Some advice on settings please

[CommonMortal]

Hello, i was wondering if one of you gurus could enlighten me. I 've done a lot of reading (wiki, this forum and handrake's forum) and i have finally got to understand quite well what the more important settings do.

However, there are some points about which i have doubts, since some people keep using them.

1) No fast pskip and partitions=all. I know that both have little benefit compared to the time they require. So i normally use fast pskip= 1 and partitions= default. Are there situations that enabling one or both of them is actually useful?

2) I backup Blue rays to 720p trying to find a good quality without waiting for 1 day to finish.

After many tries, i ended up with something like this:

profile high --level 4.1 --preset slow --tune film --pass 2 --bitrate 4553 -ref 5 --bframes 8 --b-adapt 2 --b-pyramid normal --direct auto --deblock -3:-3 --subme 9 --qpmin 1 --me umh --merange 24

I change b frames from 5 to 8 actually, depending on the source lenght and motion and merange up to 32.

I have noticed that the statistics at the end of pass 1 and pass 2 are different, with those at the end of pass 2 being "worse" (higher QP). Is that normal? Does this means that at the end, the final rate factor at 1st pass is not valid anymore and in reality it's higher? Because i take the final ratefactor as something equivalent of crf, so i always check it to see if i am doing a good quality compression.

example:

End of pass 1:

x264 [info]: frame I:1300 Avg QP:12.81 size:105972
x264 [info]: frame P:38795 Avg QP:18.04 size: 40005
x264 [info]: frame B:150028 Avg QP:19.89 size: 18430
x264 [info]: consecutive B-frames: 1.3% 2.3% 3.7% 23.0% 8.3% 41.9% 4.6% 13.2% 1.7%
x264 [info]: mb I I16..4: 49.5% 0.0% 50.5%
x264 [info]: mb P I16..4: 38.8% 0.0% 0.0% P16..4: 34.9% 0.0% 0.0% 0.0% 0.0% skip:26.4%
x264 [info]: mb B I16..4: 14.3% 0.0% 0.0% B16..8: 26.2% 0.0% 0.0% direct:19.8% skip:39.7% L0:28.8% L1:33.1% BI:38.1%
x264 [info]: final ratefactor: 16.95
x264 [info]: direct mvs spatial:100.0% temporal:0.0%
x264 [info]: coded y,uvDC,uvAC intra: 81.5% 58.5% 20.0% inter: 24.6% 21.9% 0.9%
x264 [info]: i16 v,h,dc,p: 25% 16% 43% 17%
x264 [info]: i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 16% 17% 24% 4% 10% 6% 7% 7% 9%
x264 [info]: i8c dc,h,v,p: 53% 21% 21% 4%
x264 [info]: Weighted P-Frames: Y:3.9% UV:2.3%
x264 [info]: kb/s:4494.21


End of pass 2:

x264 [info]: frame I:1300 Avg QP:14.00 size:120429
x264 [info]: frame P:38795 Avg QP:19.38 size: 40018
x264 [info]: frame B:150028 Avg QP:20.88 size: 18672
x264 [info]: consecutive B-frames: 1.3% 2.3% 3.7% 23.0% 8.3% 41.9% 4.6% 13.2% 1.7%
x264 [info]: mb I I16..4: 22.4% 66.5% 11.1%
x264 [info]: mb P I16..4: 1.9% 8.7% 1.6% P16..4: 31.8% 19.8% 10.6% 0.0% 0.0% skip:25.6%
x264 [info]: mb B I16..4: 0.1% 1.0% 0.1% B16..8: 37.8% 11.2% 3.2% direct: 4.9% skip:41.7% L0:46.5% L1:45.4% BI: 8.1%
x264 [info]: 8x8 transform intra:73.1% inter:60.6%
x264 [info]: direct mvs spatial:98.4% temporal:1.6%
x264 [info]: coded y,uvDC,uvAC intra: 84.2% 80.1% 55.5% inter: 31.1% 30.3% 2.8%
x264 [info]: i16 v,h,dc,p: 48% 21% 7% 23%
x264 [info]: i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 8% 7% 3% 11% 16% 15% 15% 12% 11%
x264 [info]: i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 7% 10% 4% 10% 15% 14% 15% 12% 13%
x264 [info]: i8c dc,h,v,p: 43% 23% 17% 17%
x264 [info]: Weighted P-Frames: Y:3.9% UV:2.3%
x264 [info]: ref P L0: 44.7% 10.4% 23.1% 11.0% 9.8% 0.9% 0.0%
x264 [info]: ref B L0: 68.6% 21.7% 7.8% 1.9%
x264 [info]: ref B L1: 88.8% 11.2%
x264 [info]: kb/s:4550.44


Thank you.

[shon3i]

Higher quantizer not mean worse quality, that's all you need to know.

[CommonMortal]

Higher quantizer not mean worse quality, that's all you need to know.

So you mean that i shouldn't be bothered by that. Thank you very much, i really appreciate it.

[nurbs]

Fast first pass is done without psy-rd. Psy-rd results in higher bitrate at the same quantizer so that alone means that the quantizer in your second pass is higher. I guess the other settings have some influence as well, but psy is very noticeable.

[CommonMortal]

Fast first pass is done without psy-rd. Psy-rd results in higher bitrate at the same quantizer so that alone means that the quantizer in your second pass is higher. I guess the other settings have some influence as well, but psy is very noticeable.

Ah, this makes perfectly sense! In deed, i use psy-rd. So i shouldn't worry about the final ratefactor being "wrong". Because i use it as "guideline" to see if i am getting a good quality or not. When i see the final ratefactor going above 20, i usually abort and increase the file size and encode again. Most of the time i keep it between 16-19.

Thank you very much.

[AnonCrow]

If increasing filesize is an option for you, why do 2-pass encoding at all when you could do a 1-pass CRF encode ?

[CommonMortal]

If increasing filesize is an option for you, why do 2-pass encoding at all when you could do a 1-pass CRF encode ?

Increasing the file size is an option, but it is the exception. I do that only when i can't do otherwise. Most of the time i try to go for 4.37GB encodes. They are usually enough for a good quality encode, they take reasonable disk space and on need, you can burn some to a few DVDs to take them with you if you go somewhere away from home with the laptop. Besides, i have begun to "predict" if a movie will require more size before starting to encode. I look at the original file size and then quickly review the movie to see if it has much grain or movement or explosions, as well as the total movie length and most of the time i get it right.

[7ekno]

I use --crf 19 --preset slow --tune film and get 720p encodes between 2-4GBs with AC3 5.1 audio ... much faster than doing 2 pass :P

7ek