DoCCE4U RoBa Method v1.0.0.4

[The_Flash]

When this template is selected, the first time CCE runs to generate the .vaf it is in VBR mode. The correct mode should be 1Pass w/ a Q of 40. The way things are setup now CCE generates a .vaf based on your avg. bitrate and not the specified Q.

[BBWoof]

Originally posted by The_Flash
When this template is selected, the first time CCE runs to generate the .vaf it is in VBR mode. The correct mode should be 1Pass w/ a Q of 40. The way things are setup now CCE generates a .vaf based on your avg. bitrate and not the specified Q.

I didn't know and am not sure that the Q factor setting is hard set at 40. I think it should be changed depending on the source.

In 1.0.0.5, I've changed the RoBa method back to a Multipass VBR run since the CCE Manual states that on a Multipass VBR the first pass is run as a 1Pass VBR to check the complexity.

BBWoof

[Doom9]

hmm.. you're correct. I used an old screenshot which still shows linear quantizer scale checked, and for simplicity's sake I left out the bias in the first pass since it would require ecl editing (at that time I didn't put anything about chapter positions in the guide) but since ecl editing is now part of the guide I guess I'll change that.

But other than that we have 2) ok, 4) not applicable since in the DVD-R case and Q40 we'll end up increasing our main movie bitrate in 99% of the cases, 5) ok and 6) is ok, too.

[BBWoof]

I guess I should have gone to the author of the method to get my info in the first place. I will once again modify the program to do a RoBa method encode in two CCE executions.

I will make the first pass 1Pass VBR with a Q factor of 40, and the second a Multipass VBR using the vaf created in the first run.

Thanks for the clarifications.

BBWoof

[Fizban99]

Shouldn't, for the First and only 1 Pass VBR that creats the vaf file for use in the subsequent pass(es), the Q value be variable, not hardcoded at 40? Or am I reading something wrong here?

I'm anxiously waiting this new version :) having the Robshot-bach method automated will be a great help to me.

[ShaneZ]

Originally posted by Fizban99
Shouldn't, for the First and only 1 Pass VBR that creats the vaf file for use in the subsequent pass(es), the Q value be variable, not hardcoded at 40? Or am I reading something wrong here?

I'm anxiously waiting this new version :) having the Robshot-bach method automated will be a great help to me.

I may be completely off base here, but the Bach's method was one-pass CBR adjusting the Q factor on a 1% pass of the source file. The reason being, adjusting the Q factor up or down will lower or raise the filesize respectively. Once you get your 1% value, you multiply that by 100 to get your projected output filesize.

The RoBa method is a bit different, which is why the Q factor is set at 40. The two methods are different.

[BBWoof]

Originally posted by Bach
The so called RoBa method uses a fixed Q.factor. However, the value of this Q.factor is a personal choice, being any value between 1 and 40. So, as I said before, "up to 40"...

The so called Bach method, which is in my opinion as good as RoBa but 94% faster, you can find in my first post HERE (http://forum.doom9.org/showthread.php?s=&postid=283536#post283536). The settings are the same of those suggested in RoBa. The only difference is that you will use only one pass instead of two.

Thanks for the info Bach, if I can figure out how to programmatically add the Bach method to my program I will do that. And for RoBa I'll leave the Q factor up to the user (which actually makes it easier for me). And the next version will use the 1Pass VBR to create the vaf file.

BBWoof