High Bitrate I-frame CBR MPEG encoding causes error

[Michael Orton]

m2v encoding of I-frame CBR mpeg at 100Mbit/sec and above gives bitrate tolerance too small for bitrate error. Below 100Mbit, error does not arise.

reproduced with XDcam demuxed m2v file:

ffmpeg.exe -i c:\tmpdir\work.m2v -vcodec mpeg2video -pix_fmt yuv422p -intra -flags +ildct+ilme -b:v 150M -minrate 150M -maxrate 150M -bufsize 47185920 -r 25 -an -y c:\tmpdir\xdcam.m2v
ffmpeg version 2.0.1 Copyright (c) 2000-2013 the FFmpeg developers

built on Sep 26 2013 01:50:16 with gcc 4.8.1 (GCC)
configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-avisynth --enable-bzlib --enable-fontconfig --enable-frei0r --enable-gnutls --enable-iconv --enable-libass --enable-libbluray --enable-libcaca --enable-libfreetype --enable-libgsm --enable-libilbc --enable-libmodplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libschroedinger --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libx264 --enable-libxavs --enable-libxvid --enable-zlib
libavutil 52. 38.100 / 52. 38.100
libavcodec 55. 18.102 / 55. 18.102
libavformat 55. 12.100 / 55. 12.100
libavdevice 55. 3.100 / 55. 3.100
libavfilter 3. 79.101 / 3. 79.101
libswscale 2. 3.100 / 2. 3.100
libswresample 0. 17.102 / 0. 17.102
libpostproc 52. 3.100 / 52. 3.100

Input #0, mpegvideo, from 'c:\tmpdir\work.m2v':

Duration: N/A, bitrate: N/A

Stream #0:0: Video: mpeg2video (4:2:2), yuv422p, 1920x1080 [SAR 1:1 DAR 16:9], 25 fps, 25 tbr, 1200k tbn, 50 tbc

[mpeg2video @ 02793a60] bitrate tolerance too small for bitrate
Output #0, mpeg2video, to 'c:\tmpdir\xdcam.m2v':

Stream #0:0: Video: mpeg2video, yuv422p, 1920x1080 [SAR 1:1 DAR 16:9], q=2-31, 150000 kb/s, 90k tbn, 25 tbc

Stream mapping:

Stream #0:0 -> #0:0 (mpeg2video -> mpeg2video)

Error while opening encoder for output stream #0:0 - maybe incorrect parameters such as bit_rate, rate, width or height

Have source file if needed.

TIA

[Carl Eugen Hoyos]

Is this problem reproducible when using -f lavfi -i testsrc as input?

Please test current git head and report back.

[Michael Orton]

Yes reproducible:

f86387bffmpeg.exe -f lavfi -i testsrc -vcodec mpeg2video -pix_fmt yuv422p -intra -flags +ildct+ilme -b:v 150M -minrate 150M -maxrate 150M -bufsize 47185920 -r 25 -an -y c:\tmpdir\xdcam.m2v
ffmpeg version N-57605-gf86387b Copyright (c) 2000-2013 the FFmpeg developers

built on Oct 29 2013 18:01:53 with gcc 4.8.2 (GCC)
configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-avisynth --enable-bzlib --enable-fontconfig --enable-frei0r --enable-gnutls --enable-iconv --enable-libass --enable-libbluray --enable-libcaca --enable-libfreetype --enable-libgsm --enable-libilbc --enable-libmodplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libschroedinger --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libx264 --enable-libxavs --enable-libxvid --enable-zlib
libavutil 52. 48.100 / 52. 48.100
libavcodec 55. 39.100 / 55. 39.100
libavformat 55. 19.104 / 55. 19.104
libavdevice 55. 5.100 / 55. 5.100
libavfilter 3. 90.100 / 3. 90.100
libswscale 2. 5.101 / 2. 5.101
libswresample 0. 17.104 / 0. 17.104
libpostproc 52. 3.100 / 52. 3.100

Input #0, lavfi, from 'testsrc':

Duration: N/A, start: 0.000000, bitrate: N/A

Stream #0:0: Video: rawvideo (RGB[24] / 0x18424752), rgb24, 320x240 [SAR 1:1 DAR 4:3], 25 tbr, 25 tbn, 25 tbc

[mpeg2video @ 027ae1a0] bitrate tolerance too small for bitrate
Output #0, mpeg2video, to 'c:\tmpdir\xdcam.m2v':

Stream #0:0: Video: mpeg2video, yuv422p, 320x240 [SAR 1:1 DAR 4:3], q=2-31, 150000 kb/s, 90k tbn, 25 tbc

Stream mapping:

Stream #0:0 -> #0:0 (rawvideo -> mpeg2video)

[Carl Eugen Hoyos]

The documentations states that the default for -bt is 4M (and that a too small value is bad for quality). For some (sufficiently large) bitrate, the default must be too small to even encode. FFmpeg is warning that you have to increase the value. I wonder what could be changed.

