Using

  • For playback, a SAMSUNG SV-J160UM
  • For A/D conversion
  • For compression, ffmpeg on Linux

A/D conversion

To optimize quality, convert to raw video using the following EZ Grabber settings

  • Video:
    • Video format: NTSC_M
    • Video source = Composite
    • Picture adjustment: All 127
  • Record:
    • Record format: AVI

The older the cassettes, the shorter the conversion interval needed to avoid synchronization problems. Intervals between 10 and 30 minutes worked OK.

Result is raw video:

$ file raw-video.avi
raw-video.avi: RIFF (little-endian) data, AVI, 720 x 480, ~30 fps,
$ ffmpeg -i raw-video.avi
Guessed Channel Layout for  Input Stream #0.0 : stereo
Input #0, avi, from 'raw-video.avi':
  Duration: 00:15:47.85, start: 0.000000, bitrate: 167435 kb/s
    Stream #0:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 48000 Hz, 2 channels, s16, 1536 kb/s
    Stream #0:1: Video: rawvideo (UYVY / 0x59565955), uyvy422, 720x480, 165669 kb/s, 29.96 fps, 29.96 tbr, 29.96 tbn, 29.96 tbc
At least one output file must be specified

Explicitly describing the channel layout of the audio stream removes the warning line Guessed Channel Layout for Input Stream #0.0 : stereo. We already know that the audio source is a stereo signal (see description of the A/D converter device above)

$ ffmpeg -ac 2 -channel_layout stereo -i raw-video.avi
Input #0, avi, from 'raw-video.avi':
  Duration: 00:15:47.85, start: 0.000000, bitrate: 167435 kb/s
    Stream #0:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 48000 Hz, stereo, s16, 1536 kb/s
    Stream #0:1: Video: rawvideo (UYVY / 0x59565955), uyvy422, 720x480, 165669 kb/s, 29.96 fps, 29.96 tbr, 29.96 tbn, 29.96 tbc
At least one output file must be specified

Both audio channels contain sound.

Compression

Compress video and pack in a Matroska container with uncompressed/raw audio. Use H.264 because of its good quality, size and compatibility.

$ ffmpeg -ac 2 -channel_layout stereo -i raw-video.avi -c:v libx264 -preset slow -crf 17 -pix_fmt yuv422p -c:a pcm_s16le video.mkv
Input #0, avi, from 'raw-video.avi':
  Duration: 00:15:47.85, start: 0.000000, bitrate: 167435 kb/s
    Stream #0:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 48000 Hz, stereo, s16, 1536 kb/s
    Stream #0:1: Video: rawvideo (UYVY / 0x59565955), uyvy422, 720x480, 165669 kb/s, 29.96 fps, 29.96 tbr, 29.96 tbn, 29.96 tbc
[libx264 @ 0x19f9f60] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.1 Cache64
[libx264 @ 0x19f9f60] profile High 4:2:2, level 3.0, 4:2:2 8-bit
[libx264 @ 0x19f9f60] 264 - core 148 r2643 5c65704 - H.264/MPEG-4 AVC codec - Copyleft 2003-2015 - http://www.videolan.org/x264.html - options: cabac=1 ref=5 deblock=1:0:0 analyse=0x3:0x113 me=umh subme=8 psy=1 psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=3 lookahead_threads=1 sliced_threads=0 nr=0 decimate=1 interlaced=0 bluray_compat=0 constrained_intra=0 bframes=3 b_pyramid=2 b_adapt=2 b_bias=0 direct=3 weightb=1 open_gop=0 weightp=2 keyint=250 keyint_min=25 scenecut=40 intra_refresh=0 rc_lookahead=50 rc=crf mbtree=1 crf=17.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
Output #0, matroska, to 'video.mkv':
  Metadata:
    encoder         : Lavf56.40.101
    Stream #0:0: Video: h264 (libx264) (H264 / 0x34363248), yuv422p, 720x480, q=-1--1, 29.96 fps, 1k tbn, 29.96 tbc
    Metadata:
      encoder         : Lavc56.60.100 libx264
    Stream #0:1: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 48000 Hz, stereo, s16, 1536 kb/s
    Metadata:
      encoder         : Lavc56.60.100 pcm_s16le
Stream mapping:
  Stream #0:1 -> #0:0 (rawvideo (native) -> h264 (libx264))
  Stream #0:0 -> #0:1 (pcm_s16le (native) -> pcm_s16le (native))
Press [q] to stop, [?] for help
frame=28397 fps=7.8 q=-1.0 Lsize= 1454961kB time=00:15:47.89 bitrate=12574.3kbits/s
video:1276353kB audio:177729kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.060433%
[libx264 @ 0x19f9f60] frame I:118   Avg QP:17.61  size:109566
[libx264 @ 0x19f9f60] frame P:7342  Avg QP:20.55  size: 69551
[libx264 @ 0x19f9f60] frame B:20937 Avg QP:23.17  size: 37418
[libx264 @ 0x19f9f60] consecutive B-frames:  0.6%  0.7%  7.7% 91.0%
[libx264 @ 0x19f9f60] mb I  I16..4:  4.2% 74.8% 21.0%
[libx264 @ 0x19f9f60] mb P  I16..4:  2.2% 31.7%  5.1%  P16..4: 18.3% 26.7% 15.8%  0.0%  0.0%    skip: 0.1%
[libx264 @ 0x19f9f60] mb B  I16..4:  0.5%  5.4%  0.7%  B16..8: 36.0% 22.6%  7.5%  direct:26.5%  skip: 0.8%  L0:36.1% L1:26.6% BI:37.2%
[libx264 @ 0x19f9f60] 8x8 transform intra:81.0% inter:63.3%
[libx264 @ 0x19f9f60] direct mvs  spatial:100.0% temporal:0.0%
[libx264 @ 0x19f9f60] coded y,uvDC,uvAC intra: 89.9% 99.9% 99.7% inter: 76.6% 99.0% 81.4%
[libx264 @ 0x19f9f60] i16 v,h,dc,p: 48% 28% 13% 11%
[libx264 @ 0x19f9f60] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 12% 14%  9%  6% 11% 11% 13%  9% 14%
[libx264 @ 0x19f9f60] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu:  9% 24%  8%  6%  9%  8% 12%  8% 16%
[libx264 @ 0x19f9f60] i8c dc,h,v,p: 51% 32%  3% 14%
[libx264 @ 0x19f9f60] Weighted P-Frames: Y:14.0% UV:8.8%
[libx264 @ 0x19f9f60] ref P L0: 32.3% 11.2% 25.6% 14.8% 13.4%  2.4%  0.1%
[libx264 @ 0x19f9f60] ref B L0: 69.3% 21.0%  7.1%  2.5%
[libx264 @ 0x19f9f60] ref B L1: 92.3%  7.7%
[libx264 @ 0x19f9f60] kb/s:11031.20

