Quantcast Interesting Avisynth Resizing Thread at Doom9.org - Page 8 - digitalFAQ.com Forums [Archives]
  #141  
03-14-2004, 01:09 PM
kwag kwag is offline
Free Member
 
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by FuPP
I'm definitevely stupid !
No FuPP, you're very bright
If you were stupid, you would have never designed such a script

However, we all bear the
and the


icons, every once in a while.
But just for laughs

-kwag
Reply With Quote
Someday, 12:01 PM
admin's Avatar
Site Staff / Ad Manager
 
Join Date: Dec 2002
Posts: 42
Thanks: ∞
Thanked 42 Times in 42 Posts
  #142  
03-14-2004, 02:42 PM
jorel jorel is offline
Invalid Email / Banned / Spammer
 
Join Date: Aug 2002
Location: Brasil - MG - third stone from the sun
Posts: 5,570
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by kwag
Quote:
Originally Posted by FuPP
I'm definitevely stupid !
No FuPP, you're very bright
If you were stupid, you would have never designed such a script

-kwag
yes FuPP, you did (and will do more) a wonderful work!


and FuPP
wrote too:
"oh, I was not talking about the chroma problem ! I was talking about the overall quality."
ok my friend, now i "see" what you mean!

and the tests using Deen like you posted?
somebody test it?
Reply With Quote
  #143  
03-14-2004, 07:37 PM
FuPP FuPP is offline
Free Member
 
Join Date: Mar 2004
Posts: 92
Thanks: 0
Thanked 0 Times in 0 Posts
ok, not completely stupid if you want

Maybe some of you would like to try HybridFuPP 0.85a... far better chroma smoothing, so colors are a lot less washed in dark scenes.

here it is : http://fupp.chez.tiscali.fr/HybridFuPP/HybridFuPP.avsi

Regards,
FuPP
Reply With Quote
  #144  
03-14-2004, 07:45 PM
audioslave audioslave is offline
Free Member
 
Join Date: Mar 2003
Location: Sweden
Posts: 725
Thanks: 0
Thanked 0 Times in 0 Posts
You bet! Will download the VOB Kwag posted and test the new HybridFuPP 0.85a.
__________________
AudioSlave
Reply With Quote
  #145  
03-14-2004, 08:19 PM
audioslave audioslave is offline
Free Member
 
Join Date: Mar 2003
Location: Sweden
Posts: 725
Thanks: 0
Thanked 0 Times in 0 Posts
Yes! The "washed out" picture is gone. Great work, man! It's looking better than ever.
__________________
AudioSlave
Reply With Quote
  #146  
03-15-2004, 12:19 AM
kwag kwag is offline
Free Member
 
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
Well, I really have to say this:
FuPP script is very good, but the time to encode a movie, is almost twice that of the MA script.
I just did a test encode to compare the quality on my Samsung 32" HDTV, and to tell you the truth, there's absolutely "zero" visual difference
I think there's a limit to how much filtering there should be done on source material.
I'll certainly use FuPP for some captures, but not for any DVD sources

-kwag
Reply With Quote
  #147  
03-15-2004, 08:55 AM
audioslave audioslave is offline
Free Member
 
Join Date: Mar 2003
Location: Sweden
Posts: 725
Thanks: 0
Thanked 0 Times in 0 Posts
Here's the reason why I'd rather use HybridFuPP to encode my movies:

~ Script Comparison ~
http://hea.port5.com/scriptcomp.htm

It's a scene (from Underworld) when most of the background is moving (subway train) but the actress (Kate Beckinsale, mmm... ) is in focus. With the MA script everything gets blurred but with HybridFuPP only the background is heavily filtered - as you can see from the screenshots. I'm sure there are better examples but this was a scene that came to mind.
__________________
AudioSlave
Reply With Quote
  #148  
03-15-2004, 08:59 AM
Dialhot Dialhot is offline
Free Member
 
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
FYI : Underworld DVD isn't interlaced.
Reply With Quote
  #149  
03-15-2004, 05:59 PM
audioslave audioslave is offline
Free Member
 
