Maximum Details fro QTGMC

[salvo00786]

Lollo, if I understood correctly, it's better to delete lossless for maximum quality? and for maximum details it's better to use lossless or not? and Sourcematch?

Yes. QTGMC goal is to deinterlace, and it performs better when it does not have to keep the orginal field.

About maximum details the same applies; however, here you do not have to refer to the original field as comparison, but the rebuilt frames in their integrity. In other words, QTGMC builds 2 frames from the interlaced fields, so the only appropriate comparison is between deinterlaced frames obtained with different options. And there you can check what is more performing, also taking into account the quality of the moving objects/parts.
In term of "details versus the original" the concept is the "smoothness" of the outcomes, and here master poisondeathray already addressed versus a parameter choice reducing this action to preserve more elements.

Is not easy to give a general rule, the source really plays a role here. And QTGMC is a complex function!

Ok, and for Sourcematch? What do you think? But if there is too smoothness, probably Topaz Nyx3 will work with less details for denoising. What do you think? Can you link me the thread of poisondeathray, please?

[lollo]

High SourceMatch values should provide better quality (i use 3 most of the times).

For what I tested, if you aim maximum quality Topaz denoising is not as good as a dedicated AviSynth processing, especially for low resolution VHS/S-VHS sources.

Integrated in a AviSynth flow, Topaz may sometimes have a sense as upscaler (replacing nnedi3 or similar), as we showed and said many times.

[salvo00786]

High SourceMatch values should provide better quality (i use 3 most of the times).

For what I tested, if you aim maximum quality Topaz denoising is not as good as a dedicated AviSynth processing, especially for low resolution VHS/S-VHS sources.

Integrated in a AviSynth flow, Topaz may sometimes have a sense as upscaler (replacing nnedi3 or similar), as we showed and said many times.

What denoiser you use in avisynth? Sometimes I use qtgmc with knlmeanscl... do you have other suggestions?

[lollo]

There are many, look in AviSynth wiki and in doom9's forums. Again, the characteristics of the source are important!

One of my favourite is TemporalDegrain2, with a temporal radius adjusted for the source. KNLMeansCL (a spatial denoiser) when needed for some (rare) sources added after TD2.

[DTL2023]

Typical options for better MVs and less bad blurring at MCompensate:
RefineMotion=true. Slow but possibly better search mode is exhaustive (3) for both all levels and pel (Search=3, PelSearch=3).
SearchParam - may be tested a bit higher from default 2 (like 3..4).
BlockSize may be tested 8 or 16. 8 is slower.
Overlap= BlockSize/2 (4 or 8).
TrueMotion - may be tested true or false (depending on content of footage).
PNew may be tested at much lower values (around 5, may be even 0, best value may significantly depend on noise levels).

From v.3.390 https://github.com/realfinder/AVS-Stuff/blob/Community/avs%202.6%20and%20up/QTGMC.avsi there were added more options for better MVs search and possible less blurring at the complex motion areas:
trymany=true (refine around all predictors, slower).
temporal=true (use temporal predictor, will force MT_SERIALIZED for MAnalyse and may be slower).
RefineMotionAll=true (force refine search of all new blocks in MRecalculate). May be about 2x slower depending on thSAD setting and noise levels.
RefineMotionSearch=3 (set exhaustive search for MCompensate in refined MVs search stage)
RefineMotionRadius=8 (set more radius for simple single level search in MCompensate, recommended to be larger or significantly larger than SearchParam for initial search up to several times). May be additionally several times slower (depending on value). Possible range to test may depend on motion speeds and noise levels (about 4 to 40+). Performance drop may be about square of this value.