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

Reply
 
Thread Tools Search this Thread Display Modes
Old 4th November 2013, 07:16   #1  |  Link
turbojet
Registered User
 
Join Date: May 2008
Posts: 1,840
Removing these combing(?) artifacts



I'm not sure if this is combing or some other artifact but after trying about 10 deinterlacers I could not find anything that worked.

Sample at http://www.sendspace.com/file/s6yyra if anyone wants to take a stab at it.
__________________
PC: FX-8320 GTS250 HTPC: G1610 GTX650
PotPlayer/MPC-BE LAVFilters MadVR-Bicubic75AR/Lanczos4AR/Lanczos4AR LumaSharpen -Strength0.9-Pattern3-Clamp0.1-OffsetBias2.0

Last edited by turbojet; 4th November 2013 at 07:18.
turbojet is offline   Reply With Quote
Old 4th November 2013, 11:00   #2  |  Link
manono
Moderator
 
Join Date: Oct 2001
Location: Hawaii
Posts: 7,273
Those aren't interlacing artifacts, but low bitrate artifacts. Maybe try Deblock_QED on it.

It IVTC's cleanly and shouldn't be deinterlaced at all
manono is offline   Reply With Quote
Old 4th November 2013, 13:21   #3  |  Link
Guest
Guest
 
Join Date: Jan 2002
Posts: 21,923
Not sure about that being low bitrate artifacts. It looks like somebody did a bad deinterlacing on it, or processed interlaced video as if it was progressive or something. So, turbojet, please tell us more about your source...where you got it and how it was captured/processed. It will be better to mitigate this at the root cause.

Last edited by Guest; 4th November 2013 at 13:26.
Guest is offline   Reply With Quote
Old 4th November 2013, 20:23   #4  |  Link
qwerty1983
Registered User
 
Join Date: Mar 2013
Posts: 44
I think it's 3:2 pulldown. Try inverse telecine

Last edited by qwerty1983; 4th November 2013 at 20:41.
qwerty1983 is offline   Reply With Quote
Old 5th November 2013, 00:25   #5  |  Link
turbojet
Registered User
 
Join Date: May 2008
Posts: 1,840
Deblock_QED didn't do much but blur the image

Source is untouched capture from local CBS affiliate, not cable/satellite

It's 3:2 pulldown and the image in OP is after ivtc. If you frame step the 2 dirty frames out of 5 are really bad looking at times in the mpg from OP. The artifacts really pop out after ivtc while playback in WMP/WMC of the original 30fps mpg it doesn't show much at all, although the dirty frames are still there.
__________________
PC: FX-8320 GTS250 HTPC: G1610 GTX650
PotPlayer/MPC-BE LAVFilters MadVR-Bicubic75AR/Lanczos4AR/Lanczos4AR LumaSharpen -Strength0.9-Pattern3-Clamp0.1-OffsetBias2.0
turbojet is offline   Reply With Quote
Old 5th November 2013, 00:27   #6  |  Link
Guest
Guest
 
Join Date: Jan 2002
Posts: 21,923
You'll get much better help if you post an unprocessed source sample. It's unbelievable that CBS would broadcast anything so bad as what you have shown. If the unprocessed source looks that bad, then you need to look at your capture process. How did you capture that?
Guest is offline   Reply With Quote
Old 5th November 2013, 00:57   #7  |  Link
turbojet
Registered User
 
Join Date: May 2008
Posts: 1,840
Original is in the OP. The capture doesn't look nearly that bad, basically the 2 dirty frames need to be cleaned up to 2 identical clean frames to remove 1. Captured with WMC with no errors according to videoredo. It's not common to find things like this but it is common in a few shows I watch like "The Crazy Ones' here and 'The Blacklist' on NBC.
__________________
PC: FX-8320 GTS250 HTPC: G1610 GTX650
PotPlayer/MPC-BE LAVFilters MadVR-Bicubic75AR/Lanczos4AR/Lanczos4AR LumaSharpen -Strength0.9-Pattern3-Clamp0.1-OffsetBias2.0

Last edited by turbojet; 5th November 2013 at 01:01.
turbojet is offline   Reply With Quote
Old 5th November 2013, 01:12   #8  |  Link
Guest
Guest
 
Join Date: Jan 2002
Posts: 21,923
Quote:
Originally Posted by turbojet View Post
Original is in the OP.
You said:

"the image in OP is after ivtc"

So which is it?

I'm interested in the *unprocessed* source.
Guest is offline   Reply With Quote
Old 5th November 2013, 01:38   #9  |  Link
turbojet
Registered User
 
Join Date: May 2008
Posts: 1,840
Below the image I wrote: Sample at http://www.sendspace.com/file/s6yyra if anyone wants to take a stab at it.

Quoting doesn't work, too short.
__________________
PC: FX-8320 GTS250 HTPC: G1610 GTX650
PotPlayer/MPC-BE LAVFilters MadVR-Bicubic75AR/Lanczos4AR/Lanczos4AR LumaSharpen -Strength0.9-Pattern3-Clamp0.1-OffsetBias2.0
turbojet is offline   Reply With Quote
Old 5th November 2013, 02:19   #10  |  Link
Guest
Guest
 
Join Date: Jan 2002
Posts: 21,923
OK, I see.