Join Date: Mar 2003
Location: Sweden
Posts: 725
Thanks: 0
Thanked 0 Times in 0 Posts
@Dialhot
Oops, you're right. The FieldDeinterlace line was already in my scripts and I forgot to comment them out. Anyways, the point was the blurring issue.
__________________
AudioSlave
Reply With Quote
  #150  
03-17-2004, 03:09 AM
bman bman is offline
Free Member
 
Join Date: Apr 2002
Posts: 356
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by FuPP
ok, not completely stupid if you want

Maybe some of you would like to try HybridFuPP 0.85a... far better chroma smoothing, so colors are a lot less washed in dark scenes.

here it is : http://fupp.chez.tiscali.fr/HybridFuPP/HybridFuPP.avsi

Regards,
FuPP
Can't download
Site is down ????
bman
Reply With Quote
  #151  
03-17-2004, 09:52 AM
bigggt bigggt is offline
Free Member
 
Join Date: Mar 2003
Location: IamCanadian
Posts: 848
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by bman
Can't download
Site is down ????
Did you right click and save as

thats the only wat to do it
Reply With Quote
  #152  
03-17-2004, 10:02 AM
bman bman is offline
Free Member
 
Join Date: Apr 2002
Posts: 356
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by bigggt
Quote:
Originally Posted by bman
Can't download
Site is down ????
Did you right click and save as

thats the only wat to do it
I did so but couldn't downloaded file !??
This time it worked for me .
Going to try it
Thank's for help
bman
Reply With Quote
  #153  
03-17-2004, 04:32 PM
audioslave audioslave is offline
Free Member
 
Join Date: Mar 2003
Location: Sweden
Posts: 725
Thanks: 0
Thanked 0 Times in 0 Posts
Good evening to you all!
This afternoon I made a comparison of two different resizing methods in the HybridFuPP (0.85a) script.

#1
Quote:
Mpeg2Source("D:\DVD Rip\Master\Master.d2v")
Crop(10, 78, 700, 422)
HybridFuPP(448, 304, Preset="High")
AddBorders(16, 136, 16, 136)
This script uses FuPP's (the creator of the excellent script ) resizing method with the Crop() line.

#2
Quote:
Mpeg2Source("D:\DVD Rip\Master\Master.d2v")
BicubicResize(448, 304, 0, 0.6, 9, 78, 700, 422)
HybridFuPP(448, 304, Preset="High")
AddBorders(16, 136, 16, 136)
This cript uses Kwag's (the creator of the excellent MA script ) method of resizing as posted earlier in this topic.

Results
Latest MA script (Nov. 21, 2003): 14 538 kB
HybridFuPP 0.85a with Crop(): 14 661 kB
HybridFuPP 0.85a with BicubicResize(): 14 821 kB

Results are based on a 02:38 sample from "Master & Commander".

This means that HybridFuPP produces files that are very close in size to those of the MA script (when using the Crop() line in the HybridFuPP script). If we're using the BicubicResize line instead we probably screw up the internal resizing algorithms within the HybridFuPP function. In other words the exact same function that FuPP made to get as small filesizes as possible. Please correct me if I'm wrong.
The MA script is of course a lot faster but if you have the time to spare I really think it's worth the wait. That's my opinion.
__________________
AudioSlave
Reply With Quote
  #154  
03-17-2004, 04:44 PM
incredible incredible is offline
Free Member
 
Join Date: May 2003
Location: Germany
Posts: 3,189
Thanks: 0
Thanked 0 Times in 0 Posts
Send a message via ICQ to incredible
I have to agree with KWAG that a little unsharper pic in case of MA you wont notice on a regular TV set. And shurely if you get rid of that second blurring process done by MergeLuma(Blur(..))! as that gives a little over all sharper picture if deleted when using MA.

THE ONLY reason for me to have a close look at HybridFupp() is that it doesnt got that "static foreground object while fast luma moving in the background" - issue! Cause that reults also in a blurry static foreground object as MA does filter the whole frame based on NF where HybFupp does filter only the moving parts.
Reply With Quote
  #155  
03-17-2004, 04:48 PM
audioslave audioslave is offline
Free Member
 
