DGDecomb

These CUDA filters are packaged into DGDecodeNV, which is part of DGDecNV.
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

Ah, yes, grant gorging, gatekeeping, and publication bias. I know exactly what you are talking about, and again, the sociology is a theme of my paper. One has to be careful talking about this but the special issue seems to invite it, so let's see how far it can be taken.

I've been thinking about this a lot recently. It's obvious that government funding has corrupted science and distorted the scientific process. But some research requires extensive equipment and facilities, so grant money cannot just be awarded randomly. The best I can think of right now, is to award some fraction based on capabilities while reserving some fraction to be randomly awarded to qualified applicants. That doesn't address gatekeeping and publication bias, however. The radical solution is to get government out of it completely; let private enterprise fund things. Then maybe we wouldn't be wasting our money and time on quantum computation and other nonsense.

You've obviously thought about this, gonca. What do you think?
User avatar
gonca
Curly Approved/Moose Approved
Posts: 952
Joined: Sun Apr 08, 2012 6:12 pm

Re: DGDecomb

Post by gonca »

Government funding could easily encourage "politically correct" results, while private funding would be more result oriented, with due care to not rushing to conclusions to fulfill the results that might be required.
The splitting of the funds might be a good alternative with the appropriate safeguards in place to protect the integrity of the funds distribution as well as the validity of the results of the research. In other words, no hanky panky on who gets the money or the results in an attempt to get more funding
DAE avatar
Aleron Ives
Posts: 113
Joined: Fri May 31, 2013 8:36 pm

Re: DGDecomb

Post by Aleron Ives »

admin wrote:The radical solution is to get government out of it completely; let private enterprise fund things.
I would argue that this would cause the opposite of the desired effect. The least reliable studies are funded by corporations with a vested interest in specific outcomes. If corporation x funds a study that says technology y is good and/or safe, and corporation x sells technology y, then it's no surprise that the study it paid for agrees with its business interests. Government funding may not be an ideal solution, but that doesn't mean the private sector has the answer, either.
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

@Aleron

Sure, but there are two sides to that coin. A company doing basic R&D in support of new/improved products is going to be much more connected to reality and focused on real results.

@all

My experiment this morning to run the CUVID postprocessor (VPP) stand-alone on raw input (that is, taking decompressed output from any source filter) was successful (with help from nVidia for one quirky thing). That means you will soon have a stand-alone PureVideo deinterlacer/bobber.
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

PVBob() is working and output is noticeably better than DGBob() (which is the yadif algorithm). I have a little wart with random access with double-rate that I have to fix. So probably tomorrow I can give it to y'all.
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

Slipstreamed PVBob().
User avatar
gonca
Curly Approved/Moose Approved
Posts: 952
Joined: Sun Apr 08, 2012 6:12 pm

Re: DGDecomb

Post by gonca »

Thanks for the new filter
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

You're welcome, Sir.
User avatar
gonca
Curly Approved/Moose Approved
Posts: 952
Joined: Sun Apr 08, 2012 6:12 pm

Re: DGDecomb

Post by gonca »

I ain't no sir :lol:
Just another guy trying to do good encodes
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

It's just a term of endearment and respect for one who has contributed much. Indulge me. :)
User avatar
gonca
Curly Approved/Moose Approved
Posts: 952
Joined: Sun Apr 08, 2012 6:12 pm

Re: DGDecomb

Post by gonca »

I know, and okay
Your skill and loyalty to DGTools users has more than earned my indulgence, it has earned my respect
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

Thank you! Just want to help guys trying to make good encodes to have some CPU left over for encoding. ;)

How's that 1080 Ti working?
User avatar
gonca
Curly Approved/Moose Approved
Posts: 952
Joined: Sun Apr 08, 2012 6:12 pm

Re: DGDecomb

Post by gonca »

Installed it yesterday
Fans don't kick on until it hits 50C
Two fans with asynchronous control
Heat dissipating back plate
Real quiet and working with no glitches, used NVidia's latest drivers and left all the extra stuff on the install disk off
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

Sweet!
User avatar
gonca
Curly Approved/Moose Approved
Posts: 952
Joined: Sun Apr 08, 2012 6:12 pm

Re: DGDecomb

Post by gonca »

Will do but back to "business"

Code: Select all

LoadPlugin("C:\Program Files (Portable)\dgdecnv\x64 Binaries\DGDecodeNV.dll")
DGSource("I:\xxx.dgi", fieldop=0)
DGBob(mode=1)
Image

Code: Select all

LoadPlugin("C:\Program Files (Portable)\dgdecnv\x64 Binaries\DGDecodeNV.dll")
DGSource("I:\xxx.dgi", fieldop=0)
PVBob(mode=1)
Image

note the green lines on pvbob script

Tried another source and same result, driver version related?
I am on 382.05
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

That's not good. Can you give me a stream that causes that (just a few frames should be enough)? If so, I can probably fix it straightaway.

