Quantcast KVCD: Nomosmooth() and File Size Prediction? - digitalFAQ.com Forums [Archives]
  #1  
11-15-2002, 06:02 PM
Daagar Daagar is offline
Free Member
 
Join Date: May 2002
Posts: 158
Thanks: 0
Thanked 0 Times in 0 Posts
Kwag and others,

I tried your latest settings (ie., NoMoSmooth() + blockbuster()), and the result seemed to be _much_ better than when I tried blockbuster() alone (with your older, 'stronger' settings). However, I noticed that the file size predicition didn't do quite as good of a job as normal...

I use DVD2SVCD, and typically have the prediction set at 0.95.

No filters: the file was 780MB
Blockbuster alone: The file was 767MB
Blockbuster+NoMoSmooth: The file was 735MB

Granted, being under is better than being over, but that also means I'm 'wasting' a decent chunk of space. Should I be adjusting the 0.95 to something else? Do I need to go higher or lower in this situation?

Just curious what everyone else is seeing.

PS - Adding NoMoSmooth() helped a lot for the CQ values as well. With blockbuster alone, CQ_VBR was dropped to 11.5 after prediction. With NoMoSmooth() added, it went back to 13.8. Modest, but shows that it is helping with compression quite a bit!
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
  #2  
11-15-2002, 09:00 PM
Daagar Daagar is offline
Free Member
 
Join Date: May 2002
Posts: 158
Thanks: 0
Thanked 0 Times in 0 Posts
Okay, duh... the 0.95 is just a scaling factor so obviously I should put it back closer to the original 0.99 value. That'll bump the CQ from 13.92 up to 14.65... yay!
Reply With Quote
  #3  
11-15-2002, 10:47 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
Yeah, the filters do make a drastic difference with the formula!. But good'ol D2S made that "factor" variable for a good reason

-kwag
Reply With Quote
  #4  
11-15-2002, 10:50 PM
snowmoon snowmoon is offline
Free Member
 
Join Date: Nov 2002
Posts: 13
Thanks: 0
Thanked 0 Times in 0 Posts
Send a message via ICQ to snowmoon Send a message via AIM to snowmoon
I've tried NoMoSmooth and have gone back to PeachSmoother.

They both do approximatly the same thing ( combines temporal / spatial smoothing based on motion mask ) but peach appears to be at least twice as fast.

Don't get me wrong I think the are both excellent filters, I just don't have cycles to burn on a PIII700.
Reply With Quote
  #5  
11-17-2002, 03:35 PM
VILLA21 VILLA21 is offline
Free Member
 
Join Date: Oct 2002
Location: Greece
Posts: 63
Thanks: 0
Thanked 0 Times in 0 Posts
The movie: "La Pianiste", PAL, 16:9, ~125min

1st attempt:
---AVS Begin---
LoadPlugin("C:\PROGRA~1\DVD2SVCD\MPEG2Dec\mpeg2dec .dll")
LoadPlugin("C:\PROGRA~1\DVD2SVCD\AVISYN~1.DLL")
mpeg2source("E:\Video\Movie\DVD2AV~1.D2V")
BilinearResize(352,216)
AddBorders(0,36,0,36)
LoadPlugin("C:\Program Files\DVD2SVCD\Avisynth\NoMoSmooth.dll")
NoMoSmooth()
LoadPlugin("C:\Program Files\DVD2SVCD\Avisynth\Blockbuster.dll")
Blockbuster( method="noise", detail_min=1, detail_max=10, lv=1 )
Blockbuster( method="sharpen", detail_min=20, detail_max=90, strength=20 )
AvisynthSubtitler("E:\Video\Movie\Subs\","permsubs .txt")
----AVS End----

Pred.Factor=0.95
Video Encoding time: 3h 30min
File size: 780MB
Comments: Very poor quality , too many blocks

2nd attempt:
---AVS Begin---
LoadPlugin("C:\PROGRA~1\DVD2SVCD\MPEG2Dec\mpeg2dec .dll")
LoadPlugin("C:\PROGRA~1\DVD2SVCD\AVISYN~1.DLL")
mpeg2source("E:\Video\Movie\DVD2AV~1.D2V")
LoadPlugin("C:\Program Files\DVD2SVCD\Avisynth\UnFilter.dll")
Unfilter(20,20)
BilinearResize(352,216)
AddBorders(0,36,0,36)
LoadPlugin("C:\Program Files\DVD2SVCD\MPEG2Dec\MPEG2DEC2.dll")
TemporalSmoother(2,1)
AvisynthSubtitler("E:\Video\Movie\Subs\","permsubs .txt")
----AVS End----

Pred.Factor=0.97
Video Encoding time: 2h 30min
File Size: 815MB
Comments: Amazing quality! blockness

Conclusion:
Using Unfilter and TemporalSmoother gave me excellent results, fast encoding and increasing the Pred.Factor to 0.97 the final image almost filled the CDR capacity.
__________________
The aRt oF FaLLing ApaRt
Reply With Quote
  #6  
11-17-2002, 04:55 PM
SansGrip SansGrip is offline
Free Member
 
Join Date: Nov 2002
Location: Ontario, Canada
Posts: 1,135
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by snowmoon
They both do approximatly the same thing ( combines temporal / spatial smoothing based on motion mask ) but peach appears to be at least twice as fast.
NoMoSmooth is as-yet unoptimized. If I get good reports from its output then I'll work at improving it, but I fear it won't get much faster than 0.1 until I learn assembler (I'm hopefully ordering a couple of books today...)

Meanwhile, you could try FluxSmooth. It should be considerably faster.
Reply With Quote
Reply




Similar Threads
Thread Thread Starter Forum Replies Last Post
KVCD: Wrong sample size in file prediction? Big_Peb Video Encoding and Conversion 11 05-24-2003 03:09 PM
File size prediction program ? genK Avisynth Scripting 3 05-24-2003 07:21 AM
File Size Prediction For PAL... Jellygoose Avisynth Scripting 8 01-01-2003 09:18 PM
KVCD: File Size Prediction with FitCD? Jellygoose Video Encoding and Conversion 3 12-17-2002 10:07 AM
KVCD: Problem With File size prediction twobit326 Video Encoding and Conversion 10 12-02-2002 10:17 PM

Thread Tools



 
All times are GMT -5. The time now is 12:53 AM  —  vBulletin © Jelsoft Enterprises Ltd