$ ffmpeg -f lavfi -i testsrc=s=1920x1080 -vcodec mpeg2video -pix_fmt yuv422p -intra -flags +ildct+ilme -b:v 150M -minrate 150M -maxrate 150M -bufsize 47185920 -r 25 -f null -
ffmpeg version N-57552-g0e85bfe Copyright (c) 2000-2013 the FFmpeg developers
  built on Oct 30 2013 00:35:19 with gcc 4.7 (SUSE Linux)
  configuration: --enable-gpl
  libavutil      52. 48.100 / 52. 48.100
  libavcodec     55. 39.100 / 55. 39.100
  libavformat    55. 19.104 / 55. 19.104
  libavdevice    55.  5.100 / 55.  5.100
  libavfilter     3. 90.100 /  3. 90.100
  libswscale      2.  5.101 /  2.  5.101
  libswresample   0. 17.104 /  0. 17.104
  libpostproc    52.  3.100 / 52.  3.100
Input #0, lavfi, from 'testsrc=s=1920x1080':
  Duration: N/A, start: 0.000000, bitrate: N/A
    Stream #0:0: Video: rawvideo (RGB[24] / 0x18424752), rgb24, 1920x1080 [SAR 1:1 DAR 16:9], 25 tbr, 25 tbn, 25 tbc
[mpeg2video @ 0x1bbabe0] bitrate tolerance too small for bitrate
Output #0, null, to 'pipe:':
    Stream #0:0: Video: mpeg2video, yuv422p, 1920x1080 [SAR 1:1 DAR 16:9], q=2-31, 150000 kb/s, 90k tbn, 25 tbc
Stream mapping:
  Stream #0:0 -> #0:0 (rawvideo -> mpeg2video)
Error while opening encoder for output stream #0:0 - maybe incorrect parameters such as bit_rate, rate, width or height

[Michael Orton]

The difficulty with adding a -bt flag (10M seems to work) is that the stream is then marked as VBR instead of CBR, and the encoded frames are no longer the expected constant size. The actual encoded size I worked out to be max about 100Mbit/sec. The aim is to achieve a CBR I frame encoding similar to IMX.

I guess if the -bt default were to scale with bitrate, that would make sense, as then the resulting bitstream would be constant frame size and marked as CBR(?)

[Carl Eugen Hoyos]

Replying to ortoni:

The difficulty with adding a -bt flag (10M seems to work) is that the stream is then marked as VBR instead of CBR,

How can I reproduce this?

and the encoded frames are no longer the expected constant size.

While CBR != constant frame size, I see constant frame size for -b:v 150M -minrate 150M -bt 6M, please elaborate.

[Michael Orton]

I could not get -b:v 150M -minrate 150M -bt 6M to work - still had "bitrate tolerance too small.." [maxrate and bufsize removed]

Increasing -bt: -b:v 150M -minrate 150M -bt 8M was allowed but m2v sequence header says BR = 104,857,200, with sequence length (i.e. # bits/frame) variable, vbv bufsize was 15,138,816. Mediainfo calculated the bitrate as 45.6Mbps.

Adding -maxrate and bufsize: -b:v 150M -minrate 150M -maxrate 150M -bt 8M -bufsize 47185920, m2v seq HDR now says BR=45,142,400, sequence length is constant, vbv bufsize is 13,631,488 and Mediainfo says "overall bitrate 149Mb/s" but Maximum Bitrate is 46.8Mbps

All tests on same git as original; and I'm very confused!

[Carl Eugen Hoyos]

Replying to ortoni:

I could not get -b:v 150M -minrate 150M -bt 6M to work - still had "bitrate tolerance too small.." [maxrate and bufsize removed]

Command line and complete, uncut console output missing.

[Michael Orton]

my bad:

f86387bffmpeg.exe -i c:\tmpdir\work.m2v -vcodec mpeg2video -pix_fmt yuv422p -intra -flags +ildct+ilme -b:v 150M -minrate 150M -bt 6M -r 25 -an -y c:\tmpdir\xdcam.m2v
ffmpeg version N-57605-gf86387b Copyright (c) 2000-2013 the FFmpeg developers

built on Oct 29 2013 18:01:53 with gcc 4.8.2 (GCC)
configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-avisynth --enable-bzlib --enable-fontconfig --enable-frei0r --enable-gnutls --enable-iconv --enable-libass --enable-libbluray --enable-libcaca --enable-libfreetype --enable-libgsm --enable-libilbc --enable-libmodplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libschroedinger --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libx264 --enable-libxavs --enable-libxvid --enable-zlib
libavutil 52. 48.100 / 52. 48.100
libavcodec 55. 39.100 / 55. 39.100
libavformat 55. 19.104 / 55. 19.104
libavdevice 55. 5.100 / 55. 5.100
libavfilter 3. 90.100 / 3. 90.100
libswscale 2. 5.101 / 2. 5.101
libswresample 0. 17.104 / 0. 17.104
libpostproc 52. 3.100 / 52. 3.100

