Avisynth: Dctfilter a Winner?

Jellygoose · #1 06-29-2003, 07:24 AM

Hi kwag and all others...

Have you already tried DCTFilter with the current optimal script? Filesize dropped for me by 0.5 MB on a Full Sampler

I looked at different frames of both scripts, and there's no detail loss. The thing that I noticed was that with DCTFilter the backgrounds got slightly darker. Well I used this line at the very end of the script:

DctFilter(1,1,1,1,1,.5,.5,0)

Try it out for yourself please...

jorel · #2 06-29-2003, 08:31 AM

Jell please,
more details!

where you use DCTFilter in the script?
begining,middle or end?
is built in avisynth252 ?

thanks in advance!

Jellygoose · #3 06-29-2003, 11:02 AM

Use it at the VERY end of the script. It's not built-in, you can download it where you can download all the filters

www.avisynth.org/~warpenterprises

jorel · #4 06-29-2003, 11:05 AM

thanks Jell,
i'm trying it now.

jorel · #5 06-29-2003, 11:49 AM

hey Jell,
i got 1 more point in CQ prediction.

i don't compare the image quality,
i'm trying to find high CQs values.

Jellygoose · #6 06-29-2003, 04:54 PM

Quote:

Originally Posted by jorel

hey Jell,
i got 1 more point in CQ prediction.

i don't compare the image quality,
i'm trying to find high CQs values.

Well I still think that high CQ values don't guarantee good quality always. Please if someone would compare the image quality of the script with and without DCTFilter, it would be really appreciated!

I cannot see that it does any big harm to the image quality, so if someone can confirm this I'm gonna PM kwag about it!

vhelp · #7 06-29-2003, 05:03 PM

@ Jellygoose..

I'll give it a try. I"m hvaing issues, but I remember a piece of source
that I can try it out on, and let you know.. before my Dial-up runs out..
I'm cutting it close the the 10th hour hehe.. I'm rushing..

I'm assumging that by "end of script" you mean after the
function FMIN() ???

I'll try both in any case.
-vhelp

Jellygoose · #8 06-29-2003, 05:06 PM

No, don't put it after the functions. Put it right after the second Limiter() line, that where it belongs!

I don't know if it matters though to put it after the functions. Just trying to sound wise

vhelp · #9 06-29-2003, 05:09 PM

@ Jellygoose..

Ok, thanks.. will do.
-vhelp

vhelp · #10 06-29-2003, 05:32 PM

@ Jelly..

Sorry for the delay.. I had issues.
I realize this is a small sample test, but I'm cutting it close to my
dissconnet - sorry.

-vhelp

In short.. here goes what I have w/ latest script.
* 20 sec clip - 1,490kb - w/out ## DctFilter(1,1,1,1,1,.5,.5,0)
* 20 sec clip - 1,499kb - with DctFilter(1,1,1,1,1,.5,.5,0)

EDIT:
9k loss, based on script below, and using DctFilter(1,1,1,1,1,.5,.5,0) filter at bottom:
------------------------------------------------------------

Code:

# ------------
 undot() 
 Limiter() 
 asharp(2, 4) 
 GripCrop(Your_GripCrop_Parameters_Here) 
 GripSize(resizer="BicubicResize") 
 STMedianFilter(8, 32, 0, 0 ) 
 MergeChroma(blur(MaxTreshold)) 
 MergeLuma(blur(0.2)) 
# 
# 

## Linear Motion Adaptive Filtering ## 
# 
# ( Portions from AviSynth's manual ) - This will apply temporalsoften to 
# very static scenes, and apply variable blur on moving scenes. 
# We also assign a variable - and this is why a line break is inserted: 

ScriptClip("nf = YDifferenceToNext()"+chr(13)+ "nf > 2 ? \ 
unfilter( -(fmin(round((nf/0.5)), 100)), -(fmin(round((nf/0.5)), 100)) ) : \ 
TemporalSoften(4, round(2/nf)*2 , round(2/nf)*50 ,1,2) ") 
# ------------end of latest script change

 Limiter() 
 DctFilter(1,1,1,1,1,.5,.5,0) 

 function fmin(float f1, float f2) { 
 return (f1<f2) ? f1 : f2 
 }

vhelp · #11 06-29-2003, 05:37 PM

@ Jellygoose..

Sorry, but the above is what I concluded from the 20 second clip.

I ment to say it grow 9k larger when using the filter

The movie was "Dogma" - the train ride.. where they are tasking at the
table.. an 1 Hour and 5 Min into the movie.

-vhelp

jorel · #12 06-29-2003, 06:09 PM

Quote:

Originally Posted by Jellygoose

Quote:

Originally Posted by jorel

hey Jell,
i got 1 more point in CQ prediction.

i don't compare the image quality,
i'm trying to find high CQs values.

Well I still think that high CQ values don't guarantee good quality always.

ok Jell, but it means that we have to
encrease the CQ to got the full space cos
we got free space with less CQ.
the prediction will give us some errors,don't?

Dialhot · #13 06-30-2003, 04:20 AM

Quote:

Originally Posted by Jellygoose

DctFilter(1,1,1,1,1,.5,.5,0)

I use this filter since several weeks (either in static or dynamic script) but I jut want to warn you : these setting are a little "too much".

By modifying the 6the parameter, you enter in "frequencies" that the human eye can see.

If you use "ffdshow" as filter for viewing you video, I suggest you to test your values with it : it has a "on the fly" DCT filter (just see in the option of ffdshow) :you can set the values you want and see directly on the video if you can notice a difference or not.

For my part I'm stick to "DCTFilter(1,1,1,1,1,1,0.5,0).