Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > Capturing and Editing Video > Avisynth Usage

Closed Thread
 
Thread Tools Search this Thread Display Modes
Old 19th November 2010, 16:44   #101  |  Link
William.Lemos.BR
Registered User
 
Join Date: Oct 2010
Posts: 18
Quote:
Originally Posted by Boulotaur2024 View Post
Sorry for not being ontopic, but "best results" compared to what ? (just out of curiosity)
Actually is kind of obvious it's better to filter the video after upscaling it and, basically, that was what I confirmed. Indeed, the noise profile generated in Neat Video has more quality as big as the video is.

Regarding Neat Video: I've tested a lot of other softwares and plugins and conclueded this was the best (of the worst) way of filtering without loosing details (since you manage how to set it's parameters appropriately). What I noticed is that all filtering aways kills the details. But I'm new to Avisynth, and I'm pretty sure that with more knowlege I'll find a better way of doing it.
William.Lemos.BR is offline  
Old 19th November 2010, 18:54   #102  |  Link
Didée
Registered User
 
Join Date: Apr 2002
Location: Germany
Posts: 5,389
Quote:
Originally Posted by William.Lemos.BR View Post
after 24 hours of researching and testing I can say with ABSOLUT certainty: VIDEO ENHANCER IS A LIEWARE!

It doesn't use super resolution, as it caims.
No, that's overboard again. It surely uses SR. The problem is that for this kind of SR to work out, you need input with special properties.

Take some random video.

a) Downscale with bicubicresize(halfwidth,halfheight,0.5,0.25). Upscale with VideoEnhancer. It will achieve "almost nothing" compared to a simple re-scaling.

b) Downscale with gaussresize(halfwidth,halfheight,p=100). Upscale with VideoEnhancer. It will achieve a good improvement over a simple re-scaling.

TSR can work pretty good when the input is of the "decimated resolution" kind. The only pity is that the vast majority of all sources doesn't fall into this category.


Quote:
But the idea behind super resolution is really interesting, do you agree? Going to give a try. In the end at least I'm going to learn usefull stuff.
Sure it's an interesting idea. But as noted above, in most everyday cases it is almost impossible to perform.

Do a forum search on the term "super resolution" / "superresolution". The topic has been discussed a few times already. Usual partipiciants have been: the naive optimists stating "THIS IS IT!", and the pessimistic naysayers stating "Y've been rickrolled". Oh, and some others that were throwing whitepapers about different sensational breakthrough algorithms.

The current status is:

We have lots of paper, but no functioning general-purpose superresolution.
__________________
- We´re at the beginning of the end of mankind´s childhood -

My little flickr gallery. (Yes indeed, I do have hobbies other than digital video!)

Last edited by Didée; 19th November 2010 at 18:58.
Didée is offline  
Old 19th November 2010, 18:59   #103  |  Link
dansrfe
Registered User
 
Join Date: Jan 2009
Posts: 1,210
Quote:
Originally Posted by Didée View Post
We have lots of paper, but no functioning superresolution
Well said.
dansrfe is offline  
Old 19th November 2010, 19:10   #104  |  Link
William.Lemos.BR
Registered User
 
Join Date: Oct 2010
Posts: 18
Quote:
Originally Posted by Didée View Post
No, that's overboard again. It surely uses SR. The problem is that for this kind of SR to work out, you need input with special properties.
Actually I know somehow it uses SR, but not the way it promisses, 'cause it doesn't ADD new information to the resulting image by analysing each frame in a sequence (for instance: a small blurry face will continue blurred, but bigger; you're not going to see eyes mouth and nose, even blurred, as I imagine a real SR process would do). For me this is not much diferent or revolutionary compared to other enlargement methods.

Quote:
Originally Posted by Didée View Post
We have lots of paper, but no functioning general-purpose superresolution.
That's it, you're right. What a pity!

Last edited by William.Lemos.BR; 19th November 2010 at 19:21. Reason: Added information
William.Lemos.BR is offline  
Old 19th November 2010, 19:56   #105  |  Link
Didée
Registered User
 
Join Date: Apr 2002
Location: Germany
Posts: 5,389
However, let me point out the term "general purpose".

