CCE SP 2.70 out!

[NaN]

Was released today. Read about it here:

http://www.cinemacraft.com/eng/topic.html#31

Cheers, NaN

[Boulder]

These two sound interesting:

- Allow adaptive quantization matrices switching for multi-pass VBR
- Automatic switching of block scan order

Now if they could get field-based encoding..

[RB]

This looks extremely cool. Ability to specify linear quantization type is back. Log file can be saved again. Finally an option to output BFF streams. And you can now do up to 99 VBR passes. God, please no! I already see discussions about whether 60 or 90 passes are enough plastering the forums everywhere :D

And some very good news about the trial version:

you can use the advanced quality settings that were previously disabled in the trials
ECL support is back! No need for EclCCE anymore :)

[wmansir]

That's odd, the readme says ecl save/load is still a trial limitation, but it's there. Sweet.

Here's some screenshots of the new GUI and options:

http://img90.exs.cx/img90/8631/CCE270_GUI.gif

http://img90.exs.cx/img90/6694/CCE270_PIC.gif

http://img90.exs.cx/img90/3233/CCE270_VID.gif

[Matthew]

Originally posted by RB
ECL support is back! No need for EclCCE anymore :)

But your command-line options + minimise/credits tweak are great. So I for one hope you'll add support for this trial version =)

[FilipeAmadeuO]

The most important for me is :

"Quality improvement in fades on a static scene"

I think it was the major problem that i had.

Where can i download it ??

[Boulder]

It's available at Cinema Craft's Japanese site or at visiblelight.com.

[hendrix]

I love that you can now select the final size before you encode - very useful if you do OPV encodes. The New GUI isn't so bad either - time to test the Inverse 3:2 pulldown and 99 passes is a bit extreme

[hobyho]

Originally posted by hendrix
99 passes is a bit extreme

Maybe not if you have a supercomputer lying around ;)

[hendrix]

Originally posted by hobyho
Maybe not if you have a supercomputer lying around ;)
or if you have nothing better to do :eek:

[Cwluc]

anything past 10 passes...can we even tell anymore?

[RB]

Originally posted by Matthew
But your command-line options + minimise/credits tweak are great. So I for one hope you'll add support for this trial version =) Command line should be working in the trial version too, somebody could check this please? As for minimize/credits tweak - yeah, it'll be sort of a "EclCCE lite" especially for CCE-SP 2.70 :) But not before next week. Real life is keeping me real busy right now...

Originally posted by Cwlucanything past 10 passes...can we even tell anymore?In my very humble opinion, the ability to do more than 1 VBR pass (aside from the VAF pass) in CCE is pure marketing fluff. After analyzing the full video (VAF pass), all the information for optimal bitrate distribution is there. Think about it, if CCE doesn't get it right the first time and really needs more passes for "optimal quality", wouldn't that mean their engine is seriously flawed?

[br408408]

Originally posted by RB

In my very humble opinion, the ability to do more than 1 VBR pass (aside from the VAF pass) in CCE is pure marketing fluff. After analyzing the full video (VAF pass), all the information for optimal bitrate distribution is there. Think about it, if CCE doesn't get it right the first time and really needs more passes for "optimal quality", wouldn't that mean their engine is seriously flawed?

Very good point!

[erdoke]

Rapid correction just one day later: 2.70.01.01
New version can be downloaded from the same location, and there is an upgrade version too (2.70.01.00 -> 2.70.01.01).

[Cwluc]

Originally posted by RB
Command line should be working in the trial version too, somebody could check this please? As for minimize/credits tweak - yeah, it'll be sort of a "EclCCE lite" especially for CCE-SP 2.70 :) But not before next week. Real life is keeping me real busy right now...

In my very humble opinion, the ability to do more than 1 VBR pass (aside from the VAF pass) in CCE is pure marketing fluff. After analyzing the full video (VAF pass), all the information for optimal bitrate distribution is there. Think about it, if CCE doesn't get it right the first time and really needs more passes for "optimal quality", wouldn't that mean their engine is seriously flawed?

That is an amazing point, I never even thought of the passes that way, which is why you have always got my respect. We should all take a note from someone who actually knows what they are talking about...

[Matthew]

Originally posted by RB
Think about it, if CCE doesn't get it right the first time and really needs more passes for "optimal quality", wouldn't that mean their engine is seriously flawed?

I don't necessarily agree with that, because I imagine a "sloppy" pass is a lot quicker than a "thorough" pass and as such a multi-pass system permits users to make a more flexible speed/quality tradeoff decision. As opposed to just a choice of CBR,1-pass VBR and (slow) 2-pass VBR.

