April 2018 - ffmpeg-user - lists.ffmpeg.org

adjusting when Matroska clusters are restarted
by Dave Rice 16 Apr '18

16 Apr '18

Hi all, The command below creates a short matroska file with a video and audio track: ffmpeg -f lavfi -i testsrc2=r=30000/1001:s=1920x1080:d=1 -f lavfi -i aevalsrc=exprs=0:d=1 -c:v v210 -c:a pcm_s24le -y -vframes 2 test.mkv ffmpeg version N-51550-g41d7c4d381 Copyright (c) 2000-2018 the FFmpeg developers built with Apple LLVM version 9.0.0 (clang-900.0.38) configuration: --prefix=/usr/local/Cellar/ffmpeg/HEAD-41d7c4d --enable-shared --enable-pthreads --enable-version3 --enable-hardcoded-tables --enable-avresample --cc=clang --host-cflags= --host-ldflags= --enable-gpl --enable-ffplay --enable-libfreetype --enable-libmp3lame --enable-libx264 --enable-libxvid --enable-opencl --enable-videotoolbox --disable-lzma libavutil 56. 9.100 / 56. 9.100 libavcodec 58. 14.100 / 58. 14.100 libavformat 58. 10.100 / 58. 10.100 libavdevice 58. 2.100 / 58. 2.100 libavfilter 7. 13.100 / 7. 13.100 libavresample 4. 0. 0 / 4. 0. 0 libswscale 5. 0.102 / 5. 0.102 libswresample 3. 0.101 / 3. 0.101 libpostproc 55. 0.100 / 55. 0.100 Input #0, lavfi, from 'testsrc2=r=30000/1001:s=1920x1080:d=1': Duration: N/A, start: 0.000000, bitrate: N/A Stream #0:0: Video: rawvideo (I420 / 0x30323449), yuv420p, 1920x1080 [SAR 1:1 DAR 16:9], 29.97 tbr, 29.97 tbn, 29.97 tbc Input #1, lavfi, from 'aevalsrc=exprs=0:d=1': Duration: N/A, start: 0.000000, bitrate: 2822 kb/s Stream #1:0: Audio: pcm_f64le, 44100 Hz, mono, dbl, 2822 kb/s Stream mapping: Stream #0:0 -> #0:0 (rawvideo (native) -> v210 (native)) Stream #1:0 -> #0:1 (pcm_f64le (native) -> pcm_s24le (native)) Press [q] to stop, [?] for help Output #0, matroska, to 'test.mkv': Metadata: encoder : Lavf58.10.100 Stream #0:0: Video: v210 (v210 / 0x30313276), yuv422p(progressive), 1920x1080 [SAR 1:1 DAR 16:9], q=2-31, 1325778 kb/s, 29.97 fps, 1k tbn, 29.97 tbc Metadata: encoder : Lavc58.14.100 v210 Stream #0:1: Audio: pcm_s24le ([1][0][0][0] / 0x0001), 44100 Hz, mono, s32, 1058 kb/s Metadata: encoder : Lavc58.14.100 pcm_s24le frame= 2 fps=0.0 q=-0.0 Lsize= 10804kB time=00:00:00.03 bitrate=2603133.6kbits/s speed=0.537x video:10800kB audio:3kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.009456% When running this through mkvalidator I get a warning: mkvalidator test.mkv WRN0C2: The timecode of the Cluster at 5530555 is not incrementing (may be intentional) mkvalidator 0.5.2: the file appears to be valid file "test.mkv" created with Lavf58.10.100 / Lavf58.10.100 With `mkvinfo -v test.mkv`, I can see how the encoded data is arranged into Clusters: |+ Cluster | + Cluster timestamp: 00:00:00.000000000 | + Simple block: key, track number 1, 1 frame(s), timestamp 00:00:00.000000000 | + Frame with size 5529600 |+ Cluster | + Cluster timestamp: 00:00:00.000000000 | + Simple block: key, track number 2, 1 frame(s), timestamp 00:00:00.000000000 | + Frame with size 3072 | + Simple block: key, track number 1, 1 frame(s), timestamp 00:00:00.033000000 | + Frame with size 5529600 So track 2 at 0 is stored in a Cluster with track 1 at 00.033. I see the Matroska muxer has options for Clusters with cluster_size_limit and cluster_time_limit but is there a way to start a new Cluster for every frame (or every keyframe) from track 1? Ideally I’d like the Cluster timestamps to be unique and timed along with the video track or at least to mux in a way that avoids the mkvalidator warning. Dave Rice