It's a different story when talking about "specialised conditions". For example, the topic of this very thread is a script that truly performs temporal superresolution.

It is just limited by two conditions:

a) the input has to be interlaced. (Interlacing is a native form of "decimated resolution")

b) the more lowpass filtering was used to create the original interlacing, the less effective the TSR part will be. (Lowpassing effectively changes "decimation" into "blurring")
__________________
- We´re at the beginning of the end of mankind´s childhood -

My little flickr gallery. (Yes indeed, I do have hobbies other than digital video!)
Didée is offline  
Old 2nd December 2010, 12:42   #106  |  Link
plonk420
amd & h.264 fanboy
 
plonk420's Avatar
 
Join Date: Jun 2002
Location: NTSC
Posts: 420
is this supposed to be a 30fps-oriented deinterlacer? it seems to have half the framerate (or at least notable motion) that mcbob gives me... on even frames, the image seems to be just slightly changed from the odd ones...

love the speed, tho!
plonk420 is offline  
Old 2nd December 2010, 13:44   #107  |  Link
aegisofrime
Registered User
 
Join Date: Apr 2009
Posts: 478
Quote:
Originally Posted by plonk420 View Post
is this supposed to be a 30fps-oriented deinterlacer? it seems to have half the framerate (or at least notable motion) that mcbob gives me... on even frames, the image seems to be just slightly changed from the odd ones...

love the speed, tho!
It's a bobbing deinterlacer. It does have a mode for processing progressive content that outputs same framerate, might you have used that by mistake? The parameters for that mode is

Code:
QuickTGMC(InputType=1)#or 2, or 3
aegisofrime is offline  
Old 5th December 2010, 07:15   #108  |  Link
dansrfe
Registered User
 
Join Date: Jan 2009
Posts: 1,210
yeah you still have to use srestore after this to remove all the extra dupes and blended frames.
dansrfe is offline  
Old 6th December 2010, 16:08   #109  |  Link
kypec
User of free A/V tools
 
kypec's Avatar
 
Join Date: Jul 2006
Location: SK
Posts: 826
How to use it for intermediate files with x264 encoding?

Well, this is my first experience with any lossless video codec. I want to use it for intermediate files that are the product of heavy processing in Avisynth. Later I plan to encode those lossless files with x264 2-pass method.
So far I was able to create AVI files
Code:
Video
ID                               : 0
Format                           : ULY0
Codec ID                         : ULY0
Duration                         : 1mn 0s
Bit rate                         : 50.7 Mbps
Width                            : 720 pixels
Height                           : 576 pixels
Display aspect ratio             : 5:4
Frame rate                       : 25.000 fps
Standard                         : PAL
Bits/(Pixel*Frame)               : 4.893
Stream size                      : 363 MiB (100%)
via this command:
Code:
avs2avi.exe myscript.avs -c ULY0
However, x264 can't recognize my AVI at all
What is the recommended way to feed UtVideo file into x264 encoder then?
kypec is offline  
Old 7th December 2010, 08:58   #110  |  Link
yup
Registered User
 
Join Date: Feb 2003
Location: Russia, Moscow
Posts: 854
kypec!
try use simple script as input for x264
Code:
AVISource("Your file.avi")
I think this different approach for 420 coding.
yup.
yup is offline  
Old 15th December 2010, 05:10   #111  |  Link
7ekno
Guest
 
Posts: n/a
Quote:
Originally Posted by William.Lemos.BR View Post
Totally agree with you, just don't have any idea of how to do it!
I would start with SimpleSlugUpscale, it works quite well on de-interlacing and upscaling Sure it's not done within QTGMC, but it a nice "all in one" !

And it's a nice place to find all the filters needed for SSU & QTGMC in the one place (thanks Robert!)

Also for those with little Avisynth experience, the walk thru Robert has on his website (to get his filter & QTGMC) working is also pretty helpful !!

Quote:
Originally Posted by William.Lemos.BR View Post
Regarding Neat Video: I've tested a lot of other softwares and plugins and conclueded this was the best
Um, yeah :P If I were you I would check out MCTemporalDenoise and/or DFTtest

A simple command like MCTemporalDenoise(settings="low", gpu=true) will beat most things for denoising and speed !!! But of course it's probably far more customisable than NEATVideo !!!

