@Quality encoding idea .... Easy to implement I mean.

[Q-W-Y]

When i want to do quality based encoding, i am doing this:


1. 1pass, quant 3 (if i need avg quality 3)
After it, I write down AVG bitrate from XviD Status window.

2. 1pass, weight 1.0

3. 2pass - at target bitrate which i've noted before.


I mean, that it could be easily automatized ... What do you mean?

[sysKin]

Originally posted by Q-W-Y
I mean, that it could be easily automatized ... What do you mean? Now this is amazing. Believe it or not, but *today* *three* people told me to do something that have been done in the last two days.

Just get Gamer's build whitch is less than two hours old...

Radek

[Q-W-Y]

Now, i have downloaded Gamr's buid.

There is Single Pass with quantizer setting, but i mean to do 3 passes. Because Single pass has only Quants at 3 or 5.
My method has better results. It uses quants 2-7.

So we have talk at cross purposes....

First 1st pass like "compressibility checker", then normal 1st,2nd with bitrate from "compressibility check"

[JimiK]

Maybe that would be a good idea for an built-in 'bitrate' calculator. As I understand it, XviD guesses the quants for the 2nd pass from the data collected in the first pass. So it knows, this frames uses X bits at quant2, so it would need approximately Y bits at quant3. So XviD could try to calculate the size if constant quant3 would be used. Then one could take this size, enter it into the final size and get a better distribution than at constant quant3, but does not have to make an additional encoding pass. (if XviD works this way)
Best regards,
JimiK

[sysKin]

The quant distribution from 2-pass is definitely not better than constant quant - it's just more random and imprecise. I doubt you get a better picture by artificially introducing quantizer fluctuations...

Radek

[Q-W-Y]

Originally posted by sysKin
The quant distribution from 2-pass is definitely not better than constant quant - it's just more random and imprecise. I doubt you get a better picture by artificially introducing quantizer fluctuations...
Radek

I mean, that if i do 2 passes, then bitrate is better granted to scenes which it needs. Am I right?

[Assault]

No! Don't forget that a constant quant encode doesn't mean a constant bitrate encode. It's the other way round. Constant quants mean a variable bitrate (difficult scenes get more bits and simple scenes get less ;)). So as sysKin already said a 2pass encode won't give you better quality than a constant quant encode with the same size.

Assault

[sysKin]

Originally posted by Q-W-Y
I mean, that if i do 2 passes, then bitrate is better granted to scenes which it needs. Am I right? Nope, don't overestimate 2-pass encoding. It doesn't know what bitrate scenes need. It just tries to achive constant quant AND desired bitrate, which is simply more difficulat than constant quant alone.