Eu quero usar o ffmpeg mais recente para codificar o conjunto de quadros JPG para o filme h.264 de 1920 x 1080 a 60 fps.
Estou à procura de um conjunto de parâmetros que permita a reprodução suave de vídeo codificado em máquinas PC razoavelmente strongs como [email protected]+Intel HD Graphics.
Eu tentei vários perfis (principal, alto) e níveis (2.1, 3.1, 4) e bitrates (6mbps, 8mbps, 10mbps) por exemplo:
ffmpeg -framerate 60 -start_number 1 -i frame-%05d.jpg -vcodec libx264 -preset slow -b:v 8M -profile:v high -level:v 3.1 -y output.mp4
Mas quando eu reproduzo output.mp4 no último VLC ou WMP eu tenho framedrops (não muitos, mas perceptíveis).
Eu reproduzi o material aleatório do youtube 60fps / 1080 na mesma máquina e a reprodução foi suave, então presumo que a máquina seja capaz de reproduzir suavemente, mas o vídeo tem que ser codificado com parâmetros apropriados.
Quais parâmetros poderiam ser usados para tornar a decodificação menos exigente no processamento?
Eu gostaria de manter a qualidade full hd e 60fps, mas o tamanho do arquivo não é relevante.
Saída do comando ffmpeg:
ffmpeg -framerate 60 -start_numb
er 1 -i d:\tmp\demo-60\test-%05d.jpg -vcodec libx264 -preset slow -b:v 8M -profi
le:v high -level:v 3.1 -y d:\tmp\insys-demo-60-noaudio.mp4
ffmpeg version N-67742-g3f07dd6 Copyright (c) 2000-2014 the FFmpeg developers
built on Nov 16 2014 22:10:05 with gcc 4.9.2 (GCC)
configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-av
isynth --enable-bzlib --enable-fontconfig --enable-frei0r --enable-gnutls --enab
le-iconv --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --
enable-libfreetype --enable-libgme --enable-libgsm --enable-libilbc --enable-lib
modplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrw
b --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libschroedinge
r --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libtwolame --en
able-libvidstab --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-libvorbis
--enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx264 --enable-
libx265 --enable-libxavs --enable-libxvid --enable-zlib
libavutil 54. 13.100 / 54. 13.100
libavcodec 56. 12.101 / 56. 12.101
libavformat 56. 13.100 / 56. 13.100
libavdevice 56. 3.100 / 56. 3.100
libavfilter 5. 2.103 / 5. 2.103
libswscale 3. 1.101 / 3. 1.101
libswresample 1. 1.100 / 1. 1.100
libpostproc 53. 3.100 / 53. 3.100
Input #0, image2, from 'd:\tmp\demo-60\test-%05d.jpg':
Duration: 00:07:01.67, start: 0.000000, bitrate: N/A
Stream #0:0: Video: mjpeg, yuvj420p(pc, bt470bg/unknown/unknown), 1920x1080
[SAR 96:96 DAR 16:9], 60 fps, 60 tbr, 60 tbn, 60 tbc
No pixel format specified, yuvj420p for H.264 encoding chosen.
Use -pix_fmt yuv420p for compatibility with outdated media players.
[libx264 @ 0000000002c47d60] using SAR=1/1
[libx264 @ 0000000002c47d60] frame MB size (120x68) > level limit (3600)
[libx264 @ 0000000002c47d60] DPB size (4 frames, 32640 mbs) > level limit (2 fra
mes, 18000 mbs)
[libx264 @ 0000000002c47d60] MB rate (489600) > level limit (108000)
[libx264 @ 0000000002c47d60] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2
AVX
[libx264 @ 0000000002c47d60] profile High, level 3.1
[libx264 @ 0000000002c47d60] 264 - core 142 r2479 dd79a61 - H.264/MPEG-4 AVC cod
ec - Copyleft 2003-2014 - http://www.videolan.org/x264.html - options: cabac=1 r
ef=2 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_pski
p=1 chroma_qp_offset=-2 threads=6 lookahead_threads=1 sliced_threads=0 nr=0 deci
mate=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=2
5 scenecut=40 intra_refresh=0 rc_lookahead=50 rc=abr mbtree=1 bitrate=8000 ratet
ol=1.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
Output #0, mp4, to 'd:\tmp\insys-demo-60-noaudio.mp4':
Metadata:
encoder : Lavf56.13.100
Stream #0:0: Video: h264 (libx264) ([33][0][0][0] / 0x0021), yuvj420p(pc), 1
920x1080 [SAR 96:96 DAR 16:9], q=-1--1, 8000 kb/s, 60 fps, 15360 tbn, 60 tbc
Metadata:
encoder : Lavc56.12.101 libx264
Stream mapping:
Stream #0:0 -> #0:0 (mjpeg (native) -> h264 (libx264))
Press [q] to stop, [?] for help
frame= 68 fps= 31 q=33.0 size= 252kB time=00:00:00.10 bitrate=20638.7kbits
frame= 87 fps= 32 q=27.0 size= 273kB time=00:00:00.41 bitrate=5358.4kbits/
frame= 102 fps= 31 q=24.0 size= 422kB time=00:00:00.66 bitrate=5191.4kbits/