Input #0, mpegvideo, from 'c:\tmpdir\work.m2v':

Duration: N/A, bitrate: N/A

Stream #0:0: Video: mpeg2video (4:2:2), yuv422p(tv, bt709), 1920x1080 [SAR 1:1 DAR 16:9], max. 50000 kb/s, 25 fps, 25 tbr, 1200k tbn, 50 tbc

[mpeg2video @ 02528b40] Warning min_rate > 0 but min_rate != max_rate isn't recommended!
[mpeg2video @ 02528b40] bitrate tolerance too small for bitrate
Output #0, mpeg2video, to 'c:\tmpdir\xdcam.m2v':

Stream #0:0: Video: mpeg2video, yuv422p, 1920x1080 [SAR 1:1 DAR 16:9], q=2-31, 150000 kb/s,

90k tbn, 25 tbc
Stream mapping:

Stream #0:0 -> #0:0 (mpeg2video -> mpeg2video)

Error while opening encoder for output stream #0:0 - maybe incorrect parameters such as bit_rate, rate, width or height

[Michael Orton]

Is there any other information I can investigate and report for this ticket @cehoyos?

Thanks and regards

[Carl Eugen Hoyos]

Why did you remove the bufsize (and the maxrate) parameter?

[Michael Orton]

Sorry was following your exact CLI above - thought that was odd. Correct output below, seems unchanged :

f86387bffmpeg.exe" -i c:\tmpdir\work.m2v -vcodec mpeg2video -pix_fmt yuv422p -intra -flags +ildct+ilme -b:v 150M -maxrate 150M -minrate 150M -bt 6M -bufsize 47185920 -r 25 -an -y c:\tmpdir\xdcam.m2v
ffmpeg version N-57605-gf86387b Copyright (c) 2000-2013 the FFmpeg developers

built on Oct 29 2013 18:01:53 with gcc 4.8.2 (GCC)
configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-avisynth --enable-bzlib --enable-fontconfig --enable-frei0r --enable-gnutls --enable-iconv --enable-libass --enable-libbluray --enable-libcaca --enable-libfreetype --enable-libgsm --enable-libilbc --enable-libmodplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libschroedinger --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libx264 --enable-libxavs --enable-libxvid --enable-zlib
libavutil 52. 48.100 / 52. 48.100
libavcodec 55. 39.100 / 55. 39.100
libavformat 55. 19.104 / 55. 19.104
libavdevice 55. 5.100 / 55. 5.100
libavfilter 3. 90.100 / 3. 90.100
libswscale 2. 5.101 / 2. 5.101
libswresample 0. 17.104 / 0. 17.104
libpostproc 52. 3.100 / 52. 3.100

Input #0, mpegvideo, from 'c:\tmpdir\work.m2v':

Duration: N/A, bitrate: N/A

Stream #0:0: Video: mpeg2video (4:2:2), yuv422p(tv, bt709), 1920x1080 [SAR 1:1 DAR 16:9], max. 50000 kb/s, 25 fps, 25 tbr, 1200k tbn, 50 tbc

[mpeg2video @ 02618b20] bitrate tolerance too small for bitrate
Output #0, mpeg2video, to 'c:\tmpdir\xdcam.m2v':

Stream #0:0: Video: mpeg2video, yuv422p, 1920x1080 [SAR 1:1 DAR 16:9], q=2-31, 150000 kb/s,

90k tbn, 25 tbc
Stream mapping:

Stream #0:0 -> #0:0 (mpeg2video -> mpeg2video)

Error while opening encoder for output stream #0:0 - maybe incorrect parameters such as bit_rate, rate, width or height

