Vapoursynth

[AOmundson]

Originally Posted by poisondeathray

Did you get the path correct ?

Are you using recent LSmash version ?

Personally, I would use neither for a DVD source. DGSource if you have a Nvidia card and license, or DGDecode / d2v.Source are more reliable

Thanks, I've got it working now. When you say to use DGSource/DGDecode instead of LSmash/FFmpeg, how would I go about implementing that? i.e. What DLLs would I need to download for the VS plugin folder, how should I write the code line in the editor itself, and what should I place in the Header, Executable, and Arguments boxes?

[LessThanJake]

I have installed VapourSynth via installer (VapourSynth64-R52.exe).

print(core.version()) for system wide python install says:
VapourSynth Video Processing Library
Copyright (c) 2012-2020 Fredrik Mellbin
Core R52
API R3.6
Options: -

If I create my own environment and add Vapoursynth via "pip install vapoursynth" says:
print(core.version())
VapourSynth Video Processing Library
Copyright (c) 2012-2020 Fredrik Mellbin
Core R51
API R3.6
Options: -

Documentation says:


If installation via pip requires installation of VapourSynth beforehand I assume it referes to the global install (R52) but it says R51.
So what version do I actually have when using pip in my own env and does it matter that it only shows R51 and can s.o. update PyPi to R52 that it matches again?

Thanks

[jinkazuya]

I am just wondering how to use the script here https://github.com/Selur/Vapoursynth...41Fun.py#L2885
if I use vapoursynth? I do not know what to do with it. Thanks and hope somebody could help. I would like to use the script or plugin of the detailsharpen.

[vxzms]

Originally Posted by jinkazuya

I am just wondering how to use the script here https://github.com/Selur/Vapoursynth...41Fun.py#L2885
if I use vapoursynth? I do not know what to do with it. Thanks and hope somebody could help. I would like to use the script or plugin of the detailsharpen.

import vapoursynth as vs
from vapoursynth import core
import G41Fun

src = core.lsmas.LWLibavSource(clip)
dsharped = G41Fun.DetailSharpen(src, z=4, sstr=1.5, power=4, ldmp=1, mode=1, med=False)

dsharped.set_output()

[jinkazuya]

Originally Posted by vxzms

import vapoursynth as vs
from vapoursynth import core
import G41Fun

src = core.lsmas.LWLibavSource(clip)
dsharped = G41Fun.DetailSharpen(src, z=4, sstr=1.5, power=4, ldmp=1, mode=1, med=False)

dsharped.set_output()

Thanks but how to use it with staxrip and what is the procedure of using it. Since it is the script of python, I do not know how to use it.

[monohouse]

I have some problem with MakeDiff in versions 50,51,52 :x

Script evaluation failed:
Python exception: MakeDiff: both clips must have constant format and dimensions, and the same format and dimensions

Traceback (most recent call last):
  File "src/cython/vapoursynth.pyx", line 2244, in vapoursynth.vpy_evaluateScript
  File "src/cython/vapoursynth.pyx", line 2245, in vapoursynth.vpy_evaluateScript
  File "/mnt/dod/04.vpy", line 11, in <module>
    clip = haf.QTGMC(clip, TFF=True, Preset="Placebo", ShowSettings=False, opencl=False, device=0, TR0=2, TR1=2, SourceMatch=0, TR2=0, Lossless=0, Sharpness=0 )
  File "/usr/lib/python3.6/site-packages/havsfunc.py", line 1281, in QTGMC
    noise = core.std.MakeDiff(clip, denoised, planes=CNplanes)
  File "src/cython/vapoursynth.pyx", line 2069, in vapoursynth.Function.__call__
vapoursynth.Error: MakeDiff: both clips must have constant format and dimensions, and the same format and dimensions

linux, all compiled today, used LD_PRELOAD to test for plugin errors and there are none, without QTGMC the script works, latest mvsfunc and havsfunc