frame= 119 fps= 31 q=22.0 size= 601kB time=00:00:00.95 bitrate=5183.7kbits/
frame= 136 fps= 31 q=21.0 size= 814kB time=00:00:01.23 bitrate=5409.1kbits/
frame= 150 fps= 31 q=21.0 size= 1047kB time=00:00:01.46 bitrate=5846.0kbits/
frame= 163 fps= 30 q=22.0 size= 1375kB time=00:00:01.68 bitrate=6690.5kbits/
frame= 177 fps= 30 q=22.0 size= 1598kB time=00:00:01.91 bitrate=6829.8kbits/
frame= 178 fps= 27 q=21.0 size= 1618kB time=00:00:01.93 bitrate=6854.9kbits/
frame= 193 fps= 27 q=21.0 size= 1810kB time=00:00:02.18 bitrate=6789.8kbits/
frame= 206 fps= 27 q=21.0 size= 2070kB time=00:00:02.40 bitrate=7065.7kbits/
frame= 220 fps= 27 q=22.0 size= 2313kB time=00:00:02.63 bitrate=7196.7kbits/
frame= 234 fps= 27 q=22.0 size= 2615kB time=00:00:02.86 bitrate=7474.0kbits/
frame= 247 fps= 27 q=22.0 size= 2793kB time=00:00:03.08 bitrate=7420.5kbits/
frame= 263 fps= 27 q=21.0 size= 2947kB time=00:00:03.35 bitrate=7205.6kbits/
frame= 279 fps= 28 q=21.0 size= 3119kB time=00:00:03.61 bitrate=7064.4kbits/
frame= 297 fps= 28 q=20.0 size= 3257kB time=00:00:03.91 bitrate=6811.5kbits/
frame= 312 fps= 28 q=20.0 size= 4109kB time=00:00:04.16 bitrate=8078.8kbits/
frame= 328 fps= 28 q=22.0 size= 4272kB time=00:00:04.43 bitrate=7893.5kbits/
frame= 345 fps= 28 q=21.0 size= 4391kB time=00:00:04.71 bitrate=7626.7kbits/
frame= 361 fps= 28 q=21.0 size= 4650kB time=00:00:04.98 bitrate=7644.7kbits/
...
No final do processamento, o ffmpeg imprime:
frame=25193 fps= 17 q=33.0 size= 395561kB time=00:06:58.85 bitrate=7736.5kbits/
frame=25211 fps= 17 q=33.0 size= 395617kB time=00:06:59.15 bitrate=7732.1kbits/
frame=25232 fps= 17 q=32.0 size= 395672kB time=00:06:59.50 bitrate=7726.7kbits/
frame=25250 fps= 17 q=32.0 size= 395735kB time=00:06:59.80 bitrate=7722.4kbits/
frame=25267 fps= 17 q=32.0 size= 395776kB time=00:07:00.08 bitrate=7718.0kbits/
frame=25285 fps= 17 q=32.0 size= 395811kB time=00:07:00.38 bitrate=7713.2kbits/
frame=25300 fps= 17 q=-1.0 Lsize= 396340kB time=00:07:01.63 bitrate=7700.6kbits
/s
video:396048kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxin
g overhead: 0.073937%
[libx264 @ 00000000003675c0] frame I:104 Avg QP:26.69 size:182027
[libx264 @ 00000000003675c0] frame P:8999 Avg QP:31.82 size: 28039
[libx264 @ 00000000003675c0] frame B:16197 Avg QP:36.70 size: 8291
[libx264 @ 00000000003675c0] consecutive B-frames: 8.4% 12.1% 19.4% 60.0%
[libx264 @ 00000000003675c0] mb I I16..4: 17.2% 51.0% 31.8%
[libx264 @ 00000000003675c0] mb P I16..4: 2.2% 4.8% 1.5% P16..4: 17.9% 9.6
% 5.9% 0.0% 0.0% skip:58.0%
[libx264 @ 00000000003675c0] mb B I16..4: 0.3% 0.5% 0.1% B16..8: 22.8% 4.8
% 1.4% direct: 0.9% skip:69.2% L0:43.6% L1:50.3% BI: 6.0%
[libx264 @ 00000000003675c0] final ratefactor: 27.73
[libx264 @ 00000000003675c0] 8x8 transform intra:55.0% inter:51.7%
[libx264 @ 00000000003675c0] direct mvs spatial:97.7% temporal:2.3%
[libx264 @ 00000000003675c0] coded y,uvDC,uvAC intra: 27.2% 19.8% 10.7% inter: 5
.8% 3.5% 1.3%
[libx264 @ 00000000003675c0] i16 v,h,dc,p: 44% 45% 5% 6%
[libx264 @ 00000000003675c0] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 23% 10% 45% 3% 4%
4% 4% 3% 4%
[libx264 @ 00000000003675c0] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 20% 17% 18% 6% 8%
7% 8% 6% 10%
[libx264 @ 00000000003675c0] i8c dc,h,v,p: 72% 16% 11% 2%
[libx264 @ 00000000003675c0] Weighted P-Frames: Y:2.0% UV:1.5%
[libx264 @ 00000000003675c0] ref P L0: 71.7% 14.6% 13.6% 0.1%
[libx264 @ 00000000003675c0] ref B L0: 87.9% 12.1%
[libx264 @ 00000000003675c0] ref B L1: 94.2% 5.8%
[libx264 @ 00000000003675c0] kb/s:7694.27
Preferências do Player VLC:
Atualizar
Após testar várias opções de GOP (-g), quadros de referência (-refs), tamanhos de buffer (-bufsize), bframes (-bf), desabilitando o cabac (-coder 0), profiles (main, high, high10, high422, high444), níveis (3.0 - 5.0) e testando vários players ativando / desativando a aceleração de hardware Cheguei à conclusão de que provavelmente não é possível codificar H.264 1920x1080 @ 60fps com dinâmica semelhante a de um jogo para que ele pudesse ser reproduzido em PC típico flawlesly. Eu testei várias máquinas, incluindo o Mac Book Pro. A única máquina que conseguiu reproduzir o material sem problemas foi o mais recente modelo de TV da Samsung (série UHD).