[nand chan]

Originally Posted by leeperry

OMG, he's unstoppable! So I can take a 0-255 gamut mapping 16bit LUT from cr3dlut, merge it w/ a .cal file and output a 8bit LUT to feed rgb3dlut() in ffdshow?

There is currently no utility for changing the bit-depth of a LUT (eg. from 16-bit down to 8-bit) but I can add a program for that by version 0.5.

Note: This would obviously only work for the /output/ bit-depth of a LUT. Changing the input bit depth is unfeasible because anything about 3x8-bit would take such obscene amounts of storage that you couldn't even load it. For example, a fully 10-bit .3DLUT will take 4 GB of memory, more than the average video card can handle. A 16-bit LUT will take several petabytes.

However, what you can do is transform a 8->16 bit LUT down to a 8->8 bit LUT (or the reverse), though at the cost of precision (in the former).

But yes, you can basically take a 8->16 bit LUT gamut mapping (like the type madVR uses) and apply .cal data to it.

I'm asking instead of trying coz my .net crapola is not up to date I'm afraid...but if that's the case I might have an urgent need for a "reset CLUT while playing" in mVR

Agreed on this. It's such a hassle to reset the CLUTs manually using dispcalGUI all the time.

But I guess you can't get it to output a YUY2 LUT, can you?

That wouldn't be in the realm of impossibility as far as I know how YUY2 LUTs work, then I'll also have to do the RGB->LUT conversion since .cal files work in RGB space. (You'd have to choose which reference decoding matrices to use). I wouldn't expect anything resembling YUV support for at least another version or two though, maybe 0.6 or 0.7, mainly because I'm only primarily concerned with madVR which uses RGB LUTs and this simplifies things greatly.