A doubt with OPV in RB

[tom942]

Hello,

I'm testing OPV and CQ since I have read several threads that it's a way to encode since it can yield similar or even better quality than multipass (correct me if I am saying a stupid thing) :).

But my question is related with which matrix to choose in DVD-RB.

I found a thread where manono answers that the bigger the file in OPV, more detail it retains.

So I did the test the way he suggested. I took AVAMAT 6 and 7 knowing that the first one is focused to bitrates of 3000 and over and the second one is focused to 3000 and under. First I did the prepare phase in DVD-RB Pro, then I used RB-OPT with AVAMAT7 and got a Q of 15. Well, I let RB encode some segments and switch to RB-OPT, to change AVAMAT7 to AVAMAT6 and this time I fixed the Q by hand to 15. As manono said, this way I can see that AVAMAT6 retains more detail than AVAMAT7, because size is bigger, but encoding with AVAMAT6 using a Q of 15, it will probably oversize.

So my doubt is when RB-OPT gives me a Q of 17 for AVAMAT6 and a Q of 15 for AVAMAT7, the size of the segments are similar.

So which one retains more detail yet? AVAMAT6 or 7? or in this situation where RB is trying to find a size within the limits (DVD-5) the behavior of both matrices is similar, I mean they retain similar details?.

And if I use other matrix like HVS-GOOD, it gives me a Q of 20, the same question, is it retaining more detail than the above or is it retaining the same that AVAMAT6 with Q 17 and AVAMAT7 with Q 15.

This doubt came to my mind when I read that the lower Q, the better, but at the same time I thought that if I try to reach a Q of 10 or less, maybe I'm throwing away a lot of details. I suppose that if I use, for instance, notch matrix I will have a low Q but how much detail I will retain?.

[~bT~]

maybe u can up some sample of both?

[Sharc]

.
.... So my doubt is when RB-OPT gives me a Q of 17 for AVAMAT6 and a Q of 15 for AVAMAT7, the size of the segments are similar.
Yes, RB-Opt provides the Q to fit the target size for a given matrix.

So which one retains more detail yet? AVAMAT6 or 7? or in this situation where RB is trying to find a size within the limits (DVD-5) the behavior of both matrices is similar, I mean they retain similar details?.
AVAMAT6 should preserve slightly more details but at the same time MAY produce slightly more artefacts due to the higher Q. The differences will be hardly noticeable, though, and the result depends also on the source.

And if I use other matrix like HVS-GOOD, it gives me a Q of 20, the same question, is it retaining more detail than the above or is it retaining the same that AVAMAT6 with Q 17 and AVAMAT7 with Q 15.
Same here. The high bitrate matrix can preserve more details, however it may introduce more artefacts. As long as Q is below say 30 (for CCE) you don't have to worry too much about encoding artefacts.

This doubt came to my mind when I read that the lower Q, the better, but at the same time I thought that if I try to reach a Q of 10 or less, maybe I'm throwing away a lot of details. I suppose that if I use, for instance, notch matrix I will have a low Q but how much detail I will retain?.
What combo of matrix and Q you find best depends much on your eyes and your personal taste, and also on the source material. You may just try. No exact science here.

[tom942]

In summary, if I get, for example, with AVAMAT6 a Q value of 15 or less, and then I say to me: "I want to try other matrix that retains a bit more of details...", say Fox, Criterion...., and they give me a Q of 20-22-24 and I don't see any artifacts, then I am not doing a stupid thing. I'm compressing more but at the same time retaining more detail.

I thought that like DVD-RB or RB-OPT try different Q values to fit the size(DVD-5) the quality at the end will always be the similar or the same independently the matrix and the Q. I mean that AVAMAT6 with Q17, AVAMAT7 with Q15 or HVS-GOOD with Q20, at the end they will give the same quality because in this apps they are not looking for the Q value to get the best quality else to fit in the disc.

Thank you for clarify me this point, Sharc.:)

@bEnGaLi tHuG

It was more a theoric question that a practical case, and what's more, I don't know how to take the samples :)