import vapoursynth as vs
import havsfunc as haf

core = vs.get_core()
core.set_max_cache_size(16000)

clip = core.ffms2.Source("//mnt//dod//movie.mkv")
clip = core.fmtc.bitdepth (clip, bits=16, fulls=False, fulld=True, dmode=1)
clip = haf.QTGMC(clip, TFF=True, Preset="Placebo", ShowSettings=False, opencl=False, device=0, TR0=2, TR1=2, SourceMatch=0, TR2=0, Lossless=0, Sharpness=0 )
clip = haf.MCTemporalDenoise(clip,  radius=3, limit=2,  twopass=False, limit2=2,  refine=True, useTTmpSm=True,  stabilize=False, maxr=3, TTstr=3,  chroma=False, MVsharp=False, sigma=0, pfMode=-1,search=3, searchparam=4, pel=4, pelsearch=4, bwbh=512, owoh=256, blksize=4, overlap=2, deblock=False, post=0,bt=4,thSAD=2000, thSAD2=2000, thSCD1=1000, thSCD2=200)
clip = core.deblock.Deblock(clip, quant=20, aoffset=0, boffset=0)
clip = core.fmtc.bitdepth (clip, bits=10, fulls=True, fulld=True, dmode=1)
clip.set_output()

tested several older versions of havsfunc no change :x
I removed this code from src/core/mergefilters.c and it started working fine and the video looks correct

    if (!isConstantFormat(d.vi) || !isSameFormat(d.vi, vsapi->getVideoInfo(d.node2))) {
        vsapi->freeNode(d.node1);
        vsapi->freeNode(d.node2);
        RETERROR("MakeDiff: both clips must have constant format and dimensions, and the same format and dimensions");
    }

strange problem

[LigH]

Originally Posted by jinkazuya

Thanks but how to use it with staxrip and what is the procedure of using it. Since it is the script of python, I do not know how to use it.

First a question to stax76 ... does StaxRip support VapourSynth filters at all, as selectable filter? Probably not easily, because StaxRip builds AviSynth scripts. I don't think you can simply mix both worlds.

Or maybe I'm wrong. I noticed that StaxRip x64 ships both frameservers. But how to select either or mix both, may be better asked in his own thread about StaxRip.

[stax76]

@LigH

I've expected a little more competence from somebody like you ... , right-click filters and choose:

Filter Setup > VapourSynth

After that:

File > Save Project As Template > Load template on startup

Vapoursynth is now your default frame server ready to use in staxrip, it currently has 60 vapoursynth plugins included, most of them with presets available in an easy-to-use menu.

https://github.com/staxrip/staxrip/wiki/Tools

[jinkazuya]

Originally Posted by stax76

@LigH

I've expected a little more competence from somebody like you ... , right-click filters and choose:

Filter Setup > VapourSynth

After that:

File > Save Project As Template > Load template on startup

Vapoursynth is now your default frame server ready to use in staxrip, it currently has 60 vapoursynth plugins included, most of them with presets available in an easy-to-use menu.

https://github.com/staxrip/staxrip/wiki/Tools

Hi stax76, would that be ok for you to include the plugin for Detailsharpen in the future release? I greatly appreciate. Staxrip is a pretty good application or program but I used to use MEGUI and now MEGUI and Avisynth+ become a bit complicated and none of the script that I used work and the encoding always throw exception when it is halfway done. But using scripts with staxrip is a little bit complicated to newbies or beginners. I hope staxrip could integrate a bit more plugins by default in the future esp I would like to use my beloved detailsharpen. Thanks a lot.

