Matrixs & DVDRB

[Mr. Monte]

O.k..my next step is to get into applying different matrixes based on the proposed bitrate. So I have a few questions if some peeps could assist and advise me if this is a viable way to do it..or if they feel theres other ways easier and or more reliable

I use CCE 2.70.2.4 for all my encodes, and in the past just use the encoder default and recently CCEAQM option on. However after reading more and more...I'm wondering if a different starting matrix might be benificial to my final output. I noticed according to RME, that the CCE default matrix is optimal for bitrates around 2800-3500. However, some of the movies I have been encoding show an average bitrate in the CCE encoding screen of over 4000. So my assumption would be that I could use a "better", aka Fox Entertainment, matrix for those type of encodes and possibly get a "better" encode.

Would this be a logical and correct assumption on my part?

If so, might it be possible for jdobbs to add a option in a future release of RBPro to look at the average or overall bitrate after the prepare phase and apply the user defined matrix at that point? So the user could set up a head of time to use the encoder default on 2800-3500 bitrate, fox for anything above 4000 and Angel Low for anything below 2500? (Just Pull-Out_Of-Ass examples..to try and get point across)

For me to get an idea of what the encoded bitrate might be..could I just run the DVD main movie VOB through bitrate viewer...see an average of say 7000, then using DVD Shrink..find out what the compression ratio is going to be..then take that times the DVD bitrate and to give me an approxiamate bitrate? (I.E..origianl DVD 7000, compression 50%, so the bitrate of a encode down to 4.7 gig DVD would be about 3500...so DVD RB would use the CCE standard default matrix)?

TIA

[apfraats]

DVD-RB-PRO !!!, already does have the ability to do this.

It has:

1) A main feature Matrix
2) A low bitrate Matrix it uses for segments with avg bitrate below LEVEL A
3) A Very Low Bitrate Matrix it uses for segments with avg below LEVEL B
4) A matrix to be used for what DVD-RB sees as EXTRA'S.

LEVEL A is above LEVEL B, I suppose , but I don't know the real averages used here.

Also, using MIN_BITRATE=0 will have more benefits then using a different matrix.

With higher averages you can indeed use FOX-MATRIX.

But DVD-RB matrices are implemented especially for use on difficult low bitrate stuff. They simply filter out higher frequencies to be quantized, so more bitrate is avialable for lower frewuencies.

This does 2 things:

1) Prevent macro blocking.
2) Smoothen the picture somewhat.

But as already mentioned, using MIN_BIRATE=0 has more benefits than using a matrix.

For higher bitrates the FOX seems to be the best.

If the FOX matrix is causing blocking to occure just use MPEG-STANDARD or CCE DEFAULT, as they deliver great resuls anyway.

Then I saw you use AUTO ADAPTIVE QUANTIZATION , and that has two main problems:

1) You can create NON STANDARD dvd's because of the adaptive switching.
2) Using a matrix will be greatly overwritten by the use of ADAPTIVE QUANTIZATION, making it kind of useless using matrices.

So the suggestion is to leave this option OFF as per default.

The simple warning from CCE that it can cause problems for ceratin DVD-player was convincing me to NOT use it.

By the way, using MATRICES is a kind of advantage fine tuning, you can't do magic with it.....

[Mr. Monte]

app,

Thanks for replying..but as I see it..DVDRB does not do what I implied above (if I'm wrong..please point it out).

If you read above..my suggestion was to:

The user picks his OWN matrixes and sets them up in DVDRB for different levels of matrixes. (Yes, I know if you do 3 mode encoding..you can prepare and then pick your matrix based off the bitrate expected output)..I'm talking in one button mode..guess I should have made that clear above.

So..in one button mode DVDRB expects the bitrate to be 3300..so it looks up in the user setup and see you want to use CCE default. Then the next batch job the expected bitrate is say 4200. Now for this job (one button mode), DVDRB looks at your setup ini file and sees for that range you want to use the Fox matrix..etc..etc

BTW..using CCE does not cause any issue to alot of users here on the board..even I have tested it on about 20 DVD players. Ans as Manano in another thread (matrixes I believe) stated..using CCE WITH AQM got the CCE default matrix closer to FOX on his test media (97 %..I believe).

I've found CCEAQM to be MUCH better than without (now thats using CCE default matrix)..especially on B & W sources

[jdobbs]

That's why I added "Templates" -- for this and other similar uses.

