] <chungy> yeah vp9 badly needs hardware acceleration (IIRC, nvidia 970/980 cards can do it)
[09:18:03 CEST] <koz_> chungy: Via CUDA, I assume?
[09:18:40 CEST] <chungy> not sure
[09:18:49 CEST] <chungy> nvenc was the API name
[09:19:07 CEST] <koz_> 970/980 is Maxwell architecture, IIRC. Will look into nvenc.
[09:19:23 CEST] <chungy> I might be totally mistaken, mind you :P
[09:19:32 CEST] <koz_> Huh, nvenc is available as early as Kepler.
[09:19:38 CEST] <koz_> It's an API for an ASIC block.
[09:19:48 CEST] <koz_> (designed for H264 apparently)
[09:21:55 CEST] <koz_> chungy: Do you know (even in rough terms) why CPUs are so bad at video encoding?
[09:22:06 CEST] <koz_> I.e. why do we even need GPU-embedded ASICs to do it fast?
[09:23:06 CEST] <chungy> I don't
[09:23:23 CEST] <chungy> most mobile devices have ASICs for H.264 and maybe HEVC too
[09:23:44 CEST] <koz_> Well, hopefully Daala will blow them out the sky.
[09:24:35 CEST] <chungy> hehe yes, it's supposed to beat VP9 too
[09:24:47 CEST] <JEEB> koz_: uhh, cpus are pretty good at encoding, unlike gpus. asics are for a different use case
[09:24:57 CEST] <chungy> actually hardware VP8 and VP9 is common on mobiles too
[09:25:01 CEST] <koz_> JEEB: Enlighten me?
[09:25:14 CEST] <koz_> I'm genuinely not knowledgeable here, and would like to know why we need those ASICs.
[09:25:48 CEST] <JEEB> it's for the people who don't want or can't use cpu cycles
[09:25:55 CEST] <JEEB> or use cases
[09:25:58 CEST] <koz_> Ah.
[09:26:06 CEST] <koz_> How does it compete with just using a CPU?
[09:26:06 CEST] <chungy> It's no secret that HEVC and VP9 encoding are extremely slow on CPUs ...
[09:26:23 CEST] <chungy> There's probably complicated math reasons
[09:26:43 CEST] <chungy> like, bitcoin mining is dominated by ASICs that do the sole job of computing sha256 hashes really, really fast.
[09:27:19 CEST] <JEEB> chungy: you can tweak that and botg implementations are in early stages. also asic-based cosing is never aimed for high comptession
[09:27:34 CEST] <koz_> botg?
[09:27:53 CEST] <JEEB> i'm on a tpuchscreen device on a bus :p
[09:28:04 CEST] <chungy> Ah, both
[09:28:18 CEST] <koz_> Right, I see.
[09:28:31 CEST] Action: koz_ is actually curious why they're so slow now.
[09:28:57 CEST] <JEEB> well remember x264 and 2006 or so?
[09:29:23 CEST] <chungy> tbh I don't remeber x264 before around 2012-2013... I'm kind of late on the encoding thing
[09:29:31 CEST] <chungy> in 2006 I knew of xvid and that was about it
[09:29:50 CEST] <JEEB> my encodes on amd pcs were roighly what i get off my 4790k now with x265, aiming at high compression
[09:30:39 CEST] <JEEB> of course, by now x264 is so well optimized in psychovisuals and speed + cpus have gone forward
[09:31:20 CEST] <JEEB> vp9 is a sad case because google doesn't care about general public encoding :p
[09:31:41 CEST] <koz_> JEEB: Have they left it to a few enthusiasts or something?
[09:31:42 CEST] <chungy> *nods*
[09:31:48 CEST] <koz_> (and plus, they're already on VP10 apparently...)