1 0

Alex Mihalev - Question on having output to a Decklink card
by Alex Mihalev 16 Apr '18

16 Apr '18

Hello there! I am new to FFMPEG and would like to apologize if I am asking an obvious question, though, I looked and didn't find much information on the topic. It is about playing a Multicast stream to Decklink Monitor Card SDI output. I am using Debian 8 with its "apt-get install ffmpeg" sort of FFMPEG, Do i need to make a compile and somehow insert the Decklink into it? Thank you!

1 0

ERROR: libfdk_aac not found
by La poste 15 Apr '18

15 Apr '18

Hi ! I was installing ffmpeg with this page : https://github.com/KhaosT/homebridge-camera-ffmpeg/wiki/Raspberry-PI <https://github.com/KhaosT/homebridge-camera-ffmpeg/wiki/Raspberry-PI> git clone https://github.com/FFmpeg/FFmpeg.git <https://github.com/FFmpeg/FFmpeg.git> cd FFmpeg ./configure --prefix=/usr/local --arch=armel --target-os=linux --enable-omx-rpi --enable-nonfree --enable-gpl --enable-libfdk-aac --enable-mmal --enable-libx264 --enable-decoder=h264 --enable-network --enable-protocol=tcp --enable-demuxer=rtsp And : ERROR: libfdk_aac not found If you think configure made a mistake, make sure you are using the latest version from Git. If the latest version fails, report the problem to the ffmpeg-user(a)ffmpeg.org mailing list or IRC #ffmpeg on irc.freenode.net.

4 3

Convert images to DNXHD 444
by Matt Kitcat 15 Apr '18

15 Apr '18

Hi All I have been trying to convert an image sequence (16 bit tiff) to DNXHD 444. It can be done with Quicktime pro on Windows but I have completely failed with ffmpeg. ffmpeg -r 23.976 -i C:\Users\Matt\Desktop\Newfolder\104_0020gn_01_X1_0%3d.tif -c:v dnxhd -profile:v 3 -s 1920x1080 -r 23.976 -y C:\Users\Matt\Desktop\Newfolder\bt.mov ffmpeg –h encode=dnxhd gives the following: General capabilities: threads Threading capabilities: frame and slice Supported pixel formats: yuv422p yuv422p10le yuv444p10le gbrp10le dnxhd AVOptions: -nitris_compat <boolean> E..V.... encode with Avid Nitris compatibility (default false) -ibias <int> E..V.... intra quant bias (from INT_MIN to INT_MAX) (default 0) -profile <int> E..V.... (from 0 to 5) (default dnxhd) dnxhd E..V.... dnxhr_444 E..V.... dnxhr_hqx E..V.... dnxhr_hq E..V.... dnxhr_sq E..V.... dnxhr_lb E..V.... So I am unsure how to achieve a pixel format of yuv444p10le I have been at this for days so if anyone can help I would be very grateful. I wonder if it is a bit rate setting but I can’t see what bit rate to use with the framerate and pixel format. I hope this question makes sense

