nnedi3 plugin for VapourSynth

[jackoneill]

v7 fixes everything. (https://github.com/dubhater/vapoursynth-nnedi3/releases/tag/v7)

[Khanattila]

v7 fixes everything. (https://github.com/dubhater/vapoursynth-nnedi3/releases/tag/v7)

Good job. I still can not understand the algorithm of nnnedi3. :D

[jackoneill]

v8 is here (https://github.com/dubhater/vapoursynth-nnedi3/releases/tag/v8). It was here a few days ago already, but I forgot to post this.


* Fix crash with more than 8 but less than 16 bits per sample (bug introduced in v6).


Anybody want to try a semi-portable binary for 64 bit Linux? It should require only Glibc 2.14 or newer. Or, I guess, other libc implementations that pretend to be glibc. You'll probably get some error message about symbols if yours is too old/incompatible.

[Muxketeer]

Sorry to be "that person," but with nnedi3_rpow2 being removed from the plugin, how do we use the python version to increase resolution? I see a python version here http://forum.doom9.org/showthread.php?p=1722246#post1722246, but that doesn't seem to offer a way to increase , both, the width and height.

Is there a already written python version nnedi3_rpow2 ? How can the one at http://forum.doom9.org/showthread.php?p=1722246#post172224 be modified to match nnedi3_rpow2's capabilities?

[Are_]

http://forum.doom9.org/showthread.php?t=172652

[Muxketeer]

http://forum.doom9.org/showthread.php?t=172652

Perfect, thanks!

[asarian]

So, what's the deal with nnedi3_weights.bin?! Seems the only source is just a Makefile. I put it in my plugins64 directory nonetheless, but got:

vapoursynth.Error: nnedi3: 'C:/Program Files (x86)/VapourSynth/plugins64/nnedi3_weights.bin' has the wrong size. Expected 13574928 bytes, got 862 bytes.

[jackoneill]

So, what's the deal with nnedi3_weights.bin?! Seems the only source is just a Makefile. I put it in my plugins64 directory nonetheless, but got:

vapoursynth.Error: nnedi3: 'C:/Program Files (x86)/VapourSynth/plugins64/nnedi3_weights.bin' has the wrong size. Expected 13574928 bytes, got 862 bytes.

If you downloaded the DLL, you must have seen the link to nnedi3_weights.bin. It's right there between the list of changes and the download links for the DLLs.

[asarian]

If you downloaded the DLL, you must have seen the link to nnedi3_weights.bin. It's right there between the list of changes and the download links for the DLLs.

No, I don't see it. :) Or, rather, I'm seeing a link to a Make.am source file, at:

https://github.com/dubhater/vapoursynth-nnedi3/blob/v6/src/nnedi3_weights.bin

My knowledge doesn't extend far enough (on Windows) on how to compile this into a new, 64-bit dll.


EDIT: Nevermind, LOL, I found it in the zip file! :)

[AzraelNewtype]

No, I don't see it. :) Or, rather, I'm seeing a link to a Make.am source file, at:

https://github.com/dubhater/vapoursynth-nnedi3/blob/v6/src/nnedi3_weights.bin

My knowledge doesn't extend far enough (on Windows) on how to compile this into a new, 64-bit dll.


EDIT: Nevermind, LOL, I found it in the zip file! :)

It's good that you found it in the source zip I suppose, but I mean...

http://i.imgur.com/IKbHv0P.png

his instructions on where to find it were not exactly unclear. Or wrong.

[asarian]

Silly me, it was in the RAW file all the time. ;P I kept clicking on the filename itself, which is just a text file. :)

[feisty2]

I think the floating point support needs some correcting, and I made a pr on github

[Pulp Catalyst]

i'm trying to incorporate a yadifmod + NNEDI3 into staxrip,

but i don't know VS, could some one just post a simple example of what the script would look like.

i want yadifmod + nnedi3 but double framerate (50fps) so to avoid the jerky (stutter) effect.

i have looked through all the posts, but havent found a combination of yadifmod and nnedi3 together.