[ChaosKing]

It seems it is aready included https://github.com/staxrip/staxrip/wiki/Tools#g41fun

[stax76]

@jinkazuya

Yes, it appears it was included by an ex maintainer, but it might not be available in the menu, for that there is documentation here:

https://github.com/staxrip/staxrip/w...deo-processing

https://github.com/staxrip/staxrip/w...%29-in-StaxRip

[jinkazuya]

Originally Posted by ChaosKing

It seems it is aready included https://github.com/staxrip/staxrip/wiki/Tools#g41fun

Originally Posted by stax76

@jinkazuya

Yes, it appears it was included by an ex maintainer, but it might not be available in the menu, for that there is documentation here:

https://github.com/staxrip/staxrip/w...deo-processing

https://github.com/staxrip/staxrip/w...%29-in-StaxRip

The github page does not exist anymore for the plugin or script of detailsharpen, which is why if possible, if the plugin is included in the menu in the future release, that would be awesome.

[stax76]

@jinkazuya

Please create a new thread (or use the staxrip thread) asking for code and default values and then add this code and values to the following wiki page:

Tool Update Requests

[jackoneill]

Originally Posted by monohouse

I have some problem with MakeDiff in versions 50,51,52 :x

Script evaluation failed:
Python exception: MakeDiff: both clips must have constant format and dimensions, and the same format and dimensions

Traceback (most recent call last):
  File "src/cython/vapoursynth.pyx", line 2244, in vapoursynth.vpy_evaluateScript
  File "src/cython/vapoursynth.pyx", line 2245, in vapoursynth.vpy_evaluateScript
  File "/mnt/dod/04.vpy", line 11, in <module>
    clip = haf.QTGMC(clip, TFF=True, Preset="Placebo", ShowSettings=False, opencl=False, device=0, TR0=2, TR1=2, SourceMatch=0, TR2=0, Lossless=0, Sharpness=0 )
  File "/usr/lib/python3.6/site-packages/havsfunc.py", line 1281, in QTGMC
    noise = core.std.MakeDiff(clip, denoised, planes=CNplanes)
  File "src/cython/vapoursynth.pyx", line 2069, in vapoursynth.Function.__call__
vapoursynth.Error: MakeDiff: both clips must have constant format and dimensions, and the same format and dimensions

linux, all compiled today, used LD_PRELOAD to test for plugin errors and there are none, without QTGMC the script works, latest mvsfunc and havsfunc

import vapoursynth as vs
import havsfunc as haf

core = vs.get_core()
core.set_max_cache_size(16000)

clip = core.ffms2.Source("//mnt//dod//movie.mkv")
clip = core.fmtc.bitdepth (clip, bits=16, fulls=False, fulld=True, dmode=1)
clip = haf.QTGMC(clip, TFF=True, Preset="Placebo", ShowSettings=False, opencl=False, device=0, TR0=2, TR1=2, SourceMatch=0, TR2=0, Lossless=0, Sharpness=0 )
clip = haf.MCTemporalDenoise(clip,  radius=3, limit=2,  twopass=False, limit2=2,  refine=True, useTTmpSm=True,  stabilize=False, maxr=3, TTstr=3,  chroma=False, MVsharp=False, sigma=0, pfMode=-1,search=3, searchparam=4, pel=4, pelsearch=4, bwbh=512, owoh=256, blksize=4, overlap=2, deblock=False, post=0,bt=4,thSAD=2000, thSAD2=2000, thSCD1=1000, thSCD2=200)
clip = core.deblock.Deblock(clip, quant=20, aoffset=0, boffset=0)
clip = core.fmtc.bitdepth (clip, bits=10, fulls=True, fulld=True, dmode=1)
clip.set_output()

tested several older versions of havsfunc no change :x
I removed this code from src/core/mergefilters.c and it started working fine and the video looks correct

    if (!isConstantFormat(d.vi) || !isSameFormat(d.vi, vsapi->getVideoInfo(d.node2))) {
        vsapi->freeNode(d.node1);
        vsapi->freeNode(d.node2);
        RETERROR("MakeDiff: both clips must have constant format and dimensions, and the same format and dimensions");
    }

strange problem

The error message gives you the line number in havsfunc.py (1281), so right before that line add

print(clip)
print(denoised)

to see what the difference is between those clips.