I'm on 381.65.

What version of Avisynth do you run?
User avatar
gonca
Curly Approved/Moose Approved
Posts: 952
Joined: Sun Apr 08, 2012 6:12 pm

Re: DGDecomb

Post by gonca »

Here is the link
http://www.mediafire.com/file/fqb918wl0 ... mple_1.mkv
Sample is from latest AVS+ 64 bit
Tried with AVISynth 2.6 32 bit and same result
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

Duplicated. Investigating...

Thanks for the report.
User avatar
gonca
Curly Approved/Moose Approved
Posts: 952
Joined: Sun Apr 08, 2012 6:12 pm

Re: DGDecomb

Post by gonca »

On a more fun note
Just used Afterburner to test the GPU
Afterburner could only get the GPU usage to 93%, stable at 70C with a clock of 1936MHz
DAE avatar
Sharc
Moose Approved
Posts: 231
Joined: Thu Sep 23, 2010 1:53 pm

Re: DGDecomb

Post by Sharc »

Thanks for the PVBob. I run first tests, and from what I have seen the results look promising.
Just for fun I did a comparison for a synthetic (nasty pathalogical) source. Not that it is representative for any natural content, but regardless how irrelevant the source is for any typical use case it shows how the various bobbers produce different results. Even QTGMC has issues with it, it seems.
Don't downscale it for watching as the subsampling would introduce new artefacts. It's up to you to guess how the original images would look.

http://www.mediafire.com/file/5bka8m44l ... nttest.mkv

Surprisingly for PVBob() the pictures seem to converge to a static picture after the first few frames before they "return to life" when the pictures start to move later on in the video. I think this is very correct as PVBob displays both oroginal fields in every frame when there is no temporal movement (no combing), hence preserving maximum details and sharpness (adaptive deinterlacing/bobbing).

Edit:
Ooops! I noticed only now that gonca has already reported another issue. Interestingly it didn't show up in my uploaded testclip but I could now reproduce it with a different source. So I'll wait for the next slipstream and redo the test.
User avatar
hydra3333
Moose Approved
Posts: 210
Joined: Wed Oct 06, 2010 3:34 am
Contact:

Re: DGDecomb

Post by hydra3333 »

admin wrote:PVBob() is working and output is noticeably better than DGBob() (which is the yadif algorithm).
Nice ! ok, 2 questions.

1. Double-framerate NV deinterlacing used to introduce a double-framerate bug per
https://forum.doom9.org/showthread.php? ... ost1391269
Based on information I received from Nvidia I believe the Nvidia chain introduces a frame delay when doing double-rate deinterlacing. There's naught I can do about it.
http://forum.doom9.org/showthread.php?p ... ost1391556
OK, so I apply trim(1,-999999) to the NV double-rate deinterlaced clip and then it appears just like the others temporally. Thanks for checking.
Does this "feature" exist in the latest DGSource when specifying double framerate deinterlacing ?

2. Does the "feature" apply to either of PVBob (stand-alone PureVideo deinterlacer/bobber) or DGBob (yadif based deinterlacer) ?

Thanks
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

After some head scratching I figured out what is going on with the broken frames. Avisynth does not guarantee that the chroma pitch is 1/2 of the luma pitch! But the VPP takes only one pitch value (the luma pitch) and uses 1/2 of it for the chroma.

I can adjust the pitches with internal copies but before doing that I have a query to nVidia to see if there is any better workaround. I want to get this fixed before addressing your other questions. Thanks for your patience.
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

All fixed. I was able to avoid the extra copying by using the cuMemcpy2D() API to re-pitch the frame as it is copied to the GPU.
User avatar
admin
Site Admin
Posts: 4447
Joined: Thu Sep 09, 2010 3:08 pm

Re: DGDecomb

Post by admin »

@Sharc

You have to be careful with that deint test file. The source version I have of it is not truly interlaced content. It is actually progressive shifted by one field so that field matching cleans it right up. Can you check that with your source?

@hydra3333

Everything is the same with the delay for DGSource and PVBob. DGBob is not affected. I will look at working around it in PVBob.
User avatar
gonca
Curly Approved/Moose Approved
Posts: 952
Joined: Sun Apr 08, 2012 6:12 pm

Re: DGDecomb

Post by gonca »

My file looks alright now, no more green lines
Post Reply