Esxi running macos For selur

[shijan]

By the way, not sure if it helps but there is plugin named TemporalDegrain2 with 8-16 int / 32 float http://vsdb.top/plugins/G41Fun (Temporal Degrain Updated by ErazorTT)

---

TimeCube now have no 8-bit banding on grayscale gradient.
TimeCube still produce color glitch artifacts. Artifacts now looks more detailed due high bit depth

Here is idea. LUTs are usually work in RGB and RGB is usually use full range data levels. Maybe we just need somehow in script convert limited to full range levels, aply TimeCube and next convert range back to video/tv/limited levels?

---

Maybe somehow try 32 bit depth for TimeCube?

[Selur]

I neither get glitches feeding limited or full range content to TimeCube in the VM,...
Will think about TemporalDegrain2, btw. the link doesn't really lead to the script, which already comes with Hybrid.
Problem is, looking at the function parameters:

def TemporalDegrain2(clip, degrainTR=2, degrainPlane=4, meAlg=5, meAlgPar=None, meSubpel=None, meBlksz=None, meTM=False,
    limitSigma=None, limitBlksz=None, fftThreads=None, postFFT=0, postTR=1, postSigma=1, knlDevId=0, ppSAD1=10, ppSAD2=5,
    ppSCD1=4, thSCD2=100, DCT=0, SubPelInterp=2, SrchClipPP=3, GlobalMotion=True, ChromaMotion=True, rec=False, extraSharp=False):

There are quite a few with no documentation about which values are allowed for these values and what they do.
-> not adding such a filter while trying to get Hybrid working on MacOS properly, since just looking up all the parameters is a lot of work.
So remind me of it after an official MacOS release of Hybrid on the main download page.

Cu Selur

[shijan]

Ok. TemporalDegrain is not a big problem because there are plenty of other denoisers.

Small idea - maybe mark 8 bit only filters in Hybrid UI with some sign like (!) or (*) or (8-bit) ?

[Selur]

may be I'll add some info to the tool-tip, first everything needs to run,...

[shijan]

Yep, info to the tool-tip sounds great!

---

TimeCubes:
Currently when i apply limited to full (TV to PC) range convert filter before TimeCube artifact is 100% gone.

In older script version when filter was operated in 8 bit depth and i apply limited to full (TV to PC) range convert filter before TimeCube, artifact was almost gone (some glitch dots where still visible). Probably it was due not enough bit and not too precise range transform.

So seems like sort of mixed problem source wrong range and not enough bit depth. LUTs love bit depth.

Also notice fro conversation with developer:
"Also, in case the issue is with the source filter, have you tried using ffms2 or some other source filter?"

---

Also if i use ColorMatrix 601 to 709, Time Cube and RangeConversion filter limited to full (TV to PC) - artifacts are visible again.
Seems like ColorMatrix somehow conflicts with RangeConversion filter.

[shijan]

From sekrit-twc: "Please try the latest TimeCube commit, which may solve the limited range issue."
Compiled version https://www.mediafire.com/folder/wvdlnja...umhhefmoiy
But it didn't help. Same artifacts with latest commit.
Just thinking... Maybe use 32 float for this filter instead of RGB48? 32 float is useful when need to bring back negative (clipped) data.

[shijan]

Bwdif deinterlace updated by l33tmeatwad WORKS
DPID & JincResize seems not available in Hybrid UI yet.

[Selur]

Nope only wrote support for JincResize for Avisynth so far, since someone requested it there.
Never tries DPID.

Just read: https://github.com/sekrit-twc/timecube/issues/4 sounds like sekrit-twc managed to reproduce the issue and might have fixed it.
I keep my fingers crossed.


Cu Selur

[shijan]

It is same version i tested. Have no idea why, but it still produce artifacts in Hybrid. Maybe something in script with bit depth-range transformation?

[Selur]

I doubt that it is Hybrid/Vapoursynths fault since sekrit-twc seems to be able to reproduce the issue.

Cu Selur