『漫游』酷论坛>『影音数码技术学习交流』>[聊天]x264参数更新
[聊天]x264参数更新
superkidx@2009-07-28 16:47
--subme / -m 9 ---》10
Add QPRD support as subme=10
Refactor trellis lambda selection to be done in analyse_init instead of in trellis.
This will allow for more easy adaption of lambda later on; for now it allows constant lambda across variable QPs.
QPRD is only available with adaptive quantization enabled and generally improves SSIM and visual quality.
Additionally, weight the SSD values from RD based on the relative QP offset for chroma; helps visually at high QPs where chroma has a lower QP than luma.
This fixes some visual artifacts created by QPRD at high QPs.
Note that this generally hurts PSNR and SSIM, and so is only on when psy-RD is on.
--aq-mode 1 -- 》2
New AQ algorithm option
"Auto-variance" uses log(var)^2 instead of log(var) and attempts to adapt strength per-frame.
Generates significantly better SSIM; on by default with --tune ssim.
Whether it generates visually better quality is still up for debate.
Available as --aq-mode 2.
ki。曲奇@2009-07-28 17:57
這個是那版本?
最近更新真快
我記得上星期還在用1173
今天己經1192
zjhray@2009-07-28 18:51
好像都还不怎么稳定，subme10这几天修修补补了好几次，还是过段时间再用
暗影翔@2009-07-28 19:49
EP参数
蓝の回忆@2009-07-28 20:00
越来越慢咯...
等各位ep人士给详细解释
roozhou@2009-07-28 20:13
subme 10还要开启trellis 2才能用
linuxyouxia@2009-07-28 20:43
感觉crf模式用了--aq-mode 2需要要把crf值调低一点，这样码率才和用--aq-mode 1时差不多。。。
似乎--aq-mode 2是用来取代--aq-mode 1 最终成为默认AQ的？
roozhou@2009-07-28 20:52
aq_mode 2就是那个AutoVAQ
linuxyouxia@2009-07-28 20:54
上doom9查了下，有AutoVAQ相关讨论：
http://forum.doom9.org/showthread.php?t=147067
ssnake@2009-07-28 22:20
反正现在觉得停在1184以下比较好。。
subme 10个人还没来得及用，反正出差在外暂时也没机会了= =
AVAQ早已有之，个人已经用了接近一个季度了，感觉是适应范围不如VAQ那么大，一定程度上替代OreAQ是可以的（如果不打算用AQDebug的话）。举例来说，这季GA用AVAQ效果不错，比VAQ画面舒服，比OreAQ省码率（以上AQ全使用默认参数）。
AVAQ的话，qp相比VAQ控制的要更好，所以参考VAQ设置的话crf可以适当开低。
52wy@2009-07-28 22:39
subme10比9慢多少？
superkidx@2009-07-29 00:55
初步测试 aq 2 = aq 0 》》》aq 1
film anime 都测了 结果一致
图片就不贴了
重点关注: QP的分配，8x8 transform，coded y,uvDC,uvAC
aq 2
x264 [info]: slice I:3 Avg QP:20.14 size: 7941
x264 [info]: slice P:176 Avg QP:21.43 size: 4888
x264 [info]: slice B:85 Avg QP:25.46 size: 1789
x264 [info]: 8x8 transform intra:35.1% inter:74.0%
x264 [info]: coded y,uvDC,uvAC intra:21.4% 40.7% 8.8% inter:20.3% 25.2% 1.4%
x264 [info]: kb/s:752.9
encoded 264 frames, 3.90 fps, 755.78 kb/s
aq 1
x264 [info]: slice I:3 Avg QP:16.18 size: 9949
x264 [info]: slice P:176 Avg QP:20.69 size: 4755
x264 [info]: slice B:85 Avg QP:23.39 size: 1809
x264 [info]: 8x8 transform intra:43.3% inter:75.5%
x264 [info]: coded y,uvDC,uvAC intra:29.2% 45.9% 10.6% inter:18.3% 23.8% 0.9%
x264 [info]: kb/s:741.4
encoded 264 frames, 3.91 fps, 744.28 kb/s
aq 0
x264 [info]: slice I:3 Avg QP:21.33 size: 7401
x264 [info]: slice P:176 Avg QP:22.71 size: 4827
x264 [info]: slice B:85 Avg QP:25.95 size: 1981
x264 [info]: 8x8 transform intra:24.4% inter:72.0%
x264 [info]: coded y,uvDC,uvAC intra:14.4% 33.7% 6.7% inter:21.2% 25.4% 2.0%
x264 [info]: kb/s:755.7
encoded 264 frames, 4.06 fps, 758.59 kb/s
ssnake@2009-07-29 06:58
且不说ssim和psnr都没有,起码得有参数和截图吧- -
ljwing@2009-07-29 09:13
恩
很久没翻翻东西了
roger有如下评论
关于subme 10，试用了两天，大概码率可以降低10%以上，而且静态比动态码率降得多，很适合压动画，
不过压制速度好像降了20%。
前些时间新加的AQ mode 2,对码率也有5%-10% 的降低，推荐使用
在发布圣机师03的一段话
-----------------------------------------------------------------------------------
嗨
本来我的U就只跑0.55fps
在速度下降就崩溃了
还有翻到点东西
"subme 10" is subme 9 combined with a 7x7 exhaustive qpel RD search. "subme 11" is subme 9 combined with an 11x11 exhaustive qpel RD search.
zjhray@2009-07-29 10:50
请问ls roger在哪个网站出没
«123»共3页
| TOP