2009/05/01

今更ですが動き検索の記事を

Diary 2005-1
まるも製作所さんのところに書いてありました。
ふむふむ・・・

とかいいつつ、先程のソースをエンコード。
んで、僕は行列が非常に苦手な制御の学生だったわけですが、量子化行列をjvtからflatにするとcrfで同値を指定したときに歴然とした差が出ることが・・・

cqm=jvt:
x264.exe --crf 22 --qpmin 12 --qpmax 32 --qpstep 8 --ipratio 1.4 --pbratio 1.6 --bframes 3 --b-adapt 2 --w
eightb --ref 4 --mixed-refs --min-keyint 1 --scenecut 60 --direct auto --8x8dct --me umh --subme 7 --no-fast-pskip --no-
dct-decimate --cqm jvt --sar 1:1 --level 4.1 --progress --threads auto --thread-input -o i:\temp.mp4 i:\temp.avs
avis [info]: 1280x720 @ 23.98 fps (2136 frames)
x264 [info]: using SAR=1/1
x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 FastShuffle SSE4.1 Cache64
x264 [info]: profile High, level 4.1
mp4 [info]: initial delay 1001 (scale 24000)
x264 [info]: slice I:54 Avg QP:18.71 size: 52576 PSNR Mean Y:46.67 U:49.80 V:50.70 Avg:47.49 Global:46.59
x264 [info]: slice P:987 Avg QP:21.09 size: 25305 PSNR Mean Y:43.80 U:47.90 V:48.37 Avg:44.79 Global:44.08
x264 [info]: slice B:1095 Avg QP:26.70 size: 3512 PSNR Mean Y:42.59 U:46.68 V:47.38 Avg:43.60 Global:42.74
x264 [info]: consecutive B-frames: 13.0% 37.2% 40.6% 9.2%
x264 [info]: mb I I16..4: 10.6% 80.0% 9.4%
x264 [info]: mb P I16..4: 2.2% 13.2% 1.5% P16..4: 55.1% 10.5% 10.4% 0.0% 0.0% skip: 7.0%
x264 [info]: mb B I16..4: 0.1% 0.4% 0.1% B16..8: 27.0% 0.8% 1.0% direct: 1.8% skip:68.8% L0:37.8% L1:60.1% BI:
2.1%
x264 [info]: 8x8 transform intra:78.1% inter:69.5%
x264 [info]: direct mvs spatial:99.1% temporal:0.9%
x264 [info]: ref P L0 73.9% 13.0% 8.5% 4.6%
x264 [info]: ref B L0 89.9% 6.4% 3.7%
x264 [info]: SSIM Mean Y:0.9869737
x264 [info]: PSNR Mean Y:43.252 U:47.321 V:47.924 Avg:44.249 Global:43.381 kb/s:2843.12

encoded 2136 frames, 17.07 fps, 2843.37 kb/s

cqm=flat:
x264.exe --crf 22 --qpmin 12 --qpmax 32 --qpstep 8 --ipratio 1.4 --pbratio 1.6 --bframes 3 --b-adapt 2 --w
eightb --ref 4 --mixed-refs --min-keyint 1 --scenecut 60 --direct auto --8x8dct --me umh --subme 7 --no-fast-pskip --no-
dct-decimate --cqm flat --sar 1:1 --level 4.1 --progress --threads auto --thread-input -o i:\temp.mp4 i:\temp.avs
avis [info]: 1280x720 @ 23.98 fps (2136 frames)
x264 [info]: using SAR=1/1
x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 FastShuffle SSE4.1 Cache64
x264 [info]: profile High, level 4.1
mp4 [info]: initial delay 1001 (scale 24000)
x264 [info]: slice I:54 Avg QP:18.72 size: 45028 PSNR Mean Y:46.83 U:49.36 V:50.15 Avg:47.53 Global:46.62
x264 [info]: slice P:987 Avg QP:21.06 size: 20642 PSNR Mean Y:43.78 U:47.15 V:47.58 Avg:44.65 Global:43.89
x264 [info]: slice B:1095 Avg QP:26.71 size: 3054 PSNR Mean Y:42.50 U:46.09 V:46.74 Avg:43.43 Global:42.51
x264 [info]: consecutive B-frames: 13.0% 37.2% 40.6% 9.2%
x264 [info]: mb I I16..4: 18.5% 61.6% 20.0%
x264 [info]: mb P I16..4: 5.0% 12.2% 3.4% P16..4: 45.3% 10.1% 9.7% 0.0% 0.0% skip:14.3%
x264 [info]: mb B I16..4: 0.3% 0.5% 0.1% B16..8: 29.8% 0.8% 1.0% direct: 2.1% skip:65.3% L0:37.6% L1:60.5% BI:
1.8%
x264 [info]: 8x8 transform intra:59.4% inter:57.5%
x264 [info]: direct mvs spatial:99.6% temporal:0.4%
x264 [info]: ref P L0 72.9% 13.7% 8.9% 4.6%
x264 [info]: ref B L0 90.4% 6.1% 3.5%
x264 [info]: SSIM Mean Y:0.9864717
x264 [info]: PSNR Mean Y:43.198 U:46.663 V:47.214 Avg:44.096 Global:43.169 kb/s:2348.09

encoded 2136 frames, 17.48 fps, 2348.33 kb/s

PSNRやSSIMはほとんど変わらないのですが、何処に違いが・・・
これが変換行列は大事言われる所以ですかね。
・・・もう少し勉強します・・・。

0 件のコメント:

コメントを投稿