Why Single Pass with "Quality Setting"?

[Yo]

Why does AutoGK encode double-pass if you select a target size, and encode single-pass if you choose a quality percentage instead?

From the little I understand double-pass encoding usually produces better results, but takes considerably longer. So there is a trade-off in encoding speed vs. quality of results. Correct?

Why does AutoGK choose first pass or second pass on the basis of whether one selected a target file size, or a percentage quality setting?

[jggimi]

The purpose of a single pass is to use a constant quantizer or to select a constant bitrate. This is regardless of the complexity of the content.

The purpose of multiple passes is to have a variable bitrate, based upon the complexity of the content. The first pass is used to analyze the content, the second (and possibly additional) pass(es) is(are) to produce variable bitrate output.

AutoGK uses two passes for variable bitrate to produce a fixed output size with maximized quality. If you don't have a specific output size in mind, a single pass using constant quantizer values works well.

[Taurus]

[QUOTE]Originally posted by Yo
[B]Why does AutoGK encode double-pass if you select a target size, and encode single-pass if you choose a quality percentage instead?

To reach a desired target size the codec has to do an analizing passage (step) which is fed afterwards via a stats file to the second ''real'' encoding step. Thus assuring the bitrate fluktuation meets the desired target rate.

In quality mode you will never reach exactly a given target size.
cos everything is encoded at a constant quantization.
This means, a clip encoded at quant 2.35, that's about 85% in AutoGk quality terms, will give you different file sizes depending on the resolution you choose and compressibility of the source.
16:9, 4:3, anamorphic, Pal or Ntsc will give you different results.
Dark or heavy coloured, high motions too.
So if you're looking for quality this is your way to go.
But be warned, size can differ extremly.
If you have to fill a desired filesize, like CD or even DVD,
a two pass encode is the only way to go, to let the codec know how to distribute the bitrate.

Hope this helps

...and its explained a thousand? times in this forum.

Cheers

Taurus

[edit]: Oops! Jggimi was faster!

[Yo]

Originally posted by Taurus

If you have to fill a desired filesize, like CD or even DVD,
a two pass encode is the only way to go, to let the codec know how to distribute the bitrate.



OK. I've read elsewhere (about encoding in general, not specifically about AutoGK) that you get better results in general by using two pass, although it takes longer. They weren't referring to fitting in a specific target size.

[jggimi]

If filesize is not an issue, your best results will be a single pass with the minimum Quantizer value allowed, usually Quantizer=2 (100% Quality).

[edeus]

what is the point of Q1 if it doesnt improve quality, only file size?

[KpeX]

Originally posted by edeus
what is the point of Q1 if it doesnt improve quality, only file size? In the case of xvid, quant 1 is used to avoid undersized complaints from users - specifically, the second pass can be larger than the first.

As far as other codecs, I'm not sure why or if they even use quant 1 - I can't remember the last time I used a codec that wasn't xvid ;)

[manono]

Hi-

I've read elsewhere (about encoding in general, not specifically about AutoGK) that you get better results in general by using two pass

My guess is that you read something old, and they were referring to 1-pass CBR encoding .vs 2-pass VBR encoding. When compared to CBR, 2-pass VBR was a huge step forward. Nowadays, 1-pass Quality based encoding, such as is possible with AutoGK, is VBR, but without the file size control possible with 2-pass methods.

[Yo]

Originally posted by jggimi
If filesize is not an issue, your best results will be a single pass with the minimum Quantizer value allowed, usually Quantizer=2 (100% Quality).

But filesize is always an issue, no? If it wasn't, just backup the whole DVD to HD as DVD folder. The reason one wishes to convert to DIVX or XVID AVI is to make a much smaller file size.

Of course, some people have very specific file size requirements, like wanting to fit it on one CD. That may be what you mean.

But getting a lower file size (per quality of video) is always a desired result, no?

[len0x]

Originally posted by Yo
But filesize is always an issue, no? If it wasn't, just backup the whole DVD to HD as DVD folder. The reason one wishes to convert to DIVX or XVID AVI is to make a much smaller file size.

some do, some don't. MPEG4 is capable of achieving much smaller sizes than MPEG2 even at 100% quality (what some ppl do). While 100% may be overkill, 70-80% will do just fine if you don't care about exact size.

[DeadMetal]

I have a question which fits into this topic (although it's an old topic).

I am trying to find the best possible quality before I start backing up all my movies.

I was wondering:

- If I create a XVID movie with, for example, 90% quality, the total filesize could become in this case, for example, 1000MB.
- Then, if I would do the same, but now I configure that AutoGK should strive for a custom size of 1000MB, which of the movies will have a higher quality?

[len0x]

They will have the same quality, but you'll spend much more time doing two passes.