[09:31:52 CEST] <chungy> I'd prefer VP9 generally because it lacks all the patent drama
[09:33:01 CEST] <JEEB> well give me a good encoder and I'll side with that :p and as vp9 is largely copypasta'd from hevc it's better than avc as a spec
[09:33:13 CEST] <JEEB> if only we had a spec, that is
[09:34:41 CEST] <JEEB> anyways asic coding is meant to not be optimized for compression but rather usually for low latency or speed. not that they don't get better with new generations of course
[09:36:14 CEST] <JEEB> intel's haswell+ avc coding seemed usable if you didn't want to use your cpu f.ex. and I think most recent nvidia weren't too far from that either
[09:37:34 CEST] <JEEB> never tested them too much because I just needed the compression and could use a cpu
[09:38:41 CEST] Action: koz_ wonders how many SLoC the vp9 encoder/decoder thing is.
[09:52:09 CEST] <noone_> hi guys
[09:52:26 CEST] <noone_> to have hls support in ffmpeg, do I need to compile it with --enable-nonfree option ?
[09:52:26 CEST] <koz_> Hi noone_.
[09:53:23 CEST] <noone_> as I always get the following error: Unrecognized option 'hls_segment_filename'.
[09:53:55 CEST] <noone_> or am I doing something wrong ?
[09:58:16 CEST] <DHE> your version of ffmpeg up to date?   ffmpeg -h full # and search for 'hls'
[10:06:02 CEST] <noone_> DHE thanks for the reply, I see it now with the `-h full` flag
[10:06:26 CEST] <koz_> How can I check what the bitrates for a .mp4 video file are?
[10:06:41 CEST] <noone_> I used the latest  static build version from the site
[10:06:58 CEST] <noone_> it's working now, thanks guys
[10:22:10 CEST] <koz_> How can I check what the bitrates for a .mp4 video file are?
[10:50:23 CEST] <koz_> Also, how can I enable the -lossless and -frame-parallel options for VP9 in FFMPEG?
[10:56:08 CEST] <relaxed> koz_: ffmpeg -h encoder=vp9
[11:02:02 CEST] <Nindustries> Hello
[11:02:16 CEST] <Nindustries> Is passmark score the best guideline to use when picking out CPUs ?
[11:06:46 CEST] <quovadis> Hi, simple question but cannot find any snippets: how to make a video form a set of raw rgb565 image ?
[11:07:14 CEST] <quovadis> "ffmpeg -framerate 25 -f image2 -pixel_format rgb565 -video_size 640x480 -i out_%04d.raw -b 65536k out.mp4" return Decoder (codec id 0) not found for input stream #0:0
[11:12:21 CEST] <durandal_1707> add -c:v rawvideo before -i
[11:16:41 CEST] <quovadis> durandal_1707: Thanks a lot !
[11:27:21 CEST] <quovadis> durandal_1707: by the way... how can I simply view an rgb565 with ffplay ? ffmpeg -f rawvideo -pixel_format rgb565 -s 640x480 -i out_0000.raw does not work
[11:28:13 CEST] <quovadis> durandal_1707: nor ffmpeg  -f image2 -pixel_format rgb565   -video_size 640x480  -c:v rawvideo -i out_0002.raw
[11:47:50 CEST] <smirky> hello
[11:48:16 CEST] <relaxed> quovadis: you probably cat them together
[11:49:31 CEST] <smirky> I'm using ffserver and ffmpeg to feed a .ffm file, recording the camera and the microphone of my laptop. The image seems to be rotated with 180 degrees. I tried toggling vflip,hflip by inserting -vf but that didn't help. If I dump the recording to a regular file, not a http:// link, the recording seems to be ok. This narrows down things to the ffserver itself I guess. Any ideas what I can set to rotate it from the ffserver.conf?
[12:01:24 CEST] <saste> smirky, ffserver won't apply filtering
[12:01:34 CEST] <saste> there is a workaround though
[12:01:47 CEST] <smirky> that's what I'm here to ask for : ))
[12:01:58 CEST] <smirky> what to apply to ffserver
[12:02:39 CEST] <smirky> if the workaround is "ffplay -vf "hflip,vflip", this won't do
[12:03:00 CEST] <saste> ffserver
[12:03:13 CEST] <saste> -override_ffserver I mean
[12:03:23 CEST] <smirky> huh?
