[FFmpeg-devel] Scale filter bug

[Alexander Bokovikov]

Hi, All,

The scale filer being applied to produce M4V or MP4 file format  
inserts wrong pixel aspect ratio box, and incorrectly sets the frame  
size in the track header box. As a result target aspect ratio is the  
same as source file aspect ratio, though I've defined target width and  
height explicitly. The same effect appears if I use -s WxH parameter  
instead of -vf scale=W:H one.

As I believe the problem is here:

[SAR 25:36 DAR 5:3]

instead of

[SAR 1:1 DAR 12:5]

Changing the values by manual boxes editing I get correct movie. I.e.  
the problem isn't in a wrong stream encoding, but just in wrong MP4  
boxes writing. Taking into account that 12:5 = (36x5)/(25x3) (see  
digits above), I believe the bug is in these calculations - wrong  
digits are at wrong places. Something like that.

Here is the source FLV file:
http://home.bokovikov.com/ffmpeg/video5.flv

Here is converted M4V file:
http://home.bokovikov.com/ffmpeg/video5.m4v

Here is the command line:

./ffmpeg -i video5.flv -f ipod -vcodec libx264 -vprofile baseline - 
level 30 -maxrate 10000000 -bufsize 10000000 -vf scale=720:300 -acodec  
copy -threads 0 -y video5.m4v

And here is the output:

ffmpeg version N-35462-g61b1d85, Copyright (c) 2000-2011 the FFmpeg  
developers
  built on Dec  7 2011 17:25:26 with gcc 4.0.1 (Apple Inc. build 5493)
  configuration: --enable-static --enable-memalign-hack --enable-gpl -- 
enable-libx264 --disable-doc --disable-ffserver --disable-ffplay -- 
disable-ffprobe --enable-pthreads --enable-runtime-cpudetect --disable- 
network --disable-shared --disable-filters --disable-decoders -- 
disable-encoders --disable-muxers --disable-demuxers --disable-parsers  
--disable-bsfs --disable-protocols --disable-devices --enable- 
filter=crop --enable-filter=scale --enable-protocol=file --enable- 
decoder=aac --enable-decoder=h263 --enable-decoder=flv --enable- 
decoder=vp6f --enable-decoder=vp6a --enable-decoder=h264 --enable- 
decoder=nellymoser --enable-decoder=mp3 --enable-encoder=libx264 -- 
enable-encoder=flv --enable-encoder=aac --enable-parser=h264 --enable- 
parser=mpegaudio --enable-demuxer=flv --enable-muxer=ipod --enable- 
muxer=mp4 --enable-demuxer=mp4 --enable-encoder=mjpeg --enable- 
demuxer=mov --enable-muxer=mjpeg
  libavutil    51. 30. 0 / 51. 30. 0
  libavcodec   53. 40. 0 / 53. 40. 0
  libavformat  53. 24. 0 / 53. 24. 0
  libavdevice  53.  4. 0 / 53.  4. 0
  libavfilter   2. 51. 0 /  2. 51. 0
  libswscale    2.  1. 0 /  2.  1. 0
  libpostproc  51.  2. 0 / 51.  2. 0
Input #0, flv, from 'video5.flv':
  Metadata:
    totalduration   : 60
    starttime       : 0
    lastkeyframetimestamp: 59
    lasttimestamp   : 60
    canseekontime   : true
    canseektoend    : true
    hasvideo        : true
    hasaudio        : true
    datasize        : 630
    bytelength      : 7077753
    author          :
    copyright       :
    title           :
    rating          :
    metadatacreator : FLV Spider for Mac OS X - www.flvspider.com
  Duration: 00:01:00.36, start: 0.019000, bitrate: 938 kb/s
    Stream #0:0: Video: h264 (Baseline), yuv420p, 720x432 [SAR 1:1 DAR  
5:3], 819 kb/s, 25 tbr, 1k tbn, 50 tbc
    Stream #0:1: Audio: aac, 48000 Hz, stereo, s16, 131 kb/s
[buffer @ 0x6014e0] w:720 h:432 pixfmt:yuv420p tb:1/1000000 sar:1/1  
sws_param:
[scale @ 0x60c7a0] w:720 h:432 fmt:yuv420p -> w:720 h:300 fmt:yuv420p  
flags:0x4
[libx264 @ 0x81b620] using SAR=25/36
[libx264 @ 0x81b620] using cpu capabilities: MMX2 SSE2Fast SSSE3 Cache64
[libx264 @ 0x81b620] profile Constrained Baseline, level 3.0
[libx264 @ 0x81b620] 264 - core 119 - H.264/MPEG-4 AVC codec -  
Copyleft 2003-2011 - http://www.videolan.org/x264.html - options:  
cabac=0 ref=3 deblock=1:0:0 analyse=0x1:0x111 me=hex subme=7 psy=1  
psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1  
8x8dct=0 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2  
threads=3 sliced_threads=0 nr=0 decimate=1 interlaced=0  
bluray_compat=0 constrained_intra=0 bframes=0 weightp=0 keyint=250  
keyint_min=25 scenecut=40 intra_refresh=0 rc_lookahead=40 rc=crf  
mbtree=1 crf=23.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4  
vbv_maxrate=10000 vbv_bufsize=10000 crf_max=0.0 nal_hrd=none  
ip_ratio=1.40 aq=1:1.00
Output #0, ipod, to 'video5.m4v':
  Metadata:
    totalduration   : 60
    starttime       : 0
    lastkeyframetimestamp: 59
    lasttimestamp   : 60
    canseekontime   : true
    canseektoend    : true
    hasvideo        : true
    hasaudio        : true
    datasize        : 630
    bytelength      : 7077753
    author          :
    copyright       :
    title           :
    rating          :
    metadatacreator : FLV Spider for Mac OS X - www.flvspider.com
    encoder         : Lavf53.24.0
    Stream #0:0: Video: h264 (avc1 / 0x31637661), yuv420p, 720x300  
[SAR 25:36 DAR 5:3], q=-1--1, 25 tbn, 25 tbc
    Stream #0:1: Audio: aac (mp4a / 0x6134706D), 48000 Hz, stereo, 131  
kb/s
Stream mapping:
  Stream #0:0 -> #0:0 (h264 -> libx264)
  Stream #0:1 -> #0:1 (copy)
Press [q] to stop, [?] for help
frame= 1509 fps= 48 q=-1.0 Lsize=    4766kB time=00:01:00.32 bitrate=  
647.3kbits/s
video:3775kB audio:942kB global headers:0kB muxing overhead 1.042347%
[libx264 @ 0x81b620] frame I:8     Avg QP:18.41  size:  8940
[libx264 @ 0x81b620] frame P:1501  Avg QP:21.42  size:  2528
[libx264 @ 0x81b620] mb I  I16..4: 56.0%  0.0% 44.0%
[libx264 @ 0x81b620] mb P  I16..4: 10.8%  0.0%  0.8%  P16..4: 46.5%  
10.2%  2.9%  0.0%  0.0%    skip:28.7%
[libx264 @ 0x81b620] coded y,uvDC,uvAC intra: 6.8% 52.7% 8.9% inter:  
8.4% 31.8% 0.3%
[libx264 @ 0x81b620] i16 v,h,dc,p: 56% 24% 13%  7%
[libx264 @ 0x81b620] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 30% 15% 27%  6%   
4%  5%  4%  7%  3%
[libx264 @ 0x81b620] i8c dc,h,v,p: 59% 19% 20%  2%
[libx264 @ 0x81b620] ref P L0: 77.4% 13.9%  8.7%
[libx264 @ 0x81b620] kb/s:512.31