Thanks,

[foxyshadis]

That would be something like:

yadifmod.Yadifmod(yourvid, edeint=nnedi3.nnedi3(yourvid), mode=1)

If it's not working right, then your source isn't setting _FieldBased correctly, and you'll have to preface that with:

core.std.SetFrameProp(yourvid, prop="_FieldBased", intval=X) where X is 1 for bff and 2 for tff.

[Pulp Catalyst]

thankyou foxy, much appreciated.

[jackoneill]

https://github.com/dubhater/vapoursynth-nnedi3/releases/tag/v9


* Float pixels are no longer clamped to the usual ranges of [0.0, 1.0]/[-0.5, 0.5].
* Fix a crash in OS X and possibly 64 bit Linux.
* Fix a bug in the NEON intrinsics.
* Always set the _FieldBased frame property to progressive.
* Ignore the _FieldBased frame property when "dh" is True and use the _Field property instead.
* Remove the deprecated parameters "Y", "U", and "V".
* Replace the bitfield parameter "fapprox" with the parameters "int16_prescreener", "int16_predictor", and "exp".


The fapprox parameter was too annoying to use.

[hydra3333]

x64, replaced nnedi3 v8 with the new nnedi3 v9 and received this error:
"C:\SOFTWARE\Vapoursynth-x64\VSPipe.exe" "T:\HDTV\WDTVlive\MP4-VS\x64-1\test.mpg.2016.08.24.22.56.27.68.vpy" - --y4m | "C:\SOFTWARE\ffmpeg\0-homebuilt-x64\x264-mp4.exe" - --stdin y4m --thread-input --frames 45810 --profile high --level 4.1 --preset slow --interlaced --tff --no-cabac --crf 16 --sar 64:45 --colormatrix bt470bg -o "s:\HDTV\WDTVlive\MP4-VS\x64-1\test-temp.2016.08.24.22.56.27.68.h264"
Script evaluation failed:
Python exception: nnedi3: Function does not take argument(s) named U, V
Traceback (most recent call last):
File "src\cython\vapoursynth.pyx", line 1491, in vapoursynth.vpy_evaluateScript (src\cython\vapoursynth.c:26905)
File "T:\HDTV\WDTVlive\MP4-VS\x64-1\test.mpg.2016.08.24.22.56.27.68.vpy", line 496, in <module>
main()
File "T:\HDTV\WDTVlive\MP4-VS\x64-1\test.mpg.2016.08.24.22.56.27.68.vpy", line 122, in main
video = haf.QTGMC(video, TFF=True, Preset="Very Slow", Sharpness=1.2, SLMode=2, EZKeepGrain=1.2, NoiseProcess=2) # result is double framerate progressive, so re-interlate it later
File "C:\SOFTWARE\Vapoursynth-x64\havsfuncTS.py", line 1130, in QTGMC
edi1 = QTGMC_Interpolate(ediInput, InputType, EdiMode, NNSize, NNeurons, EdiQual, EdiMaxD, bobbed, ChromaEdi, TFF)
File "C:\SOFTWARE\Vapoursynth-x64\havsfuncTS.py", line 1411, in QTGMC_Interpolate
interp = core.nnedi3.nnedi3(Input, field=field, U=CEed, V=CEed, nsize=NNSize, nns=NNeurons, qual=EdiQual)
File "src\cython\vapoursynth.pyx", line 1356, in vapoursynth.Function.__call__ (src\cython\vapoursynth.c:24822)
vapoursynth.Error: nnedi3: Function does not take argument(s) named U, V