[12:03:43 CEST] <smirky> oh, nice
[12:03:47 CEST] <smirky> let me check
[12:05:30 CEST] <saste> it's an ffmpeg option
[12:07:32 CEST] <smirky> I know
[12:07:41 CEST] <smirky> but I face invalid data found when processing input
[12:07:57 CEST] <smirky> this is probably because I have to pass correct variables
[12:08:08 CEST] <smirky> and before, ffserver was fixing it
[12:12:26 CEST] <saste> smirky, what's you command? use pastebin
[12:14:21 CEST] <smirky> saste: https://paste.smirky.net/view/raw/69799d21
[12:14:33 CEST] <smirky> this is the script only, I'll show you the ffserver.conf
[12:15:07 CEST] <smirky> saste: https://paste.smirky.net/view/raw/16d3e0ad
[12:19:44 CEST] <saste> smirky, try to remove all AVCodec options from the ffserver file
[12:20:06 CEST] <saste> also, what kind of error do you get? from ffmpeg or from ffserver?
[12:20:32 CEST] <koz_> What arguments to ffmpeg should I add if I want to include a subtitle stream?
[12:20:40 CEST] <smirky1> http://192.168.1.105:8090/live.flv: Invalid data found when processing input
[12:20:45 CEST] <smirky1> saste: ^
[12:21:44 CEST] <smirky1> oops
[12:21:51 CEST] <smirky1> those errors are not related to ffmpeg or ffserver
[12:21:56 CEST] <smirky1> that's ffplay
[12:22:16 CEST] <saste> smirky, mmh ok
[12:22:47 CEST] <smirky1> [flv @ 0x55e26e978b90] Codec for stream 0 does not use global headers but container format requires global headers
[12:22:47 CEST] <smirky1> [flv @ 0x55e26e978b90] Codec for stream 1 does not use global headers but container format requires global headers
[12:22:47 CEST] <smirky1> [flv @ 0x55e26e978b90] FLV does not support sample rate 48000, choose from (44100, 22050, 11025)
[12:22:47 CEST] <smirky1> [flv @ 0x55e26e978b90] Audio codec mp3 not compatible with flv
[12:22:49 CEST] <smirky1> there we go
[12:23:03 CEST] <smirky1> I think I found it, need to set bitrate to 44100
[12:23:06 CEST] <saste> allright add those flags to the ffmpeg command
[12:23:16 CEST] <saste> smirky, yes
[12:23:18 CEST] <smirky1> I had ffserver 2>/dev/nill
[12:23:31 CEST] <smirky1> and I remembered that it spits out stuff :)
[12:24:25 CEST] <smirky1> on the other hand,  Audio codec mp3 not compatible with flv
[12:24:27 CEST] <smirky1> hm.....
[12:27:53 CEST] <saste>  Audio codec mp3 not compatible with flv?
[12:27:55 CEST] <saste> sure it is
[12:29:01 CEST] <smirky1> fixed that
[12:29:07 CEST] <smirky1> I got it working properly
[12:29:22 CEST] <smirky1> now some sync and bitrate settings
[12:29:27 CEST] <smirky1> and I'm done with the script
[12:29:57 CEST] <saste> smirky, do you know why you was getting that error about MP3 and FLV?
[12:30:14 CEST] <smirky1> no, but -ar 44100 made it go away
[12:31:38 CEST] <smirky1> saste: any thoughts?
[12:32:24 CEST] <koz_> This AAC thing is annoying the shit out of me.
[12:33:08 CEST] <saste> smirky, I was checking the code, it is silly error reporting
[12:33:51 CEST] <saste> i'll send a patch if $gods want
[12:33:54 CEST] <koz_> Also, how would I add a subtitle stream to something using ffmpeg?
[12:34:16 CEST] <saste> koz_, you reading from files?
[12:34:28 CEST] <saste> what's wrong with -i subs.fmt
[12:36:56 CEST] <koz_> saste: Yeah, from a file, adding to a video which I'm also encoding to a different format and into a different container.
[12:36:59 CEST] <koz_> My sub file is an srt.