7ek

Last edited by 7ekno; 15th December 2010 at 05:22. Reason: added links
 
Old 18th December 2010, 14:50   #112  |  Link
Desmodeus
Compulsive Tinkerer
 
Desmodeus's Avatar
 
Join Date: Dec 2010
Location: Trapped in the Matrix
Posts: 4
Comment Removed

Last edited by Desmodeus; 18th December 2010 at 15:41. Reason: too complicated to properly explain
Desmodeus is offline  
Old 20th December 2010, 07:51   #113  |  Link
Mounir
Registered User
 
Join Date: Nov 2006
Posts: 773
Is YUY2 /RGB support planned for QTMC ? I'm not a big fan of YV12
Mounir is offline  
Old 20th December 2010, 22:30   #114  |  Link
-Vit-
Registered User
 
Join Date: Jul 2010
Posts: 448
Quote:
Originally Posted by Mounir View Post
Is YUY2 /RGB support planned for QTMC ? I'm not a big fan of YV12
Adding other colorspaces wasn't on my to-do list, but I guess YUY2 should be fairly straightforward. It won't make it into the next update (due soon), perhaps the one after.
I don't think MVTools2 supports RGB input and MVTools2 is at the core of the script.
-Vit- is offline  
Old 20th December 2010, 23:18   #115  |  Link
Didée
Registered User
 
Join Date: Apr 2002
Location: Germany
Posts: 5,389
If it has to be by all means, you can do RGB, too. Take RGB in, clone it to YV12 to do the motion search. Then separate the three RGB channels to three greyscale YV12 clips, perform basically 3 times TGMC on that bunch, and join back to RGB.

"Mischief managed."
__________________
- We´re at the beginning of the end of mankind´s childhood -

My little flickr gallery. (Yes indeed, I do have hobbies other than digital video!)
Didée is offline  
Old 24th December 2010, 00:14   #116  |  Link
palwan
Registered User
 
Join Date: Nov 2010
Posts: 9
Any idea, why I get slower speed using QTGMC(preset="v.fast") ? what I mean I get faster speed using the options < "v.fast".

Last edited by palwan; 24th December 2010 at 00:16.
palwan is offline  
Old 24th December 2010, 01:00   #117  |  Link
Gavino
Avisynth language lover
 
Join Date: Dec 2007
Location: Spain
Posts: 3,431
Quote:
Originally Posted by aman5 View Post
Any idea, why I get slower speed using QTGMC(preset="v.fast") ? what I mean I get faster speed using the options < "v.fast".
It should be preset="Very Fast".
"v.fast" is not recognised, so it reverts to the default of "Slower".
__________________
GScript and GRunT - complex Avisynth scripting made easier
Gavino is offline  
Old 24th December 2010, 06:38   #118  |  Link
dansrfe
Registered User
 
Join Date: Jan 2009
Posts: 1,210
Quote:
Originally Posted by aman5 View Post
Any idea, why I get slower speed using QTGMC(preset="v.fast") ? what I mean I get faster speed using the options < "v.fast".
With that logic I should be getting extreme or super duper fast speeds using QTGMC(preset="e.fast") or QTGMC(preset="s.d.fast")
dansrfe is offline  
Old 24th December 2010, 08:39   #119  |  Link
palwan
Registered User
 
Join Date: Nov 2010
Posts: 9
as it as been showed in script V.Fast,S.Fast,U.Fast, so I went according to that, but now I will try it with "Very Fast" and see the result.
palwan is offline  
Old 24th December 2010, 09:58   #120  |  Link
Gavino
Avisynth language lover
 
Join Date: Dec 2007
Location: Spain
Posts: 3,431
Quote:
Originally Posted by aman5 View Post
as it as been showed in script V.Fast,S.Fast,U.Fast, so I went according to that
You were looking at a comment in the code that uses abbreviations to fit on the line for the table below it.
For the actual values, see the comments in the Presets/Tuning section.
__________________
GScript and GRunT - complex Avisynth scripting made easier
Gavino is offline  
Closed Thread

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 22:06.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.