Then it's a case of garbage in garbage out. There's nothing useful you can do for this.
Guest is offline   Reply With Quote
Old 6th November 2013, 00:57   #11  |  Link
turbojet
Registered User
 
Join Date: May 2008
Posts: 1,840
Had a feeling that might be the case after looking at it frame by frame, thanks for taking the time to look.
__________________
PC: FX-8320 GTS250 HTPC: G1610 GTX650
PotPlayer/MPC-BE LAVFilters MadVR-Bicubic75AR/Lanczos4AR/Lanczos4AR LumaSharpen -Strength0.9-Pattern3-Clamp0.1-OffsetBias2.0
turbojet is offline   Reply With Quote
Old 7th November 2013, 15:44   #12  |  Link
feisty2
I'm Siri
 
feisty2's Avatar
 
Join Date: Oct 2012
Location: Los Angeles, California
Posts: 2,146
typical mpeg2 compression artifacts
use strong mdegain
something like
Code:
a=last
a.smdegrain (tr=12,thsad=2000,thsadc=2000).repair (a,mode=13)
should clean the mess

Last edited by feisty2; 7th November 2013 at 15:48.
feisty2 is offline   Reply With Quote
Old 9th November 2013, 08:51   #13  |  Link
turbojet
Registered User
 
Join Date: May 2008
Posts: 1,840
SMDegrain is really slow, I suppose MT could help but I don't plan on spending all that much time for the few times this issue occurs, thanks for the suggestion though.

TDecimate(hybrid=1) does something interesting, it discards the bad frames, mostly, and blends the surrounding good frames. This would work great if instead of blending it would interpolate the surrounding 'good' frames, is that possible?
__________________
PC: FX-8320 GTS250 HTPC: G1610 GTX650
PotPlayer/MPC-BE LAVFilters MadVR-Bicubic75AR/Lanczos4AR/Lanczos4AR LumaSharpen -Strength0.9-Pattern3-Clamp0.1-OffsetBias2.0
turbojet is offline   Reply With Quote
Old 9th November 2013, 12:43   #14  |  Link
feisty2
I'm Siri
 
feisty2's Avatar
 
Join Date: Oct 2012
Location: Los Angeles, California
Posts: 2,146
Quote:
Originally Posted by turbojet View Post
SMDegrain is really slow, I suppose MT could help but I don't plan on spending all that much time for the few times this issue occurs, thanks for the suggestion though.

TDecimate(hybrid=1) does something interesting, it discards the bad frames, mostly, and blends the surrounding good frames. This would work great if instead of blending it would interpolate the surrounding 'good' frames, is that possible?
yes, it's doable,you can simply remove unfixable bad frames and create a good frame by using MFlowFps in MVTool2,but if the frame is in the middle of a high motion, the interpolated frame would be very ugly (artifacts like image dilaceration) so I still suggest using MDegrain instead of MFlowFps, MDegrain can almost fix any kind of bad frames if "thSAD" is high enough

Last edited by feisty2; 9th November 2013 at 12:51.
feisty2 is offline   Reply With Quote
Old 9th November 2013, 12:46   #15  |  Link
Taurus
Registered User
 
Taurus's Avatar
 
Join Date: Mar 2002
Location: Krautland
Posts: 875
Quote:
Originally Posted by feisty2 View Post
typical mpeg2 compression artifacts
use strong mdegain
something like
Code:
a=last
a.smdegrain (tr=12,thsad=2000,thsadc=2000).repair (a,mode=13)
should clean the mess
@feisty:
your code snippet gives me:
Code:
MAnalyse does not have a named argument "multi"" (SMDegrain v.2.2d)
Using Avisynth 2.6.0.5 MT (SET's build)
mvtools2 v.2.5.11.13
mt_masktools26 2.0.48.0
A simple call "smdegrain()" works fine.
Taurus is offline   Reply With Quote
Old 9th November 2013, 12:55   #16  |  Link
feisty2
I'm Siri
 
feisty2's Avatar
 
Join Date: Oct 2012
Location: Los Angeles, California
Posts: 2,146
Quote:
Originally Posted by Taurus View Post
@feisty:
your code snippet gives me:
Code:
MAnalyse does not have a named argument "multi"" (SMDegrain v.2.2d)
Using Avisynth 2.6.0.5 MT (SET's build)
mvtools2 v.2.5.11.13
mt_masktools26 2.0.48.0
A simple call "smdegrain()" works fine.
you should use MVtools2mod from ditherpackage if tr>3 or lsb=true
feisty2 is offline   Reply With Quote
Old 9th November 2013, 13:08   #17  |  Link
Taurus
Registered User
 
Taurus's Avatar
 
Join Date: Mar 2002
Location: Krautland
Posts: 875
@feisty2:

Thank you
Taurus is offline   Reply With Quote
Old 6th August 2016, 18:15   #18  |  Link
HeadFi
Registered User
 
HeadFi's Avatar
 
Join Date: Feb 2015
Location: London
Posts: 2
Hi

Quote:
Originally Posted by feisty2 View Post
typical mpeg2 compression artifacts
use strong mdegain
something like
Code:
a=last
a.smdegrain (tr=12,thsad=2000,thsadc=2000).repair (a,mode=13)
should clean the mess
Hahaha that is insane radius mate. It will take weeks to encode )
HeadFi is offline   Reply With Quote
Old 6th August 2016, 18:38   #19  |  Link
feisty2
I'm Siri
 
feisty2's Avatar
 
Join Date: Oct 2012
Location: Los Angeles, California
Posts: 2,146
well...
do you really have to resurrect a 3 yrs ago thread...?
__________________
If I got new ideas, will post here: https://github.com/IFeelBloated
feisty2 is offline   Reply With Quote
Old 8th August 2016, 09:23   #20  |  Link
kuchikirukia
Registered User
 
Join Date: Oct 2014
Posts: 448
If OP tried tr=12 he still might be encoding it.
kuchikirukia is offline   Reply With Quote
Reply

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 10:24.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, vBulletin Solutions Inc.