[12:37:21 CEST] <saste> koz_, show commandline and output in a pastebin
[12:37:25 CEST] <smirky1> how should I calculate what bitrate I need
[12:37:29 CEST] <smirky1> for the webcam
[12:37:40 CEST] <smirky1> not just inserting random numbers
[12:39:11 CEST] <koz_> saste: http://dpaste.com/059MPKP
[12:39:48 CEST] <saste> koz_, show the ffmpeg output
[12:39:59 CEST] <saste> also, you already have the subtitle stream in the input file?
[12:40:23 CEST] <saste> I mean the *console* ffmpeg output
[12:40:44 CEST] <saste> !paste
[12:41:34 CEST] <koz_> saste: No, the subs are in a separate file.
[12:42:19 CEST] <saste> koz_: so you need to add this ffmpeg -i subs.srt -i in.mp4 ...
[12:42:49 CEST] <koz_> Ah, is that all? Neat.
[12:43:29 CEST] <koz_> saste: Is there a way to shut this up? [aac @ 0x562c110c7ea0] element type mismatch 3 != 0
[12:43:35 CEST] <koz_> It's annoying as hell.
[12:44:03 CEST] <saste> koz_, I don't think so
[12:44:16 CEST] <saste> it could be a broken file or something
[12:45:38 CEST] <koz_> Well, there's only one input file (I did tests without subs).
[12:45:57 CEST] <koz_> I have no idea if this is fixable or something...
[13:15:46 CEST] <quovadis> Does anyone knows how to display and rgb565 raw image with ffplay ?
[13:19:27 CEST] <durandal_1707> Just replace ffmpeg with ffplay
[13:23:15 CEST] <satiender> Is any one compile ffmpeg for arm
[13:23:28 CEST] <satiender> Or use ffmpeg API for arm
[13:23:38 CEST] <satiender> ??
[13:23:41 CEST] <satiender> API
[13:23:47 CEST] <satiender> ??
[13:26:52 CEST] <JEEB> yes, it can be compiled and used on ARM just fine
[13:26:55 CEST] <JEEB> see fate.ffmpeg.org
[13:27:02 CEST] <JEEB> there's plenty of ARM testing
[13:37:50 CEST] <satiender> ok thanks
[13:38:33 CEST] <smirky1> saste: forgot to thank you, btw
[13:46:34 CEST] <quovadis> durandal_1707: I tired but get error: "Failed to set value 'rawvideo' for option 'c:v'" for the command ffplay -framerate 25 -f image2 -pixel_format rgb565 -video_size 640x480 -c:v rawvideo -i out_0000.raw
[13:49:34 CEST] <durandal_1707> -f rawvideo not image2
[13:50:14 CEST] <relaxed> omit -c:v rawvideo
[13:50:22 CEST] <durandal_1707> or -codec instead of -c:v
[13:50:56 CEST] <quovadis> durandal_1707: thanks it work well
[13:54:39 CEST] <daggs1> greetings, how can I replace a video track within a file?
[17:15:18 CEST] <DrSlony> Hey, does ffmpeg need sequential input image names when creating a video from image files, or cna a glob *.jpg work even if the filenames have no order?
[17:16:55 CEST] <c_14> glob works
[17:20:00 CEST] <DrSlony> thanks
[18:55:26 CEST] <Bermond> relaxed: are you there?
[19:23:44 CEST] <Bermond> relaxed: hi
[19:44:36 CEST] <waressearcher2> one more time, will it increase quality if I change "-c:v h264 -vb 2000k -preset ultrafast" to "-c:v h264 -vb 2000k -preset medium" ?
[19:44:55 CEST] <manel2020> hello
[19:44:56 CEST] <waressearcher2> it will be encoded 7 times longer but will it brings better quality ?
[19:45:02 CEST] <c_14> waressearcher2: yes
[19:45:21 CEST] <waressearcher2> c_14: have you seen terminator
[19:45:22 CEST] <waressearcher2> ?
[19:45:28 CEST] <c_14> yes?
[19:45:37 CEST] <waressearcher2> did they sent T-800 back in future in that movie ?
[19:45:48 CEST] <waressearcher2> back in past
[19:45:59 CEST] <waressearcher2> to where "terminator 2" begins
[19:47:17 CEST] <c_14> No, what does this have to do with ffmpeg?