Join Date: Mar 2003
Location: Sweden
Posts: 725
Thanks: 0
Thanked 0 Times in 0 Posts
Incredible wrote:
Quote:
THE ONLY reason for me to have a close look at HybridFupp() is that it doesnt got that "static foreground object while fast luma moving in the background" - issue! Cause that reults also in a blurry static foreground object as MA does filter the whole frame based on NF where HybFupp does filter only the moving parts.
My point exactly. And I agree with both Kwag and yourself about not being able to see the difference with MA's blurrier look on a TV-screen. BUT, it nice to have only the moving parts of a frame blurred...
__________________
AudioSlave
Reply With Quote
  #156  
03-17-2004, 11:45 PM
fabrice fabrice is offline
Free Member
 
Join Date: Mar 2003
Location: Madrid-Spain
Posts: 515
Thanks: 0
Thanked 0 Times in 0 Posts
Hi,

So, why not trying to play with maskedMerge, to get a similar effect (blurring the moving part), and replace the MA part? (i already post that before, but I know it's a bit OT...).

We just could begin a new thread for the maskedmerge version of the MA version...

CU
Fabrice
Reply With Quote
  #157  
03-18-2004, 12:00 AM
kwag kwag is offline
Free Member
 
Join Date: Apr 2002
Location: Puerto Rico, USA
Posts: 13,537
Thanks: 0
Thanked 0 Times in 0 Posts
Hi Fabrice,

Did you get my PM this morning about that

-kwag
Reply With Quote
  #158  
03-18-2004, 12:19 AM
fabrice fabrice is offline
Free Member
 
Join Date: Mar 2003
Location: Madrid-Spain
Posts: 515
Thanks: 0
Thanked 0 Times in 0 Posts
Hi karl,

just answer right now. Don't know why it doesn't popup me today...

CU
Fabrice
Reply With Quote
  #159  
03-18-2004, 04:12 AM
incredible incredible is offline
Free Member
 
Join Date: May 2003
Location: Germany
Posts: 3,189
Thanks: 0
Thanked 0 Times in 0 Posts
Send a message via ICQ to incredible
I also gave that hint to kWAG some time ago, but there is the compromise between the "simpler" MA compared to a more complex masking routine.

Masking routines do need much more time than a simple "whole" frame filter.

BUT what we can try is to implementate a maskingroutine which only works/performs on the y (luma) channel.

I did some tests this week on noisy captures where I just let run some routines only on luma and other routines only on the chroma channel less static grain on moving walls etc. and on the other hand much more faster as the filter do perform more effective according to the Y/C channels.

"Something" like this...

SourceImport("xxxxxxxxxx.avi")
Mergeluma(filter(xxxxx))
Mergechroma(filter(xxxxx))

... just an example....

So that also "could" speed up HybFupp if we let apply for example only the sharpening unfilter on the luma channel, also the spatial filtering of plain surfaces could be done on the luma only.
only! Softresizings on chroma, but according to HybFupps internal masking ... soft/mid/sharp resizing on Luma only.

For explaining:

SoftAction=Bicubicresize(xx,yy,0,0).blur(x)
SharpAction=Lanczosresize(xx,yy).TemporalSoften(xx x)
MergeLuma(SharpAction)
MergeChroma(SoftAction)

There are filters which do have ettings separately on luma and chroma, but I noticed that its a bit diff. case if really some filters do work ONLY on that specific channel, separated by the MergeXXXX command
Reply With Quote
  #160  
03-18-2004, 04:36 AM
Dialhot Dialhot is offline
Free Member
 
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by fabrice
just answer right now. Don't know why it doesn't popup me today...
You moved to mozilla or firefox and didn't there is an anti popup feature in them ?
Reply With Quote
Reply




Similar Threads
Thread Thread Starter Forum Replies Last Post
Avisynth: Interesting results with YlevelsS supermule Avisynth Scripting 2 08-06-2006 11:59 PM
Avisynth: Interesting ASharp phenomenon... audioslave Avisynth Scripting 12 10-23-2003 06:36 AM
Avisynth: Latest Scripts Thread EvilFactor Avisynth Scripting 3 06-16-2003 11:53 PM
Avisynth: Resizing? CheronAph Avisynth Scripting 5 03-01-2003 08:41 PM
AviSynth 2.05 filter Compare for debugging is interesting... black prince Avisynth Scripting 0 09-25-2002 08:57 PM

Thread Tools



 
All times are GMT -5. The time now is 11:09 AM  —  vBulletin Jelsoft Enterprises Ltd