Set the matrices you want for a specific range, then save your settings as a template. Then load that template when you run into a source that matches. That goes for other settings also, like bias, qual_prec...

If you wanted me to find out what the incompatible players are, all I'd have to do is just put AQM in as the default... the emails would be only a few minutes away. ;)

[archaeo]

If you wanted me to find out what the incompatible players are, all I'd have to do is just put AQM in as the default... the emails would be only a few minutes away. ;)

Is CCEAQM that risky? I've heard some rumors, but haven't heard much that's concrete from those who have used it. What is it about an AQM generated matrix that actually causes the problem?

[jdobbs]

I really can't say... but when Cinema Craft themselves posts the warning, it certainly makes me hesitant. My guess would be that it might only cause issues on older players -- but that is still just a guess.

[Mr. Monte]

That's why I added "Templates" -- for this and other similar uses.

Set the matrices you want for a specific range, then save your settings as a template. Then load that template when you run into a source that matches. That goes for other settings also, like bias, qual_prec...

jdobbs,

Do the templates work in one button mode though?

I.E> DVDRB makes the decision based on the prepare phase automatically? Or do you need to check the approximate bitrate of you output a head of time and pick the template before starting the process?

[jdobbs]

Yes. Templates are similar to project files in that they load settings... but the settings are limited to those options that are transient. So loading a template won't change you paths, skin, etc. like a project would.

[Boulder]

Adaptive quantization is DVD compliant so if you have a standalone that doesn't like it, it's not supposed to be called a DVD player at all.

[Mr. Monte]

If you wanted me to find out what the incompatible players are, all I'd have to do is just put AQM in as the default... the emails would be only a few minutes away. ;)

Yes boulder..but I tend to agree with jdobbs on this one. I think those that have tried it...see the benefits.

[jdobbs]

There's also apparently some difference between Adaptive Quantization and the workings of CCEAQM... someone here on Doom9 pointed it out -- but I can't remember who it was of the exact content.

[Mr. Monte]

I believe Manano talk ed about a test he did with CCE's standard with and without CCEAQM versus Fox Home Entertainment. He said CCE standard with CCEAQM had a result of about 95% detail versus the same encode with Fox home Entertainment.

[Boulder]

Let's just say that if the major studios use the feature on their DVDs, I think it's safe to use it yourself as well. Then again, you have the option to enable it if you like so no one can complain here :)

[apfraats]

I'm always in for better.

But what does CCEAQM precisely ???

Does it change a matrix depending on encoding charasteristics ???

IOW: If I use Matrixeditor to extract the matrix used, I won't get it back ?

Or am I at risk getting an other matrix back ?

Just what does it precisely do ???

Does it adapt quantization charestistics only or adjust the encoder matrix dynamically ?

And what do I get out of it ?

Can it help on difficult low average high compressed stuff to prevent macro-blocking, as that is the most important issue for me. (It seems impossible to tell an encoder to give priority to preventing macroblocking at the costs of loosing details for a moment, which should be much less disturbing)

Still if it were the perfect solution for my problems, I hesitate to use it, because of CCE's own warnings I have also red.

Sure older players are more difficult, especially sensitive for exceeding MAX_BITRATE and other stuff.

It's like an older bug in HC were vectors were calculated 'to point half a pixel out of the frame' that worked fine in lot's of DVD-players but did not on some of them.

First I have to understand what it exactly is doing CCEAQM ?

Someone, please ?

[Boulder]

I think manono explained it quite well recently in some other thread either in this forum or the CCE forum.

[Gibuk]

apfraats -

Encoding Ninja Scroll (anime) with high bitrate, I have listened your advices: using min_bitrate=0, vbr_bias=0, new Fox matrix, I have got a video more clear and than original.. why?
Should I have to use vbr_bias > 0, for example = 10?

[archaeo]

Adaptive quantization is DVD compliant so if you have a standalone that doesn't like it, it's not supposed to be called a DVD player at all.

To try to find out more, I sent an e-mail to CCE tech help to ask specifically what is the problem that they're warning about. They said that it's not that the player won't play it - it's that some (older) players would show block noise if the feature is enabled and it doesn't like it. So it appears that the symptoms someone would look for w/AQM incompatibility would be DCT blocking in playback mode.