『漫游』酷论坛>『影音数码技术学习交流』>关于x264编码psy_rd参数 ..
superkidx@2008-08-19 23:02
引用
最初由 leeyoung 发布
呃 某大就免了 [/han]
我不明白 你指的是 什么AQ？
就是现在官方版本中的aq啊
另外那个 翻成英文比较好懂 呵呵 果然在线翻译还是最重视e文啊
The object is seen or is not dimly slightly seen or be of the correspondence to "Shade part" in the state of the dark of
extent. QP is raised if there is no image and data is saved, and it leaves untouched or QP is lowered and the image quality is raised when there is something.
It is a scene of the solid black to correspond to "Super-shade part".The state that Y cannot check anything visually unlimitedly and in the vicinity to 0 generates the color noise in UV, and consumes data sadly on the data structure named YUV somewhat. This raises and saves QP because it is useless.
When Y value is low, it is practical there though the verge in the shade part and the super-shade part is severe, "There is no other way" though QP is raised by assumption "When seeing as animation, it is not possible to check it visually".
Such brightness is not judged in VAQ when MixAQ is compared with OreAQ, it designs , saying that "
The shade part is weak to OreAQ" to protect it neatly by comparing it with AQ unused because HAQ is
AQ originally made by the patch purpose of the shade part though it is likely to be received.
ps MythCreator 你的签名超标咯
ps2 ly你那个vob是哪里找的啊 我怎么找都只有vcd级别的
huhuyayas@2008-08-20 20:50
一般还在用 2pass~~(看来这个做法有点虐机） 感觉不是很明显~不过 x264在低码率的时候，其实表现并不是太尽如人意思~~ 中码率的时候， 因为psy_rd 如此导致的 码率变化， 相对于码率不太明显了。。。
再 希望 x264日趋完美
zjhray@2008-08-20 21:10
psy_rd感觉一定要用在干净的源上，否则就像罩着一层裂开的玻璃一样
セイバー@2008-08-21 10:44
谁还有那个日本用户修改的版本啊?
今天去那个日本人网站下了个938的,发现什么都没了.....
leeyoung@2008-08-21 11:21
引用
最初由 セイバー 发布
谁还有那个日本用户修改的版本啊?
今天去那个日本人网站下了个938的,发现什么都没了.....
08.08.21 AM11：18
检查这个链接：
http://seraphy.fam.cx/~seraphy/program/x264/おまけ/
r938 等等都在
セイバー@2008-08-21 11:45
原来是我下错了.....
leeyoung@2008-08-21 11:53
引用
最初由 セイバー 发布
原来是我下错了.....
LS的同学 能否对OreAQ的原理 谈谈你的理解？ 不胜感激
セイバー@2008-08-21 12:02
我不知道,我只是个超级菜鸟.....
leeyoung@2008-08-21 12:04
引用
最初由 セイバー 发布
我不知道,我只是个超级菜鸟.....
其实 我想找个 人肉翻译机:o
qyqgpower@2008-08-22 22:54
x264的psy-rd“优化”得越来越不适合动画了，ringing变本加厉，低对比度轮廓一团模糊
uc0083@2008-08-22 23:23
引用
最初由 qyqgpower 发布
x264的psy-rd“优化”得越来越不适合动画了，ringing变本加厉，低对比度轮廓一团模糊
不是现在有cartoon的预设了吗，x264默认看来就不是用来压动画的
superkidx@2008-08-22 23:42
引用
it's AQ that is messing with edges...
qyqgpower@2008-08-23 08:27
it's AQ2 or noAQ that is messing with low contrast edges.
Ringing is mainly introduced by psy-rd and enhanced by AQ2.
bmp samples
http://www.mediafire.com/?80tdgk4iaia
comparing the edge of the shadow
noAQnoPsy OUT
2AQnoPsy OUT
noAQPsy11 OUT
comparing around her collar
1AQPsy11 OUT
2AQPsy11 OUT
best result: 1AQnoPsy
full set of the samples
http://www.namipan.com/d/46beb82b9bf935da24b4e5ce30f18b4b02e0204f26a97205
52wy@2008-08-23 09:03
引用
最初由 uc0083 发布
不是现在有cartoon的预设了吗，x264默认看来就不是用来压动画的
为啥我更新了米发现cartoon的预设=。=
superkidx@2008-08-23 09:21
引用
最初由 qyqgpower 发布
it's AQ2 or noAQ that is messing with low contrast edges.
Ringing is mainly introduced by psy-rd and enhanced by AQ2.
bmp samples
http://www.mediafire.com/?80tdgk4iaia
comparing the edge of the shadow
noAQnoPsy OUT
2AQnoPsy OUT
noAQPsy11 OUT
comparing around her collar
1AQPsy11 OUT
2AQPsy11 OUT
best result: 1AQnoPsy
full set of the samples
http://www.namipan.com/d/46beb82b9bf935da24b4e5ce30f18b4b02e0204f26a97205
我自己只测试过aq 就低对比度的边缘而言 最好的是noaq 其次是aq1
«34567»共7页
| TOP