Q factor questions...

[nautilus7]

I am new to cce encoding and i would like to help me understand what exacly is the Q factor at the 1 pass VBR and how it is set correctly.

For example: In AutoQ 2 i put the size of the audio streams etc, i used 4% of my movie and it predict that the right Q value was 2. Then, with this value i done a 1 pass vbr to the whole movie, but the file size was 4,36 GB, which is too big.

Do i have to set another Q value and try again or i can procced to multipass vbr encoding as it is?
What is the muxing/authoring value in AutoQ 2?

[auenf]

a Q setting of 60 means a CQ VBR where the Quantizer of the encoded m2v will pretty much stay at a 9, it will only go higher if you dont let the bitrate go high enough

the lower the Q setting, the lower the quantizer will be in the encoded m2v.

Enf...

[Amnon82]

Seems u did something wrong. AutoQ2 prediction is normaly ok. As u say that ur final q-value was 2 u had a movie with low action or many dark parts. (just guessing)

Higher the sample % value. Higher the minimum bitrate. lower the maximum bitrate. Try the different prediction modes.

For the q-value: the lower the q-value the higher the bitrate.

[nautilus7]

Thanks for your reply. The movie is "the league of extraordinary gentlemen". I 'll do what you said.
What will happen (in quality) if i use the .vaf file from this "1 pass vbr" (the one i created with the wrong Q value) at the forthcoming multi vbr passes?

[Amnon82]

U'll save u the first pass of the multipasses (vaf creation). For example if u choose 2 pass CCE runs only 2 times when u use the vaf-file, otherwise it will run 3 times. (1 Pass for VAF, 1. pass of multipass, 2. pass of multipass)

ATM I only do 1 pass vbr cos I don't see any quality differences. I used only 1multipass for Day after Tomorrow, cos of the lightnings. (0 pass for vaf and 1 multipass)

[Trahald]

q changes that under 6 tend to have large jumps between them.. meaning say a q of 3 for your movie would likely be well undersize .. its hard for any q predictor to be accurate with that low of a q.