7 18

Supported pix_fmts for decoders using -h decoder
by Kieran O Leary 15 Apr '18

15 Apr '18

Hi, I was wondering if there was a reason why supported pix_fmts are rarely seen when calling `ffmpeg -h decoder=$codec`. I did a quick audit and it looks like pix_fmts are declared in the help section of some decoders (see below), but for the most part it is left out. One benefit that I could see would be that it would allow scripts to check if an installed version of ffmpeg can decode a pixel format that might have been added recently, rather than just catching the errors. Would this be of any benefit, and would patches be welcome? Decoder flashsv [Flash Screen Video v1]: General capabilities: dr1 Threading capabilities: none Supported pixel formats: bgr24 Decoder flashsv2 [Flash Screen Video v2]: General capabilities: dr1 Threading capabilities: none Supported pixel formats: bgr24 -- Decoder flv [FLV / Sorenson Spark / Sorenson H.263 (Flash Video)]: General capabilities: horizband dr1 Threading capabilities: none Supported pixel formats: yuv420p -- Decoder h263 [H.263 / H.263-1996, H.263+ / H.263-1998 / H.263 version 2]: General capabilities: horizband dr1 trunc delay Threading capabilities: none Supported pixel formats: vaapi_vld vdpau yuv420p Decoder h263i [Intel H.263]: General capabilities: horizband dr1 Threading capabilities: none Supported pixel formats: yuv420p Decoder h263p [H.263 / H.263-1996, H.263+ / H.263-1998 / H.263 version 2]: General capabilities: horizband dr1 trunc delay Threading capabilities: none Supported pixel formats: vaapi_vld vdpau yuv420p -- Decoder mpeg4 [MPEG-4 part 2]: General capabilities: horizband dr1 trunc delay threads Threading capabilities: frame Supported pixel formats: vaapi_vld vdpau yuv420p -- Decoder msmpeg4v1 [MPEG-4 part 2 Microsoft variant version 1]: General capabilities: horizband dr1 Threading capabilities: none Supported pixel formats: yuv420p Decoder msmpeg4v2 [MPEG-4 part 2 Microsoft variant version 2]: General capabilities: horizband dr1 Threading capabilities: none Supported pixel formats: yuv420p Decoder msmpeg4 [MPEG-4 part 2 Microsoft variant version 3]: General capabilities: horizband dr1 Threading capabilities: none Supported pixel formats: yuv420p -- Decoder rv10 [RealVideo 1.0]: General capabilities: dr1 Threading capabilities: none Supported pixel formats: yuv420p Decoder rv20 [RealVideo 2.0]: General capabilities: dr1 delay Threading capabilities: none Supported pixel formats: yuv420p Decoder rv30 [RealVideo 3.0]: General capabilities: dr1 delay threads Threading capabilities: frame Supported pixel formats: yuv420p Decoder rv40 [RealVideo 4.0]: General capabilities: dr1 delay threads Threading capabilities: frame Supported pixel formats: yuv420p -- Decoder svq1 [Sorenson Vector Quantizer 1 / Sorenson Video 1 / SVQ1]: General capabilities: dr1 Threading capabilities: none Supported pixel formats: yuv410p Decoder svq3 [Sorenson Vector Quantizer 3 / Sorenson Video 3 / SVQ3]: General capabilities: horizband dr1 delay Threading capabilities: none Supported pixel formats: yuvj420p -- Decoder vc1 [SMPTE VC-1]: General capabilities: dr1 delay Threading capabilities: none Supported pixel formats: vaapi_vld vdpau yuv420p Decoder vc1image [Windows Media Video 9 Image v2]: General capabilities: dr1 Threading capabilities: none Supported pixel formats: yuv420p -- Decoder wmv1 [Windows Media Video 7]: General capabilities: horizband dr1 Threading capabilities: none Supported pixel formats: yuv420p Decoder wmv2 [Windows Media Video 8]: General capabilities: horizband dr1 Threading capabilities: none Supported pixel formats: yuv420p Decoder wmv3 [Windows Media Video 9]: General capabilities: dr1 delay Threading capabilities: none Supported pixel formats: vaapi_vld vdpau yuv420p Decoder wmv3image [Windows Media Video 9 Image]: General capabilities: dr1 Threading capabilities: none Supported pixel formats: yuv420p -- Decoder xface [X-face image]: General capabilities: none Threading capabilities: none Supported pixel formats: monow

