digitalFAQ.com Forums [Archives] - Avisynth: CQ Matic static script?

digitalFAQ.com Forums [Archives] (http://www.digitalfaq.com/archives/)

- Avisynth Scripting (http://www.digitalfaq.com/archives/avisynth/)

- - Avisynth: CQ Matic static script? (http://www.digitalfaq.com/archives/avisynth/6329-avisynth-cq-matic.html)

Quote:

Originally Posted by bigggt

@ Dialhot

could you please post your script here that you use for dvd rips.there was too many different versions in this thread so i'm not sure which one you use now.

Thanx

The one I use is not, it never was, and probably never be MA.

So I can't answer to your question the way you meant it.

I mean can you post the script that you use instead of the MA on cleand dvd material i just want to compare for myself

Quote:

Originally Posted by bigggt

I mean can you post the script that you use instead of the MA on cleand dvd material i just want to compare for myself

For KVCD I use the script I did for avi->kvcd.
For KDVD I use this one :

Code:

Mpeg2Source("PATH\NAME.d2v")

GripCrop(704, 576, overscan=0, source_anamorphic=true)

GripSize(resizer="LanczosResize")

FluxSmooth()

TemporalCleaner(ythresh=5,cthresh=7)

Undot()

DCTFilter(1,1,1,1,1,1,0.5,0)

GripBorders()

LetterBox(16, 16, 16, 16)

Note that in reality I do not use any resizing (just do a Crop/Addborder manually, withotu using Gripfit) so perhaps you should try Bicubicresize insteed of LanczosResize.

Thanx Phil

Let me just get this right so i understand

even when you do a dvd rip and want it to go to kvcd you use the same script that you use when you do an avi.

so dvd and avi -

Code:

AviSource("PATH\NAME.avi",false or Mpeg2Source"C:\PATH\NAME.d2v")

BlindPP(cpu=4) 

Blockbuster(method="noise",detail_min=1,detail_max=3,variance=0.1,seed=1) 

Convolution3D(1, 6, 12, 6, 8, 2.8, 0) 

GripCrop(WIDTH, HEIGHT, overscan=1, source_anamorphic=false) 

GripSize(resizer="LanczosResize") 

Undot() 

TemporalSoften(2,7,7,3,2) 

DCTFilter(1,1,1,1,1,1,0.5,0) 

#Blockbuster(method="noise",detail_min=1,detail_max=10,variance=0.3,seed=5623) 

GripBorders()

i thought that you only used that script for avi :D

The only thing you can remove in the script concerning clean DVD is the "blindPP" line.
This line is there to remove the DCT blocks that are in the source and is not very usefull for DVD.
(However, the blockbuster and convolution3D that are after do the same work ;-))

Quote:

Originally Posted by Dialhot

If this is wrong, could you just post the script you use for DVD material and for AVI material?

No need, what you wrote is correct.

Quote:

You mentioned that Convolution3d and Blockbuster both do the same thing.. which is better in your opinion?

Let me clear this : BlindPP(CPU=4) removes the DCT blocks by a method called "Deblocking" (you can use also "Deringing" that is stronger - too much according to me - with "cpu=6").

Blocks are also removed by the PAIR of lines "Blockbuster(noise) + C3D". These two lines must be put or removed together and can't be splited. In this case, the blocks are removed by a "spay some noise on the blocks to break their borders then remove the noise" proccesing.

For DVD, blindPP=4 does not do a lot of things are blocks are far less visible than on avi and so, the deblocking method is inefficient. But the "blockbuster+C3D" one still does something, that's why I rather let them here and remove the blindPP line.

Quote:

Also,what of MergeChroma and MergeLuma? Do you not use these at all? Audioslave mentioned that he was getting higher compression using these..?

avis are already too much blurred that"s why I do not use that in the script. I tried once to add it for a DVD but the image seems to soft and I never did other tests. But you can try.

Quote:

From what I've read in this thread so far, your scripts are getting better results than the Optimal 2.5 script, yes?

Far better in my eyes, that's why I do not use MA. But all is matter of taste (and test).

Quote:

Also, what are you all referring to when you say MA?

Motion adaptative script. The one with "nf=YDifferenceToNext()". The parameters of the filters on this line aren't constant but adapted to what will be in the next frame (that's the purpose of the YDifferenceToNext".

See it there :

http://www.kvcd.net/forum/viewtopic.php?t=3483
In other words, the scripts is adapted to the motion of the scene. The opposite is a "static" script (all filters work the same way on all frames).

And for sure, one more time, I edited post insteed of quoting it :-(

So The previous post is mine, and it answer to a post of gonzopdx.

(Kwag, can't you do something to separate "quote" and "reply" button in moderator panel ??? Or at least keep the quote button to its correct place !)

Quote:

Originally Posted by Dialhot

I have done more encodes with a static script rather than with the MA one,
at least recently.

Also, today I tried an encode with the classic MA script and the time left was increasing rather than decreasing! I decided to switch to the static one again... at least it will finish!

Quote:

Originally Posted by cweb

Also, today I tried an encode with the classic MA script and the time left was increasing rather than decreasing! I decided to switch to the static one again... at least it will finish!

Yes, but if you have an action movie, that's where the MA script will pay off, and you'll get excelent quality and far less file size than any static script, of course at the cost of encoding time ;)

-kwag

Quote:

Originally Posted by kwag

Quote:

Originally Posted by cweb

Also, today I tried an encode with the classic MA script and the time left was increasing rather than decreasing! I decided to switch to the static one again... at least it will finish!

Sorry kwag, so perhaps that was the reason. My sources today weren't action movies - more like comedy etc (e.g. 'my hero' involves a superhero but there isn't much action in most episodes, it's mostly a comedy).

But the slowness was not a 3% extra like someone mentioned... it was much more! And when TMPGENC was idle (while some other process was active), it would slow down much more, indicating several hours left (always increasing of course!). Odd behaviour...

Thanks kwag for the tip then - you helped me see why this was happening.