[Patman]

Hello everybody. I need your help with the following problem:

I have this code in my avs.c:

    h->bit_depth = h->func.avs_bits_per_component(vi);
    FAIL_IF_ERROR( h->bit_depth < 8 || h->bit_depth > 16, "unsupported bit depth `%d'\n", h->bit_depth );
    if( h->bit_depth & 7 )
    {
        AVS_Value arg_arr[2];
        arg_arr[0] = res;
        arg_arr[1] = avs_new_value_int( 16 );
        const char *arg_name[] = { NULL, "bits" };
        AVS_Value res2 = h->func.avs_invoke( h->env, "ConvertBits", avs_new_value_array( arg_arr, 2 ), arg_name );
        FAIL_IF_ERROR( avs_is_error( res2 ), "couldn't convert to 16 bits: %s\n", avs_as_error( res2 ) );
        res = update_clip( h, &vi, res2, res );
    }

I would now like to store a code with an identical function in my vpy.c, but unfortunately it doesn't work that well. I have the following approach:

    h->bit_depth = vi->format->bitsPerSample;
    FAIL_IF_ERROR( h->bit_depth < 8 || h->bit_depth > 16, "unsupported bit depth `%d'\n", h->bit_depth );
    FAIL_IF_ERROR( vi->format->sampleType == stFloat, "unsupported sample type `float'\n" );
    if (h->bit_depth & 7)
    {
        VSMap* args = h->vsapi->createMap();
        const char* error;
        VSPlugin* fmtcPlugin = h->vsapi->getPluginById("fmtconv", core);

        h->vsapi->propSetNode(args, "clip", h->node, paReplace);
        h->vsapi->freeNode(h->node);
        const VSFormat* new_format = h->vsapi->registerFormat(vi->format->colorFamily, vi->format->sampleType, 16, vi->format->subSamplingW, vi->format->subSamplingH, core);
        h->vsapi->propSetInt(args, "csp", new_format->id, paReplace);
        VSMap* ret = h->vsapi->invoke(fmtcPlugin, "bitdepth", args);
        error = h->vsapi->getError(ret);
        if (error)
        {
            FAIL_IF_ERROR(1, "failed to convert node to 16 bits: `%s'\n", error);
            h->vsapi->freeMap(args);
            h->vsapi->freeMap(ret);
            return -1;
        }
        h->node = h->vsapi->propGetNode(ret, "clip", 0, NULL);
        h->vsapi->freeMap(ret);
    }

The aim is to consider certain plugins when processing. With avs it is 'ConvertBits' and with vpy 'fmtconv'. Does somebody has any idea?

[feisty2]

if you have a C++20 compiler (I guess GCC11 is currently the only option), you can play with my C++ wrapper for VS API and coding a plugin is no more complicated than python scripting.

if you need to invoke an external filter as some sort of preprocessing, you can do it directly in the constructor of the filter

MyFilter(auto Arguments, auto Core) {
    auto PreprocessedClip = Core["fmtc"]["bitdepth"]("clip", Arguments["clip"], "bits", 16);
}

alternatively, you can define InitiateCallGraph() if you need to run a more complex filter sequence, where you can invoke external filters for postprocessing, or recursively invoke the filter that is being defined.

[feisty2]

and no error checking of any sort is required (unless you need to change how your filter behaves based on if an evaluation fails), the error message (plugin not found, evaluation failed, or whatever) will automatically propagate to the console (vsedit, vspipe, etc.)

[DJATOM]

Since it's about x264, it uses C compiler, not C++.

[feisty2]

He/She didn't mention what it is that he/she's working on so I wouldn't have known. and since x264 already relies on GCC, it should be pretty handy for him/her to use the C++ wrapper (C99, the standard that x264 seems to be using, is mostly a subset of C++20 after all).

[Patman]

I'll give your wrapper a try.

As DJATOM already mentioned, the files are written in C (C compiler) and are used to directly support avs / vpy scripts in x264.