2 1

Lavfi video generator slow performance
by Elliott Balsley 13 Apr '18

13 Apr '18

Hello. I'm trying to use ffmpeg to generate test images, and I'm not getting very fast performance. Is there any way to improve speed? If Lavfi is the problem, then is there any other high-performance virtual input I can use instead? I would actually prefer to use something more complex, like mandelbrot, but I just changed to solid red color for this test. Example command: ffmpeg -f lavfi -i color=color=red -t 60 -pix_fmt rgb48le -y -s 3840x2160 ffmpeg\%4d.tiff ffmpeg version N-90610-g2accdd3871 This runs at 6fps. My CPU usage is only about 1% and my disk has plenty of bandwidth (3000 MBps). Thanks, Elliott

8 15

Finding / filtering 25 fps content on a 720p50 TS video, while preserving 50 fps content as 50 fps
by Mattias Brunschen 12 Apr '18

12 Apr '18

Hi Folks, I often record films on TV using "TV Headend" and use ffmpeg (h.264 with -crf18) do compress the videos to better file sizes for permanent storage. Source is a public service TV channel, permanentely transmitting in 720p50. The recorded files are *.ts files in 720p50, video is aac / h.264 encoded. Now, some newer films have true 50 frames per second, so that's okay. Some older films, esp. filmed on chemical film stock and hires-scanned to video, have effectively 25 true *different* frames per second, allthough the video stream is still in 50 fps. These videos save file size, if I can re-code them with only 25 fps, thus throwing away the 25 redundant "filler" frames. I would like to have a generic re-coding engine, where I just drop the files and the script analyzes the file and calls ffmpeg with the needed options. Currently I watch a part of each film frame by frame, and if I see changes in only every second frame, I add the option "-framerate 25" th the ffmpeg call. So: Is there a generic filter option in ffmpeg, to say: "Transcode 50p stream with 50 different fps and keep 50 fps, while in 50p stream with 25 different is downgraded to 25 fps video."? Or is there a way for ffprobe to tell me, if the video has effectively 25 or 50 fps, even if the video stream is allways in p50? In this way, I could first determine the real fps and include an option "-framerate 25" in the following ffmpeg call, if needed. Regards, Mattias

1 0

How to drop frames to limit memory usage (or keep latency low)?
by Adam Nielsen 11 Apr '18

11 Apr '18

Hi all, I am transcoding footage from a security camera from MJPEG into MPEG2, but the machine I'm doing it on can't quite manage it in real-time. It's pegged at 100% CPU usage but after 10 minutes it exhausts the available memory and the output video becomes full of visual artefacts. As far as I can tell, it's not a memory leak but rather the input frames from the camera being buffered, because the transcoding is happening slower than real-time. So I am looking for some way to prevent this buffer from growing so large that it causes problems. Since it's security camera footage, I don't mind if a few frames are dropped here and there. If I could set a memory limit and have frames dropped once this limit is reached, that would be good. Alternatively a low latency/realtime option that dropped frames if the output was getting too far behind the input would perhaps be even better. I've had a look through the manpage but I can't find any options that can do this. The -r and -vsync options seem to only drop frames to match a constant framerate (rather than trying to minimise latency) and -frame_drop_threshold looks to be similar, keeping up with timestamps rather than affecting latency. Is there some option that will effectively drop frames in order to prevent the latency from becoming too great? This is what I'm running now: ffmpeg -overrun_nonfatal 1 -y -i udp://1.2.3.4 -c mpeg2video -b:v 2M -vf transpose=2 -preset ultrafast -an -r 30 out.mkv Many thanks, Adam.