Using -bt 8M proceeds:

f86387bffmpeg.exe" -i c:\tmpdir\work.m2v -vcodec mpeg2video -pix_fmt yuv422p -intra -flags +ildct+ilme -b:v 150M -maxrate 150M -minrate 150M -bt 8M -bufsize 47185920 -r 25 -an -y c:\tmpdir\xdcam.m2v
ffmpeg version N-57605-gf86387b Copyright (c) 2000-2013 the FFmpeg developers

built on Oct 29 2013 18:01:53 with gcc 4.8.2 (GCC)
configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-avisynth --enable-bzlib --enable-fontconfig --enable-frei0r --enable-gnutls --enable-iconv --enable-libass --enable-libbluray --enable-libcaca --enable-libfreetype --enable-libgsm --enable-libilbc --enable-libmodplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libschroedinger --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libx264 --enable-libxavs --enable-libxvid --enable-zlib
libavutil 52. 48.100 / 52. 48.100
libavcodec 55. 39.100 / 55. 39.100
libavformat 55. 19.104 / 55. 19.104
libavdevice 55. 5.100 / 55. 5.100
libavfilter 3. 90.100 / 3. 90.100
libswscale 2. 5.101 / 2. 5.101
libswresample 0. 17.104 / 0. 17.104
libpostproc 52. 3.100 / 52. 3.100

Input #0, mpegvideo, from 'c:\tmpdir\work.m2v':

Duration: N/A, bitrate: N/A

Stream #0:0: Video: mpeg2video (4:2:2), yuv422p(tv, bt709), 1920x1080 [SAR 1:1 DAR 16:9], max. 50000 kb/s, 25 fps, 25 tbr, 1200k tbn, 50 tbc

Output #0, mpeg2video, to 'c:\tmpdir\xdcam.m2v':

Metadata:

encoder : Lavf55.19.104
Stream #0:0: Video: mpeg2video, yuv422p, 1920x1080 [SAR 1:1 DAR 16:9], q=2-31, 150000 kb/s,90k tbn, 25 tbc

Stream mapping:

Stream #0:0 -> #0:0 (mpeg2video -> mpeg2video)

Press [q] to stop, ? for help
frame= 500 fps= 58 q=1.6 Lsize= 364771kB time=00:00:19.96 bitrate=149709.6kbits/s
video:364771kB audio:0kB subtitle:0 global headers:0kB muxing overhead 0.000000%

but this encoding above shows in MediaInfo:
Bit rate : 149 Mbps
Maximum bit rate : 46.8 Mbps

and my m2v essence tool shows at bitrate of 45,142,400 and VBV size= 13,631,488 in the Sequence header

But 99M is happy:

f86387bffmpeg.exe" -i c:\tmpdir\work.m2v -vcodec mpeg2video -pix_fmt yuv422p -intra -flags +ildct+ilme -b:v 99M -maxrate 99M -minrate 99M -bt 6M -bufsize 47185920 -r 25 -an -y c:\tmpdir\xdcam.m2v
ffmpeg version N-57605-gf86387b Copyright (c) 2000-2013 the FFmpeg developers

built on Oct 29 2013 18:01:53 with gcc 4.8.2 (GCC)
configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-avisynth --enable-bzlib --enable-fontconfig --enable-frei0r --enable-gnutls --enable-iconv --enable-libass --enable-libbluray --enable-libcaca --enable-libfreetype --enable-libgsm --enable-libilbc --enable-libmodplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libschroedinger --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libx264 --enable-libxavs --enable-libxvid --enable-zlib
libavutil 52. 48.100 / 52. 48.100
libavcodec 55. 39.100 / 55. 39.100
libavformat 55. 19.104 / 55. 19.104
libavdevice 55. 5.100 / 55. 5.100
libavfilter 3. 90.100 / 3. 90.100
libswscale 2. 5.101 / 2. 5.101
libswresample 0. 17.104 / 0. 17.104
libpostproc 52. 3.100 / 52. 3.100

Input #0, mpegvideo, from 'c:\tmpdir\work.m2v':

Duration: N/A, bitrate: N/A

Stream #0:0: Video: mpeg2video (4:2:2), yuv422p(tv, bt709), 1920x1080 [SAR 1:1 DAR 16:9], max. 50000 kb/s, 25 fps, 25 tbr, 1200k tbn, 50 tbc

Output #0, mpeg2video, to 'c:\tmpdir\xdcam.m2v':

Metadata:

encoder : Lavf55.19.104
Stream #0:0: Video: mpeg2video, yuv422p, 1920x1080 [SAR 1:1 DAR 16:9], q=2-31, 99000 kb/s, 90k tbn, 25 tbc

And for this one MediaInfo reports:
Overall bit rate : 98.4 Mbps
Bit rate : 99.0 Mbps

m2v sequence header bitrate = 99,000,000 and vbv = 13,631,488

Its just anything over 100Mbps seems not to encode correctly

[Michael Niedermayer]

CBR is not the same as constant frame sizes.
Which do you want ? and why ?

and i wouldnt trust mediainfo when it comes to CBR bitrates

[Carl Eugen Hoyos]

Please reopen this ticket if you can answer Michael's questions.