PDA

View Full Version : Why only 1 pass when using Target Quality?


dwfromm
14th August 2005, 05:39
While I am quite generally quite happy with AutoGK,
it does seem to me that at least having any option to
do a 2 pass encode when using Target Quality setting
would be a welcome enhancement for those of us that
leave our encodes on a hard drive.

I do understand that the purpose of a multipass encoding
is to maximize quality for a particular file size, but given
a target quality I believe that a 2 pass encode would also
deliver the smallest file size for the specified quality. And
for the growing number of folks streaming video to settop
boxes isn't that what is generally desired? At least for me,
that would be the holy grail of DVD->DIVX encoding. :)

Is there much chance of seeing a 2-pass quality target
encode option for AutoGK?

Thanks,
Dwight

manono
14th August 2005, 08:14
Hello, and welcome to the forum-

...it does seem to me that at least having any option to
do a 2 pass encode when using Target Quality setting
would be a welcome enhancement

Enhancement? In what way? You think you'll get better quality that way? You won't. You'll just be wasting the time it takes to do the second pass.

...but given a target quality I believe that a 2 pass encode would also
deliver the smallest file size for the specified quality.

It'll deliver the same file size, or very close to it.

Is there much chance of seeing a 2-pass quality target

I can't speak for len0x, but an educated guess says no.

If anything, traditional 2-pass encoding delivers arguably worse quality. If you, like me, want the entire video to have equal quality, complex scenes and static scenes both, then it doesn't get any better than 1-pass quality based encoding. The 2-pass mode of AutoGK, because of the way it restricts the quants, delivers better quality than the traditional methods of leaving the quants within the range of 2-31, Now, if you want the complex scenes to have worse quality than the static scenes, then the quant jumps that take place during such scenes will free up more bits to improve the static scenes. I just don't happen to want that. As has been repeated ad nauseam around here, the only reason for 2-pass encoding is to fit the video to a specific file size.