aWarpSharp causing green on the sides of video

[killerteengohan]

aWarpSharp(depth=7.5, blurlevel=1)

I have been using this filter for years and in the past this has never happened to me.

It doesn't happen on my older computer when I use this, but when I use it on my newer computer it all of a sudden has green appearing on the sides like this. If you look close enough you can see it most under his arm and close to his elbows.




Its easier to see with a dark background like in my album with more examples in it.
http://imgur.com/a/0aOYW#0



I'm using the same exact plugins, the same exact programs and the same encoding method. Why would these only be showing up on my newer computer and not my older one?

Whats the cause of these coming up all of a sudden?

Its not just this source either, its happening on every source I try on my newer computer but not on my older computer.

[poisondeathray]

No idea why it only occurs on some setups, but it's a known issue, use AwarpSharp2

http://forum.doom9.org/showthread.php?t=147285

[killerteengohan]

Originally Posted by poisondeathray

No idea why it only occurs on some setups, but it's a known issue, use AwarpSharp2

http://forum.doom9.org/showthread.php?t=147285

That filter gives alot more trouble than its worth and didnt do that great of job when I tried it a while back. I tried it before and didn't like it, that's why I stuck with the original. Its more damaging than the original one is by over thinning lines or making it look bad.


I've never seen or had a problem with the green on the sides before until just now on this newer computer.

Same source on old computer with same plugins and programs, no green on the sides but there sure is on the newer computer.

There's got to be some kind of reason for this coming out of nowhere I would think.

[poisondeathray]

Originally Posted by killerteengohan

That filter gives alot more trouble than its worth and didnt do that great of job when I tried it a while back. I tried it before and didn't like it, that's why I stuck with the original. Its more damaging than the original one is by over thinning lines or making it look bad.

Did you read the instructions on mapping from original AwarpSharp parameters ?

I've never seen or had a problem with the green on the sides before until just now on this newer computer.

Same source on old computer with same plugins and programs, no green on the sides but there sure is on the newer computer.

There's got to be some kind of reason for this coming out of nowhere I would think.

I had the same problem on a few computers, and could never track down the issue about what caused it, what conditions, what plugins might be contributing.

Eitherway, Awarpsharp2 fixes everything .

[killerteengohan]

Originally Posted by poisondeathray

Did you read the instructions on mapping from original AwarpSharp parameters ?




I had the same problem on a few computers, and could never track down the issue about what caused it, what conditions, what plugins might be contributing.

Eitherway, Awarpsharp2 fixes everything .

No I saw no such instructions about mapping. Can you tell me where to find them?

I did however find a 2012 update recently posted not too long ago and that one seems to work better than the 2009 one I had tried a couple years ago.

I used to have to type awarpsharp2() and the parameter depth= didnt exist but it seems like it works in the newer one. Can I use it just like the old awarpsharp with the same parameters and not get the green?

[poisondeathray]

Originally Posted by killerteengohan

No I saw no such instructions about mapping. Can you tell me where to find them?

In the awarpsharp2 instructions, there is a section on "Mapping from original aWarpSharp parameters"

[Reel.Deel]

For some reason I could never get the original aWarpSharp to coincide with aWarpSharp2's aWarpSharp (different outputs). From my understanding there was an update to the original aWarpSharp some time ago.

Originally Posted by Didée

I heard a rumour those DLLs were "slightly modified". But looking at the file dates, and remembering that MarcFD wasn't the busyiest with releasing sources ...

Originally Posted by Soulhunter

Ok, I asked pruned for the exact changes...

- Chroma to luma warper

- Improved stability

- Improved precision

- Other small fixes

Unfortunately all of the download links in that thread are dead. Does anyone have the updated aWarpSharp?

[killerteengohan]

Originally Posted by Reel.Deel

For some reason I could never get the original aWarpSharp to coincide with aWarpSharp2's aWarpSharp (different outputs). From my understanding there was an update to the original aWarpSharp some time ago.


Unfortunately all of the download links in that thread are dead. Does anyone have the updated aWarpSharp?

I only have a 2003 version for awarpsharp and 2009 awarpsharp2, but I did find a 2012 one here. http://forum.doom9.org/showthread.php?t=147285

I realize its awarpsharp2 but from the updates description it might do what you said you couldnt get it to do just above.

Originally Posted by SEt

While I'm at it, let's post some improvements that were implemented years ago but never made it to public: new version 2012.03.28.

Changes:
1) Added new final part aWarp4 that takes 4x upsampled source for better final subpixel interpolation. Allows producing much sharper result with lower depth, so less deformations. Examples:

aWarp4(Spline36Resize(width*4, height*4, 0.375, 0.375), aSobel().aBlur(), depth=3)
aWarp4(nnedi3_rpow2(rfactor=2).Spline36Resize(width*4, height*4, 0.25, 0.25), aSobel().aBlur(), depth=3)
aWarp4(nnedi3_rpow2(rfactor=2).nnedi3_rpow2(rfactor=2), aSobel().aBlur(), depth=2)

Note that upsampling should be left-top aligned, so Spline36Resize(width*4, height*4) or nnedi3_rpow2(rfactor=4) won't produce correct results.