2 2

MPEG-DASH
by Jeff Jones 10 Apr '18

10 Apr '18

I'm trying to decipher the documentation to create a manifest for a series of mpeg-dash files. I've created the h.264 encoded videos and the aac encoded audio file. It is't real clear on how to create the manifest file associated with this format. There are a few examples for webm formatted manifests, but nothing for mpeg. I have the following from the documentation: ffmpeg -re -i <input> -map 0 -map 0 -c:a libfdk_aac -c:v libx264 -b:v:0 800k -b:v:1 300k -s:v:1 320x170 -profile:v:1 baseline -profile:v:0 main -bf 1 -keyint_min 120 -g 120 -sc_threshold 0 -b_strategy 0 -ar:a:1 22050 -use_timeline 1 -use_template 1 -window_size 5 -adaptation_sets "id=0,streams=v id=1,streams=a" -f dash /path/to/out.mpd It appears to be creating the h.264 video and aac audio from a live stream as well as creating the manifest, but the arguments don't appear to be defined. I don't see any info about the what the -use_timeline 1 - use_template 1 options are or if the manifest can be generated separately. Can anyone break this down for me? Thanks, Jeff...

2 1

x-axis range of drawgraph picture mode
by Dave Rice 09 Apr '18

09 Apr '18

Hi ffmpeg-user, I’m using drawgraph with slide=picture in order to make some still images of audiovisual metrics upon a timeline. I’d like to use drawgraph and drawtext in order to overlay some x-axis markers and label the timeline; however I’m unclear how the x-axis is represented in this mode. For instance, the example below plots lavfi.astats.1.Min_level for a 60 second sample, but the resulting image appears to have a x-axis that represents more than 60 seconds. The result appears to roughly plot about 80 seconds with the 60-80 second range being blank. ffmpeg -f lavfi -i "aevalsrc=d=60:exprs=sin(PI*t):s=48000,aformat=sample_fmts=dbl,astats=metadata=1:reset=1,adrawgraph=m1=lavfi.astats.1.Min_level:max=1:min=-1:slide=picture:s=1920x240" -y test.png Is there some math that I can use to determine the maximum value of the x-axis from the plotted image? Full output: ffmpeg -f lavfi -i "aevalsrc=d=60:exprs=sin(PI*t):s=48000,aformat=sample_fmts=dbl,astats=metadata=1:reset=1,adrawgraph=m1=lavfi.astats.1.Min_level:max=1:min=-1:slide=picture:s=1920x240" -y test.png ffmpeg version N-51550-g41d7c4d381 Copyright (c) 2000-2018 the FFmpeg developers built with Apple LLVM version 9.0.0 (clang-900.0.38) configuration: --prefix=/usr/local/Cellar/ffmpeg/HEAD-41d7c4d --enable-shared --enable-pthreads --enable-version3 --enable-hardcoded-tables --enable-avresample --cc=clang --host-cflags= --host-ldflags= --enable-gpl --enable-ffplay --enable-libfreetype --enable-libmp3lame --enable-libx264 --enable-libxvid --enable-opencl --enable-videotoolbox --disable-lzma libavutil 56. 9.100 / 56. 9.100 libavcodec 58. 14.100 / 58. 14.100 libavformat 58. 10.100 / 58. 10.100 libavdevice 58. 2.100 / 58. 2.100 libavfilter 7. 13.100 / 7. 13.100 libavresample 4. 0. 0 / 4. 0. 0 libswscale 5. 0.102 / 5. 0.102 libswresample 3. 