digitalFAQ.com Forums [Archives]

digitalFAQ.com Forums [Archives] (http://www.digitalfaq.com/archives/)
-   Avisynth Scripting (http://www.digitalfaq.com/archives/avisynth/)
-   -   Sizes with scripts for avisynth 2.08 vs avisynth 2.52 ? (http://www.digitalfaq.com/archives/avisynth/4438-sizes-scripts-avisynth.html)

jorel 07-10-2003 10:42 AM

Sizes with scripts for avisynth 2.08 vs avisynth 2.52 ?
 
someone saw this :?:

same source (.d2v),33 seconds sample with 1001 frames:

with best script with avisynth252
:arrow: 2,54mb (2.611kb)

with best script with avisynth208
:arrow: 2,38mb (2.448kb)

using the script with avisynth252 is faster with more final size.
using the script with avisynth208 is slow with less final size.
8O
:?:
any comment?

Jellygoose 07-10-2003 10:43 AM

8O
Can you post a frame by frame comparison please? (screenshots)

jorel 07-10-2003 10:46 AM

Quote:

Originally Posted by Jellygoose
8O
Can you post a frame by frame comparison please? (screenshots)

i don't have space to post it Jell,
i can send by mail.


ps:
i have space(50mb) but i'm stup(you all know)
and i don't know how to use the sapce to put my files...
:oops:

ovg64 07-10-2003 10:52 AM

Re: sizes with scripts for avisynth208 vs avisynth252
 
Quote:

Originally Posted by jorel
someone saw this :?:

same source (.d2v),33 seconds sample with 1001 frames:

with best script with avisynth252
:arrow: 2,54mb (2.611kb)

with best script with avisynth208
:arrow: 2,38mb (2.448kb)

using the script with avisynth252 is faster with more final size.
using the script with avisynth208 is slow with less final size.
8O
:?:
any comment?

I would say thats the way it suppose to be Jorge'
Faster= Bigger file size
Slow= Smaller file size
thats y u get better Quality even with low CQ :)

jorel 07-10-2003 11:00 AM

maybe yes Osvaldo,
in full movie with 208 the CQ can encrease to got
the same final size that 252 will give.
i only did this samples twice using CQ60 and got this results.
the quality in 252 is a little more sharp but give a few more blocks...
but the 2 the scripts are cool :!:

i don't did but maybe is a good idea;
encode the same sample with each version of avisynth using only
resize to see that the final size is caused for the script
or for the differents versions of avisynth.
:!:

ovg64 07-10-2003 11:16 AM

Hey jo, maybe you pulsing problem is whats causeing the blocks, cause i sure dont see any blocks in my encodes even when i lower the max bitrate to 2300 to gain more CQ. like Kwag says this MA Script is bullet proof. :mrgreen:

jorel 07-10-2003 11:21 AM

My "breathing" problems is caused for GOP 30 cos
if i adjust the Gop to 24 or 18,the sample "pulse" less.
but what can i do?
my souce is 29,97 interlaced!
than i'm lost with this problem,don't know what to do.
the blocks and the ugly "dark green" remains.
:(

kwag 07-10-2003 11:44 AM

Hi jorel,

Have you tried a regular encode without any filters and using one of the standard TMPEG templates :idea:
If the problem is still present, then there's nothing you can do about it :!:

-kwag

jorel 07-10-2003 11:59 AM

i think you are right as always Kwag,
i did samples using only resize and the "breathing" problem
remains in avisynth208 and 252.
the incredible is that the vobs are very clean.
you and more friends are trying to help me with this problem and
in the end of tons tests,i can be right that:
is not the version of tmpgenc,
is not the scripts,
is not the avisynth version.

and if the source is 29,27 interlaced ntsc,i can't use forced film,right?
or can i encode with GOP 18 or 24?
give me some hints or i delete this vobs cos
with the horrible result ,i can't encode and burn it :!:

some more hints or is the end of the road for this movie?
:?:

thanks in advance.
:)

Dialhot 07-10-2003 12:25 PM

[quote="jorel"]
and if the source is 29,27 interlaced ntsc[quote]

Jorel : when you test tmpgenc without any filters : how did you deinterlace you source ?

Can't the problem come from there ?

Dialhot 07-10-2003 12:28 PM

Re: sizes with scripts for avisynth208 vs avisynth252
 
Quote:

Originally Posted by jorel
someone saw this :?:

same source (.d2v),33 seconds sample with 1001 frames:

with best script with avisynth252
:arrow: 2,54mb (2.611kb)

with best script with avisynth208
:arrow: 2,38mb (2.448kb)

using the script with avisynth252 is faster with more final size.
using the script with avisynth208 is slow with less final size.
8O
:?:
any comment?