x64, reinstated nnedi3 v8 and received this error which I expected since it was a test script under development:
"C:\SOFTWARE\Vapoursynth-x64\VSPipe.exe" "T:\HDTV\WDTVlive\MP4-VS\x64-1\test.mpg.2016.08.24.22.59.13.71.vpy" - --y4m | "C:\SOFTWARE\ffmpeg\0-homebuilt-x64\x264-mp4.exe" - --stdin y4m --thread-input --frames 45810 --profile high --level 4.1 --preset slow --interlaced --tff --no-cabac --crf 16 --sar 64:45 --colormatrix bt470bg -o "s:\HDTV\WDTVlive\MP4-VS\x64-1\test-temp.2016.08.24.22.59.13.71.h264"
Script evaluation failed:
Python exception: bitdepth: you cannot specify both a colorspace and a pixel format.
Traceback (most recent call last):
File "src\cython\vapoursynth.pyx", line 1491, in vapoursynth.vpy_evaluateScript (src\cython\vapoursynth.c:26905)
File "T:\HDTV\WDTVlive\MP4-VS\x64-1\test.mpg.2016.08.24.22.59.13.71.vpy", line 496, in <module>
main()
File "T:\HDTV\WDTVlive\MP4-VS\x64-1\test.mpg.2016.08.24.22.59.13.71.vpy", line 143, in main
video = core.fmtc.bitdepth(video, csp=vs.FLOAT, bits=32, flt=1) #convert to 32bit float
File "src\cython\vapoursynth.pyx", line 1383, in vapoursynth.Function.__call__ (src\cython\vapoursynth.c:25212)
vapoursynth.Error: bitdepth: you cannot specify both a colorspace and a pixel format.

This seems to indicate nnedi3 v9 is not backward compatible with v8 and thus breaks QTGMC per the statement Remove the deprecated parameters "Y", "U", and "V". For the time being I'll need to stick with v8 until havsfunc is updated ...
Script:
# preliminary stuff
video = haf.QTGMC(video, TFF=True, Preset="Very Slow", Sharpness=1.2, SLMode=2, EZKeepGrain=1.2, NoiseProcess=2) # result is double framerate progressive, so re-interlate it later
# Vine only accepts 32bit float input
video = core.fmtc.bitdepth(video, csp=vs.FLOAT, bits=32, flt=1) #convert to 32bit float
video = Vine.Dehalo(video, radius=[2, None])
video = core.fmtc.bitdepth(video, csp=vs.INTEGER, bits=8, flt=0) #convert to 8 bit integer
# other stuff

[jackoneill]

Oops, v9 has a bug that makes it process 16 bit and float input with just the cubic interpolation.

https://github.com/dubhater/vapoursynth-nnedi3/releases/tag/v10

[Sm3n]

Thank you for your work!

[ShogoXT]

Hi all!
Noob here been learning vapoursynth through Staxrip.
I thought I was doing things wrong adding in nnedi3 for resizing, which needed the nnedi3_rpow2 linked in the OP.

I tried to use it, but kept getting fapprox errors. I noticed that setting is gone in v8 and rpow2 hasnt been updated in 10 months. I probably will try the older version for now, but I guess it wont work til adjusted?

Thanks for your hard work!

EDIT: Nvm, Staxrip doesnt allow me to use the non v10 version of nnedi3, so I guess im stuck.

[jackoneill]

Hi all!
Noob here been learning vapoursynth through Staxrip.
I thought I was doing things wrong adding in nnedi3 for resizing, which needed the nnedi3_rpow2 linked in the OP.

I tried to use it, but kept getting fapprox errors. I noticed that setting is gone in v8 and rpow2 hasnt been updated in 10 months. I probably will try the older version for now, but I guess it wont work til adjusted?

Thanks for your hard work!

EDIT: Nvm, Staxrip doesnt allow me to use the non v10 version of nnedi3, so I guess im stuck.

nnedi3_rpow2 doesn't really do anything with the fapprox parameter, so just delete it.

[jackoneill]

https://github.com/dubhater/vapoursynth-nnedi3/releases/tag/v11


* Allow the use of the int16 prescreeners with 9..16 bit input. This
means the faster "new" prescreener (pscrn=2..4) can be used and so
9..16 bit input is processed faster with default parameters.
* Allow the use of the int16 predictor with 9..15 bit input. This
means 9..15 bit input is processed faster (theoretically) with
default parameters.
* Fix memory leak when nnedi3_weights.bin is missing or has the
wrong size (bug introduced in v3).
* readme.rst finally explains what all the parameters are.