0.101 / 3. 0.101 libpostproc 55. 0.100 / 55. 0.100 [Parsed_aevalsrc_0 @ 0x7f8035500000] EOF timestamp not reliable Input #0, lavfi, from 'aevalsrc=d=60:exprs=sin(PI*t):s=48000,aformat=sample_fmts=dbl,astats=metadata=1:reset=1,adrawgraph=m1=lavfi.astats.1.Min_level:max=1:min=-1:slide=picture:s=1920x240': Duration: N/A, start: 0.000000, bitrate: N/A Stream #0:0: Video: rawvideo (RGBA / 0x41424752), rgba, 1920x240 [SAR 1:1 DAR 8:1], 48k tbr, 48k tbn, 48k tbc Stream mapping: Stream #0:0 -> #0:0 (rawvideo (native) -> png (native)) Press [q] to stop, [?] for help [image2 @ 0x7f8036801400] Frame rate very high for a muxer not efficiently supporting it. Please consider specifying a lower framerate, a different muxer or -vsync 2 Output #0, image2, to 'test.png': Metadata: encoder : Lavf58.10.100 Stream #0:0: Video: png, rgba, 1920x240 [SAR 1:1 DAR 8:1], q=2-31, 200 kb/s, 48k fps, 48k tbn, 48k tbc Metadata: encoder : Lavc58.14.100 png frame= 1 fps=0.0 q=-0.0 Lsize=N/A time=00:00:00.00 bitrate=N/A speed=0.00116x video:14kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown [Parsed_astats_2 @ 0x7f8035500640] Channel: 1 [Parsed_astats_2 @ 0x7f8035500640] DC offset: -0.016786 [Parsed_astats_2 @ 0x7f8035500640] Min level: -0.033504 [Parsed_astats_2 @ 0x7f8035500640] Max level: 0.000000 [Parsed_astats_2 @ 0x7f8035500640] Min difference: 0.000065 [Parsed_astats_2 @ 0x7f8035500640] Max difference: 0.000065 [Parsed_astats_2 @ 0x7f8035500640] Mean difference: 0.000066 [Parsed_astats_2 @ 0x7f8035500640] RMS difference: 0.000066 [Parsed_astats_2 @ 0x7f8035500640] Peak level dB: -29.498054 [Parsed_astats_2 @ 0x7f8035500640] RMS level dB: -34.255905 [Parsed_astats_2 @ 0x7f8035500640] RMS peak dB: -34.255905 [Parsed_astats_2 @ 0x7f8035500640] RMS trough dB: -34.255905 [Parsed_astats_2 @ 0x7f8035500640] Crest factor: 1.729388 [Parsed_astats_2 @ 0x7f8035500640] Flat factor: 0.000000 [Parsed_astats_2 @ 0x7f8035500640] Peak count: 1 [Parsed_astats_2 @ 0x7f8035500640] Bit depth: 59/64 [Parsed_astats_2 @ 0x7f8035500640] Dynamic range: 60.204373 [Parsed_astats_2 @ 0x7f8035500640] Overall [Parsed_astats_2 @ 0x7f8035500640] DC offset: -0.016786 [Parsed_astats_2 @ 0x7f8035500640] Min level: -0.033504 [Parsed_astats_2 @ 0x7f8035500640] Max level: 0.000000 [Parsed_astats_2 @ 0x7f8035500640] Min difference: 0.000065 [Parsed_astats_2 @ 0x7f8035500640] Max difference: 0.000065 [Parsed_astats_2 @ 0x7f8035500640] Mean difference: 0.000066 [Parsed_astats_2 @ 0x7f8035500640] RMS difference: 0.000066 [Parsed_astats_2 @ 0x7f8035500640] Peak level dB: -29.498054 [Parsed_astats_2 @ 0x7f8035500640] RMS level dB: -34.255905 [Parsed_astats_2 @ 0x7f8035500640] RMS peak dB: -34.255905 [Parsed_astats_2 @ 0x7f8035500640] RMS trough dB: -34.255905 [Parsed_astats_2 @ 0x7f8035500640] Flat factor: 0.000000 [Parsed_astats_2 @ 0x7f8035500640] Peak count: 1.000000 [Parsed_astats_2 @ 0x7f8035500640] Bit depth: 59/64 [Parsed_astats_2 @ 0x7f8035500640] Number of samples: 512 Best Regards, Dave Rice

1 1