Also, your command-line switches are better than the real thing e.g. -nocancel :/

[Boulder]

Maybe CCE works on several sections, I mean, if you have a 2-hour video and there's a lot of low-bitrate scenes at the beginning (like credits etc.) the saved bits may not show up in the end of the video but near the section where they were saved. Maybe the encoder shifts the bits pass-by-pass so that the saved bits are more evenly spread among the whole video. After three to four passes the bits are spread as well as they can be?

[GZZ]

I still think that more passes are better then one, because the encoder don't have a briliant AI, so it does what is most logic. It got though the video and collect infomationer about movement what color those the picture consists of. It then encode the video with the info it have collected. Then it go though the video again (second pass) and look at the previous picture and maybe have an advanced algorithm which look at the picture and try looks for pixels or other flaws in the picture and then try to raise or lower the bitrate. The more time it does this process the better the picture will look. I know there is a max number of pass before the changes in the picture will not be visible anymore...Thanks why Cinemacraft writes 3-4 pass as max in the help file, but my expirence in encoding movies shows that 5 or 6 pass can improved quality on video at very low bitrate. Like if your avg bitrate are around 3000 kbit or under. Some people will properly disagree with me and other will say more passes are better. I don't know why the change the max number of pass to 99, if they haven't change the way it encode the video, so more passes are need to have the same quality as before at low passes, like 5 pass in CCE 2.67 are equel to 10 pass in CCE 2.70 ?? Don't know if this is true, its pure speculation on why they change it from max 9 passes before to 99 passes. :/


Sorry for the bad english, hope you all understand my ideas.

GZZ

[dragongodz]

After analyzing the full video (VAF pass), all the information for optimal bitrate distribution is there. Think about it, if CCE doesn't get it right the first time and really needs more passes for "optimal quality", wouldn't that mean their engine is seriously flawed?
depends. is the VAF pass doing the equivilant of a constant quant pass to gain statistics ?
if yes then a second pass should be all it needs with maybe a third to fix maybe minor mistakes.

even in their own documentation(someone quoted it in the forum some time ago) they say anything over 3 or 4 passes is pretty pointless.

99 passes ? either its for marketing or they have really screwed the engine up somewhere. maybe one of the programmers was typing 9 and accidentily double tapped it. :D

[hendrix]

Originally posted by GZZ
so more passes are need to have the same quality as before at low passes, like 5 pass in CCE 2.67 are equel to 10 pass in CCE 2.70 ??
i dont think it would take 10 passes in 2.70 to have the same quality as 5 passes in 2.67 - if that were the case then Cinema Craft seriously downgraded the encoding engine - vice versa i can see.

im not going be trying the 99 pass encoding - i pass on that one (get it?) :D

[MLS]

"Quality improvement in fades on a static scene"

Thank god is all I can say. CCE really had problems with these.

/MLS

[Boulder]

Originally posted by dragongodz
depends. is the VAF pass doing the equivilant of a constant quant pass to gain statistics ?
if yes then a second pass should be all it needs with maybe a third to fix maybe minor mistakes.


In some older version, probably v2.50, the encoder ran the VAF creation as constant bitrate, the bitrate set at the desired average. In v2.67.00.27 the bitrate varies during the creation of the VAF file but the quantizers are not constant. The quants are only constant when OPV is used.

Maybe a conclusion can be drawn from this : OPV + sizing pass should yield optimal bitrate distribution provided that the OPV pass produces an average bitrate which is close enough to the desired one. Thus a D2SRoBa approach might be beneficial in DVD-RB?

[crespo80]

is there any improvement on the audio quality?

[Arky]

Originally posted by crespo80
is there any improvement on the audio quality?

Respectfully, does this even matter?

What I mean by this is that anyone serious about DVD authoring (and CCE is for serious MPEG 2 encoding work...) would also encode their audio to .ac3, for maximum compatibility. Since CCE does not encode to .ac3, it seems useful only for MPEG 2 video encoding, anyway, wouldn't you agree?



Arky ;o)

[auenf]

Originally posted by Arky
Respectfully, does this even matter?

What I mean by this is that anyone serious about DVD authoring (and CCE is for serious MPEG 2 encoding work...) would also encode their audio to .ac3, for maximum compatibility. Since CCE does not encode to .ac3, it seems useful only for MPEG 2 video encoding, anyway, wouldn't you agree?



Arky ;o)

exactly, and reminds me that i still need to wait for Apple to fix the A.Pack dual processor bug that causes encoding to fail unless you disable a processor, or do something else on the machine at the same time.

Enf...

[Angelus]

Is there any way to use this CCE with BatchCCEWS? I'm guessing not because the eclcce doesn't support this version yet. Or is there an *easy* manual way to encode it directly? Thanks and sorry for my ingnorance ;)