PDA

View Full Version : how much can 2pass/1pass's ratio reach?


jeffhuge
5th February 2005, 04:58
the highest and the lowest
mine is up to 79% high,is that normal ?

btw:with 0.15 weight in the zone setting in the credit scence in 1pass

tripnotik
5th February 2005, 05:27
It depends on the bitrate or target size you choose. If you set the quantizer range to be 1-31, you can have a second pass larger than the first pass.

Personally, I encode my movies at 60% of first pass.

jeffhuge
5th February 2005, 07:12
Originally posted by tripnotik
It depends on the bitrate or target size you choose. If you set the quantizer range to be 1-31, you can have a second pass larger than the first pass.



can or must , i had set the quantizer at those numbers several times but I didnt ger the result as you said .

ATM
5th February 2005, 07:15
Originally posted by jeffhuge
can or must , i had set the quantizer at those numbers several times but I didnt ger the result as you said .

1-31 quantizer range has a possibility of giving you an oversized file which is why people use the 2-31 range, but at the same time this has a chance of giving you an undersized file (though I always use 2-31 and I have never had a problem with either undersized or oversized files).

jeffhuge
5th February 2005, 08:07
Originally posted by ATM
1-31 quantizer range has a possibility of giving you an oversized file which is why people use the 2-31 range, but at the same time this has a chance of giving you an undersized file (though I always use 2-31 and I have never had a problem with either undersized or oversized files).

thanks

I cant understand the difference between 1-31 and 2-31.They all let xvid decide the biggest quantizer and donot restict,right?

ATM
5th February 2005, 08:13
Originally posted by jeffhuge
thanks

I cant understand the difference between 1-31 and 2-31.They all let xvid decide the biggest quantizer and donot restict,right?

A frame with a quantizer of 1 usually will be significantly bigger then a quantizer of 2 so the problem is that if the codec starts using alot of quant 1s in the encode it can bloat the filesize, leading to people having oversized files over the years, which is why it is usually recommended that you use 2-31.

Koepi
5th February 2005, 08:31
People don't "often" get oversized files when using a quantiser range from 1 to 31.

That case is in the contrary quite seldom, you'd read _a lot more_ posts about oversized files like we do now.

Just set the overflow treatment values to 10% and leave the quantiser range at 1-31 in the case of oversized encodes, they'll vanish then.

Cheers,
Koepi

jeffhuge
5th February 2005, 09:41
thanks
ATM
KOEPI

Originally posted by Koepi
People don't "often" get oversized files when using a quantiser range from 1 to 31.

That case is in the contrary quite seldom, you'd read _a lot more_ posts about oversized files like we do now.

Just set the overflow treatment values to 10% and leave the quantiser range at 1-31 in the case of oversized encodes, they'll vanish then.

Cheers,
Koepi


the overflow treatment values to 10% means
overflow control strength
max overflow improvment
max overflow degration

all the 3 set to 10?

and leave the quantizer to 1-31




I'll rip it again coz since I change to 1.1 there is always a oversize problem confusing me.(always 10-30mb bigger),

thanks again
JeFfHugE

Zep
11th February 2005, 19:56
Originally posted by Koepi
People don't "often" get oversized files when using a quantiser range from 1 to 31.

That case is in the contrary quite seldom, you'd read _a lot more_ posts about oversized files like we do now.

Just set the overflow treatment values to 10% and leave the quantiser range at 1-31 in the case of oversized encodes, they'll vanish then.

Cheers,
Koepi

did that, didn't help. still over by 6 megs
Xvid 1.1 beta. i tried 20 20 20 also still 6 megs over.

the Q average was 5.14.
42 minutes long encode.

why would a quantiser range at 1-31 be better and stop
oversized files than a quantiser range at 2-31? The problem
clearly is HUGE Q=1 frames.


any other suggestions?

jon.schaffer
12th February 2005, 00:56
Zep, could you try these settings and report the results, please?:

Leave Q: 1-31 for each frame-type
And:
- Overflow control strength 0
- Max. overflow improvement 4
- Max. overflow degradation 9

(consider the quality too: is there some "jumps" in quality...)