digitalFAQ.com Forums [Archives] - KVCD: Does CQ value effect the encoding time?

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

- Video Encoding and Conversion (http://www.digitalfaq.com/archives/encode/)

- - KVCD: Does CQ value effect the encoding time? (http://www.digitalfaq.com/archives/encode/3416-kvcd-cq-value.html)

I tried mpeg2dec 1,2 and 3 and got the messsage not a valid 2.5 filter, no I didn't burn the m1v file off to cd didn't think when I saw the quality. I tried using kwags latest script but it comes up saying unfilter is not a valid 1.0 filter.

Quote:

Originally Posted by andybno1

download unfilter from here,you got the 2 versions...2.x and 2.5x
http://www.kvcd.net/forum/viewtopic....3c32b6faf06339

and...
maybe we have to wait a few days cos avisynth 2.51 is beta
and got some problems using this version.
:(

cheers jorel, downloaded it and it works fine with 2.08 managed to get it going.

after thousands tests:
a hint for filters, size, CQ and time to encode:

1- every filter that you put in the script(how many)

2- more resize(352x240/288, more than this..etc)

3- more CQ

will got more time to encode.ever!
(the order of the filters in the script change the final time and size)
:!:

Code:

LoadPlugin("C:\fitcd\MPEG2DEC.dll")

LoadPlugin("C:\fitcd\fluxsmooth.dll")

LoadPlugin("C:\fitcd\GripFit_preview.dll")

LoadPlugin("C:\fitcd\STMedianFilter.dll")

LoadPlugin("C:\fitcd\MergeChroma.dll")

LoadPlugin("C:\fitcd\MergeLuma.dll")

LoadPlugin("C:\fitcd\UnFilter.dll")

LoadPlugin("C:\fitcd\TemporalCleanerOld.dll")

LoadPlugin("C:\fitcd\sampler.dll")

LoadPlugin("C:\fitcd\legalclip.dll")



mpeg2source("C:\DVD Movie\Batman Forever.d2v")

LegalClip()

GripCrop( 352, 576 )

GripSize(resizer="BicubicResize")

STMedianFilter(10, 50, 0, 0, 10, 50)

SpaceDust() # Optional - for some "not so clean" DVDs.

unfilter(50,50)

temporalsmoother(1,2)

mergechroma(blur(1.58))

mergeluma(blur(0.3))

# Convolution3d or DCTFilter for extreme worse cases.

GripBorders()

LegalClip()

#sampler()

there is my script, what is the best order for the filters??

:wink:

it's ok!

a question:
do you really need spacedust in the script
to encode this movie?...is not from a dvd source?
....i think that you can remove this line:

SpaceDust() # Optional - for some "not so clean" DVDs
or use
#SpaceDust() # Optional - for some "not so clean" DVDs

and your encode will be faster.

:!:

it is a dvd source, I'm doing a copy for a mate I only put spacedust in to try and get rid of the fuzzyness oin the picture.

Personally, I would probably try not to use both STMedianFilter AND SpaceDust because, at least to me, it just smoothes the picture a bit too much using both of them (making it blurry). Also, Kwag recently dropped STMedianFilter from his recommended Optimal Script because some people were saying they thought it decreased the visual quality too much. In my opinion however, I think it should be left there and used by people only when necessary (on longer movies or when you need to raise your CQ level a bit more). Using ANY kind of filters usually diminishes the quality to some degree...but it's the price you hafta pay sometimes to fit a long movie onto one disc.

Again, this is only my opinion, but to me the best usage of kwag's optimal script (for clean dvd sources) would be something like this (keep in mind that with the unfilter over-sharpening and subsequent mergeluma/chroma blurring methods used by kwag's current script, you can get away with using a pretty low CQ level, possibly in the range of mid-to-high-50's in my experience, and it should still look pretty good on standard TVs, because the noise looks more natural than it used to in his older recommended scripts that used lots of filtering):

Code:

mpeg2source("C:\movie.d2v") 

LegalClip() 

GripCrop(width=544, height=480, overscan=2) #use x3 if/when possible

GripSize(resizer="BicubicResize") 

#STMedianFilter(10, 50, 0, 0, 10, 50) #1st choice for more compression

#SpaceDust() #Optional on "unclean" DVDs. try not to use w/ STMedian

unfilter(50,50) 

temporalsmoother(1,2) 

mergechroma(blur(1.58)) 

mergeluma(blur(0.3)) #use lower values for a less blurry image

# Convolution3d or DCTFilter for extreme worse cases. 

GripBorders() 

LegalClip() 

#sampler()

Using the above script as is, trying to use as few extra filters as possible, looks the best to me (but this is usually only possible with short movies). If I'm encoding a longer movie and need more compression so I can hit my target CQ level during filesize prediction, then the first filter I would turn to (other people's opinion may differ) is STMedianFilter. I think the picture looks slightly better with STMedianFilter than it does using a similar smoother like SpaceDust to gain compression (like I said, I try to only use 1 or the other of those 2 filters because using both just doesn't look as pleasing to me...too much smoothing...but if I'm really desperate, I'll use both).

Anyways, hope you find something useful in all of that. You can always give it a try and see if you agree or not :wink: . Let us all know how everything works out for ya now that you're using prediction and have figured out your avisynth woes and are well on your way! :)

-d&c

p.s. if your friend's dvd player will handle the x3 template resolution, you should try to use that one..it will look much better than 352x576. It will give you the best compromise between picture quality and the amount of minutes you can fit onto one disc.

again i wrote dazed&confused:

you're not confused or dazed my friend!
:lol:

you're correct here too:
"(keep in mind that with the unfilter over-sharpening and subsequent mergeluma/chroma blurring methods used by kwag's current script,
......"
and:
"Using ANY kind of filters usually diminishes the quality to some degree..."

great observations.
:!:

the script there is kwags latest script from the script section, I just activated spacedust to try and sort the fuzz out