2) aWarpSharp2 can now be also called by aWarpSharp() with internal parameters conversion for compatibility with ancient scripts. Direct aWarpSharp2 call is encouraged though.
3) New chroma values: 5 and 6: the same as 3 and 4, but luma isn't processed and just copied. Useful for filtering sources with certain chroma problems.
4) Changed some defaults.

[Reel.Deel]

Originally Posted by killerteengohan

I only have a 2003 version for awarpsharp and 2009 awarpsharp2, but I did find a 2012 one here. http://forum.doom9.org/showthread.php?t=147285

I realize its awarpsharp2 but from the updates description it might do what you said you couldnt get it to do just above.

No that's not it, I'm already using aWarpSharp2 from 2012. I tried to get similar output from aWarpsharp (2003) and aWarpSharp2 (2012) but I never could (still can't). The updated aWarpsharp I'm talking about was released around 2005.

[poisondeathray]

It probably won't match up perfectly when you use decimal values, because depth parameter in awarpsharp2 only takes integer values. I think it rounds the value in the auto conversion when you call awarpsharp instead of awarpsharp2 (which gives you an error). That might be the reason behind the green borders in the first place

The mapping
depth = depth*blurlevel/2

eg. old awarpsharp
aWarpSharp(depth=7.5, blurlevel=1)

7.5 * 1/2 = 3.75

[Reel.Deel]

Yeah that's probably one the reasons. In addition to that, the old aWarpSharp cm parameter defaults to 3 (included documentation is wrong) and there seems to be no valid mapping in aWarpSharp2.

Mapping
chroma = 0->2, 1->4, 2->3

[killerteengohan]

I did notice if you crop a multiple of 16 off the sides like 10 off the right and 6 off the left, or whatever gets you a multiple of 16, the green goes away mostly or reduces alot but theres still obvious marks on the video, its just not bright green and overly apparent.

[2Bdecided]

The easy solution is to add borders before using awarpsharp and crop them off again afterwards.

The greening is content dependent.

Cheers,
David.

[foxyshadis]

Originally Posted by Reel.Deel

Yeah that's probably one the reasons. In addition to that, the old aWarpSharp cm parameter defaults to 3 (included documentation is wrong) and there seems to be no valid mapping in aWarpSharp2.

Mapping
chroma = 0->2, 1->4, 2->3

No. 0=>1, 1=>4, 2=>3, 3=>2. aWarpSharp2's were rearranged to be in line with Masktools (though I wish he'd added the Masktools2 string parameters).

[raffriff42]

Originally Posted by 2Bdecided

The easy solution is to add borders before using awarpsharp and crop them off again afterwards.

Better than borders: duplicate edges with Padding()

[SEt]

aWarpSharp2 is exactly the same algorithm as aWarpSharp, only with some bugfixes. If you are getting significantly different results – you are doing something wrong. Latest aWarpSharp2 supports aWarpSharp alias that remaps the parameter values automatically if you absolutely can't read the documentation.

Btw, aWarpSharp2 has integer parameters because algorithm uses actually integers, so again aWarpSharp2 is exactly the same as aWarpSharp, just removes the false feeling that slight change in float number changes something when it actually maps to the same integer inside.

[Reel.Deel]

Originally Posted by foxyshadis

No. 0=>1, 1=>4, 2=>3, 3=>2. aWarpSharp2's were rearranged to be in line with Masktools (though I wish he'd added the Masktools2 string parameters).

Actually I was talking about the mapping.

Originally Posted by aWarpSharp2 documentation

Original aWarpSharp compatibility:
Mapping from original aWarpSharp parameters:
thresh = thresh*256
blur = blurlevel
depth = depth*blurlevel/2
chroma = 0->2, 1->4, 2->3

This version is able to do the conversion automatically when called as aWarpSharp,
but errors in case of out of bounds values will be displayed in new terms.

Originally Posted by SEt

aWarpSharp2 is exactly the same algorithm as aWarpSharp, only with some bugfixes. If you are getting significantly different results – you are doing something wrong.

Yeah, it's been a while since I tried it, I'll double check myself when I get some time.

[SEt]

aWarpSharp2's documentation is correct regarding how aWarpSharp version "5 fev 2003" actually works. Though aWarpSharp indeed accepts cm=3, it's just another bug as it works exactly as cm=1 (and cm=1 is actual default).

[AzraelNewtype]

Originally Posted by raffriff42

Better than borders: duplicate edges with Padding()

In what way is this better than adding on black pixels that you expect to be corrupted and never actually want to view?

[raffriff42]

Because since WarpSharp works by averaging* pixels over a certain area, it can "drag" some of the black pixels into view.

*EDIT not averaging of course, poor choice of word. For any kind of sharpening or blurring, each output pixel will be affected by all its neighbor pixels.