[21:08:11 CEST] <JodaZ> is it possible when seeking to not have it touch the audio packets at all and just cut them cleanly?
[21:16:58 CEST] <crot^^^> Hello all, Im having some issues on a fresh ubuntu 15.04 with a compile following the guide complie guide
[21:17:43 CEST] <crot^^^> the compile goes well but when I execute ffmpeg it seems to stall after it presents libpostproc.....
[21:18:49 CEST] <crot^^^> has anyone expirenced this
[21:22:02 CEST] <aiena> I have some video tutorials whose audio qulaity is not very good for comfortable listening I would like to split the audio component from the file and then enhance it and combine it again with the video to create a new video. FFmpeg has lots of options and I am not sure with methodology is appropriate.
[21:22:17 CEST] <aiena> *which
[21:22:42 CEST] <aiena> can someone please guide me. I am planning to use audacity to enhance the audio and then merge it back with the video
[21:31:24 CEST] <crot^^^> http://pastebin.com/jQnb57EA     is a showing of what I'm executing with and output in console
[21:36:04 CEST] <aiena> googling has me just more confused because I am finding variety of approaches and no clear answer
[21:38:31 CEST] <DelphiWorld> hi ffmpegsters
[21:38:42 CEST] <DelphiWorld> i'm sponing ffmpeg process from nginx rtmp server
[21:38:49 CEST] <DelphiWorld> to transcode udp ts to rtmp
[21:38:54 CEST] <DelphiWorld> mpeg to h264/AAC
[21:39:08 CEST] <DelphiWorld> but sometime channel dont have video
[21:39:15 CEST] <DelphiWorld> i'm transcoding to h264 baseline 3.0
[22:27:51 CEST] <crot^^^> anyone around with closed caption experience
[22:28:37 CEST] <DHE> what kind?
[22:29:52 CEST] <crot^^^> CEA-608
[22:31:03 CEST] <DHE> not me then
[22:31:48 CEST] <crot^^^> what we are doing is taking an mpeg2 dumped live out of an hdhomerun and transcoding it into x264/mpeg4 but after the transcode I cant see it in the stream
[22:32:16 CEST] <DHE> oh. oh that's easy. I have a patch for that
[22:32:36 CEST] <crot^^^> you rock!
[22:34:07 CEST] <DHE> I run a mostly uptodate Git checkout. Add this patch onto it: https://github.com/DeHackEd/FFmpeg/commit/05d5069ca3
[22:34:13 CEST] <DHE> I've been meaning to offer it up to ffmpeg
[22:34:35 CEST] <crot^^^> let me give it a shot
[22:48:17 CEST] <crot^^^> I'll let you know once this finishings recompiling if it works
[23:17:45 CEST] <DelphiWorld> yo
[23:18:05 CEST] <DelphiWorld> how should i know what passtrue codecs is supported on my tv over hdmi?
[23:32:38 CEST] <fred1807> I am looking for a GUI only for the cut job , any good minimalistic one ?
[23:33:34 CEST] <DHE> crot^^^: still building? :)
[23:37:06 CEST] <crot^^^> just finished
[23:37:19 CEST] <crot^^^> do i need any special switches
[23:37:29 CEST] <crot^^^> or arguments
[23:41:07 CEST] <DHE> add "-a53cc 1" before the output filename
[23:42:02 CEST] <fred1807> exactly what I was looking for http://www.popmedic.com/popmsplit/
[23:42:40 CEST] <fred1807> what a wonderfull piece of softwware to add to another wonderful code that is ffmpeg
[23:42:58 CEST] <fred1807> what a wonderfull time to be living in
[23:43:38 CEST] <crot^^^> showing 4 subtitle streams in media info no output of subs on vlc however
[23:44:27 CEST] <crot^^^> take that back there they are
[23:45:11 CEST] <DHE> so it works?
[23:45:47 CEST] <crot^^^> yeah once I find the correct language :)
[23:46:13 CEST] <crot^^^> thanks a bunch we really apprecaite
[00:00:00 CEST] --- Sat Sep 26 2015