That quite normal : the last optimal script use heavy sharpening (asharp(2,4)) and low bluring (blur(0.1)) where old script was a little lighter (blur(0.2) if I remember well ?).

Moreover, the new script use "Motion dectect" where old use "Hight precision", and that make a big difference on the size (and speed).

FYI, I droped temporalsoften (and use temporalcleaner) because the size is smaller with it. And sometimes it's WAY smaller !

jorel 07-10-2003 12:33 PM

[quote="Dialhot"][quote="jorel"]
and if the source is 29,27 interlaced ntsc
Quote:


Jorel : when you test tmpgenc without any filters : how did you deinterlace you source ?

Can't the problem come from there ?
i don't do anything,no deintelace...only for test samples i do it.
i need another explanations here Phil,i thing i'm loose in your question. :?

jorel 07-10-2003 12:36 PM

Re: sizes with scripts for avisynth208 vs avisynth252
 
Quote:

Originally Posted by Dialhot
Quote:

Originally Posted by jorel
someone saw this :?:

same source (.d2v),33 seconds sample with 1001 frames:

with best script with avisynth252
:arrow: 2,54mb (2.611kb)

with best script with avisynth208
:arrow: 2,38mb (2.448kb)

using the script with avisynth252 is faster with more final size.
using the script with avisynth208 is slow with less final size.
8O
:?:
any comment?


That quite normal : the last optimal script use heavy sharpening (asharp(2,4)) and low bluring (blur(0.1)) where old script was a little lighter (blur(0.2) if I remember well ?).

Moreover, the new script use "Motion dectect" where old use "Hight precision", and that make a big difference on the size (and speed).

FYI, I droped temporalsoften (and use temporalcleaner) because the size is smaller with it. And sometimes it's WAY smaller !

yes Phil,
i post here that i see a "bit" better quality with temporalcleaner
and the file size is a "bit" small too.
thanks for explanations.
:wink:

kwag 07-10-2003 12:59 PM

Re: sizes with scripts for avisynth208 vs avisynth252
 
Quote:

Originally Posted by Dialhot

FYI, I droped temporalsoften (and use temporalcleaner) because the size is smaller with it. And sometimes it's WAY smaller !

Taking notes :)
Will give it a shot sometime today after I compare screenshots ;)

-kwag

bman 07-10-2003 02:07 PM

Hi all !
@ Jorel !
I'm sorry but i can't agree with your conclusion that Best AVS208 script will give u smaller file and looks better than AVS252 with all kind of sources.
I'm encoding my doughters wedding movie which is capture from DV .
What I see is that both of scripts on the same CQ produced file with identical size => 14080kb .
Quality >>>> I can't see any visual difference 8O 8O 8O
I want to notice that script AVS208 includes PIXIE dust (heavy filtering)
======
About "Breathing" effect !!!
As I remember I hade same prob with encoding of AVI files when we were using long GOP.
Then I come to conclusion that min value for GOP to vanish this prob was 24 .
I think u can't couse any damage if u encode 29.97fps movie with max GOP 24 .
Try and see yourself 8)
==========
I want to share with some observations :
Laterly I'm using this script :

Mpeg2Source("I:\E&D-1\E&D.d2v")
BicubicResize(352, 240, 0, 0, 0, 0, 720, 576)
#STMedianFilter(30, 50, 0, 0 )
SwitchThreshold = (Width<=352) ? 4 : (Width<=480) ? 3 : 2
ScriptClip("nf = YDifferenceToNext()"+chr(13)+ "nf >= SwitchThreshold ? \
unfilter( -(fmin(round(nf)*2, 100)), -(fmin(round(nf)*2, 100)) ) : \
TemporalSoften( fmin( round(2/nf), 6), round(1/nf) , round(3/nf) , 1, 1) ")
#ScriptClip("Subtitle(String(nf),10,10)")
LetterBox(32,32,8,8)

AssumeFPS(23.976, true)

sampler(length=72)

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

What I see is :
1 ) If u remove Limiter() , MergeLuma() , MergeChroma() , aSharp()
and as resize filter u use BicubicResize(352, 240, 0, 0, 0, 0,
720, 576) than u get better quality , get smaller file size ( gaine 2-3
CQ units ) and SPEED is ~30% faster ( importent for old comp
owners ?!!)
I've tried this script for 352x240 only and can't be sure for another resolutions .
I think it is interesting , how do u think 8)
bman

ovg64 07-10-2003 03:03 PM

Quote:

FYI, I droped temporalsoften (and use temporalcleaner) because the size is smaller with it. And sometimes it's WAY smaller !
_________________
@ Dialhot,
i really haven't check to see if TemporalCleaner give smaller file size, all i know is i gotten same CQ on source w TCleaner or TSoften. However i turn to Temporalcleaner in my MAS cause like the filter says it gives a Cleaner picture and better Quality overall. I just did Die Another Day in 2 CDs cause of the long n heavy action, and it look like i put a dvd on cds. :D

jorel 07-10-2003 04:04 PM

Quote:

Originally Posted by bman
Hi all !
@ Jorel !
I'm sorry but i can't agree with your conclusion that Best AVS208 script will give u smaller file and looks better than AVS252 with all kind of sources.
I'm encoding my doughters wedding movie which is capture from DV .
What I see is that both of scripts on the same CQ produced file with identical size => 14080kb .
Quality >>>> I can't see any visual difference 8O 8O 8O
I want to notice that script AVS208 includes PIXIE dust (heavy filtering)
======
About "Breathing" effect !!!
As I remember I hade same prob with encoding of AVI files when we were using long GOP.
Then I come to conclusion that min value for GOP to vanish this prob was 24 .
I think u can't couse any damage if u encode 29.97fps movie with max GOP 24 .
Try and see yourself 8)
==========
I want to share with some observations :
Laterly I'm using this script :

Mpeg2Source("I:\E&D-1\E&D.d2v")
BicubicResize(352, 240, 0, 0, 0, 0, 720, 576)
#STMedianFilter(30, 50, 0, 0 )
SwitchThreshold = (Width<=352) ? 4 : (Width<=480) ? 3 : 2
ScriptClip("nf = YDifferenceToNext()"+chr(13)+ "nf >= SwitchThreshold ? \
unfilter( -(fmin(round(nf)*2, 100)), -(fmin(round(nf)*2, 100)) ) : \
TemporalSoften( fmin( round(2/nf), 6), round(1/nf) , round(3/nf) , 1, 1) ")
#ScriptClip("Subtitle(String(nf),10,10)")
LetterBox(32,32,8,8)

AssumeFPS(23.976, true)

sampler(length=72)

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

What I see is :
1 ) If u remove Limiter() , MergeLuma() , MergeChroma() , aSharp()
and as resize filter u use BicubicResize(352, 240, 0, 0, 0, 0,
720, 576) than u get better quality , get smaller file size ( gaine 2-3
CQ units ) and SPEED is ~30% faster ( importent for old comp
owners ?!!)
I've tried this script for 352x240 only and can't be sure for another resolutions .
I think it is interesting , how do u think 8)
bman

just a minute my friend bman.... :!:
clarifications:

the scripts used are exact as posted for Kwag for 2.x and for 252
and was used for each version of avisynth :!:

i post that i got less size with 208 avisynth, :arrow: not better quality :!:

i post about quality when change the temporal smoother
to temporalcleaner in the best script for 252 :!:

about "breathing" :
:arrow: you are the first (after me) that post about
"hade same prob with encoding....when we were using long GOP."
nobody told me this before.

:arrow: you are the first (too) that post:
"Then I come to conclusion that min value for GOP to vanish this prob was 24 .
I think u can't couse any damage if u encode 29.97fps movie with max GOP 24 ."
i never knew it! nobody told me this before too!

conclusion with question:
can i use GOP 24 for a source 29,97,ntsc interlaced :?:
:? :? :? :?

thanks for advices and scripts!
:wink:

jorel 07-10-2003 07:18 PM

bman,Phil,Osvaldo,Kwag,boulder,bp,zooter,Jell,frie ndalunit,d&c,
GFR,plautzer,Red-m,netten,bchteam,J-Wo,logan555,'Z,kane,pyrofire,
totonho,bud,.....all friends here:
can one of you answer my question in the last post?

:arrow: can i use GOP 24 for a source 29,97,ntsc interlaced :?:

this movie is boring me for 4 months,
or i find a way to do a good job or i trash it.....
believe,more than 4 months finding solutions :!:

kwag 07-10-2003 07:20 PM

Sure jorel :)
You can even use a short (12) GOP too :!:

-kwag

jorel 07-10-2003 07:34 PM

Quote:

Originally Posted by kwag
Sure jorel :)
You can even use a short (12) GOP too :!:

-kwag

BIG thanks my dear friend :!:
:D

i will test this GOP values,one in each time:
12,than another with15,another with18, and...24
and compare the results.

the last:
use force film for 29,97 interlaced is a bad idea? :?


All times are GMT -5. The time now is 02:47 PM  —  vBulletin © Jelsoft Enterprises Ltd

Site design, images and content © 2002-2024 The Digital FAQ, www.digitalFAQ.com
Forum Software by vBulletin · Copyright © 2024 Jelsoft Enterprises Ltd.