If the command didn’t include the options ac, channel_layout and/or pix_fmt, ffmpeg would have guessed them while printing warnings. It’s best to be explicit about it since we already know the arguments to such options (we know that the pixel format is yuv422p from the output of ffmpeg -i video.mkv)

The result:

$ file video.mkv
video.mkv: Matroska data
$ ffmpeg -ac 2 -channel_layout stereo -i video.mkv
Input #0, matroska,webm, from 'video.mkv':
  Metadata:
    ENCODER         : Lavf56.40.101
  Duration: 00:15:47.89, start: 0.000000, bitrate: 12574 kb/s
    Stream #0:0: Video: h264 (High 4:2:2), yuv422p, 720x480, SAR 1:1 DAR 3:2, 29.96 fps, 29.96 tbr, 1k tbn, 59.92 tbc (default)
    Metadata:
      ENCODER         : Lavc56.60.100 libx264
      DURATION        : 00:15:47.846000000
    Stream #0:1: Audio: pcm_s16le, 48000 Hz, stereo, s16, 1536 kb/s (default)
    Metadata:
      ENCODER         : Lavc56.60.100 pcm_s16le
      DURATION        : 00:15:47.890000000
At least one output file must be specified

Script

Compress raw video into Matroska container, see ffmpeg & Digitizing Hi8 cassettes.

Audio

To extract the raw audio from the Matroska container without encoding

$ ffmpeg -i video.mkv -f s16le -acodec pcm_s16le video_audio.raw

Note that the arguments to f and acodec options match the description of the audio stream (#0.0) in the output of ffmpeg -i video.mkv. The extracted audio is literally raw data:

$ file video_audio.raw
video_audio.raw: data

Neither ffmpeg nor vlc know how to deal with such file without the user providing some help. ffmpeg thinks it’s video data

$ ffmpeg -i video_audio.raw
[image2 @ 0x9c1fe60] Format image2 detected only with low score of 5, misdetection possible!
[rawvideo @ 0x9c20e40] Invalid pixel format.
[image2 @ 0x9c1fe60] Failed to open codec in av_find_stream_info
[rawvideo @ 0x9c20e40] Invalid pixel format.
[image2 @ 0x9c1fe60] Could not find codec parameters for stream 0 (Video: rawvideo, none): unspecified size
Consider increasing the value for the 'analyzeduration' and 'probesize' options
video_audio.raw: could not find codec parameters
Input #0, image2, from 'video_audio.raw':
  Duration: 00:00:00.04, start: 0.000000, bitrate: N/A
    Stream #0:0: Video: rawvideo, none, 25 tbr, 25 tbn, 25 tbc
At least one output file must be specified

vlc will just show you an error. Providing a little guidance works, for example below options will result in vlc correctly playing the file

$ vlc --demux=rawaud --rawaud-samplerate 48000 --rawaud-channels 2 video_audio.raw &

Specify the input format to ffmpeg like this

$ ffmpeg -f s16le -ar 48k -ac 2 -i video_audio.raw
[s16le @ 0x9bd7e80] Estimating duration from bitrate, this may be inaccurate
Guessed Channel Layout for  Input Stream #0.0 : stereo
Input #0, s16le, from 'video_audio.raw':
  Duration: 00:15:47.89, bitrate: 1536 kb/s
    Stream #0:0: Audio: pcm_s16le, 48000 Hz, 2 channels, s16, 1536 kb/s
At least one output file must be specified

Note that you have to be pretty explicit about the nature of the audio data to both vlc and ffmpeg and that the description of the stream matches the output of ffmpeg -i video.mkv in the previous section.

References