Show Posts

241

[Max] Resolved Feature Requests / Re: ACEScg

« on: 2021-12-15, 12:36:58 »

Quote from: olivierc on 2021-10-01, 13:29:36

Hi Guys,

So what's up with ACES implemented in Corona 7?

Vray does it.. :/

Cheers,
Olivier

sure. VRay does it. please have a look at some tests I did recently (and consider the fact Corona has wide rgb). internal colour space comparison sRGB vs ACEScg. the only adjustment between the two is white balance and exposure (about 1500K and 1 stop):

242

[Max] I need help! / Re: Identifying this kind of aesthetic

« on: 2021-12-08, 12:11:32 »

I like it. subtlety lets the composition play first fiddle.

243

[Max] General Discussion / Re: Tonemapping - Plz Halp

« on: 2021-12-08, 00:05:36 »

Quote from: burnin on 2021-12-07, 22:46:57

PS
What I want to convey is, first a proper IRL reference must be established and only after a search for best CGI practice can commence.

heh, and I was going for something simpler like: there is nothing wrong with tone mapping in Corona.

244

[Max] General Discussion / Re: Tonemapping - Plz Halp

« on: 2021-12-07, 12:09:02 »

apologies, forgot about 2.2 >> srgb.
corrected E and F from CIE (rough 2.2 >> srgb done in photoshop).

245

[Max] General Discussion / Re: Tonemapping - Plz Halp

« on: 2021-12-07, 11:23:03 »

Quote from: cjwidd on 2021-12-07, 02:51:56

@piotrus3333 do you have LUTs for these, I'd like to test them out(?)

this is output from CIE with mP_LogCorona2VRayAMPAS.3dl and 1 stop of exposure correction (I did the same in VRay - srgb raw was brighter than acescg).
conversion was: VRay's raw acescg >> srgb >> CIE >> -1 exp >> log lut

edit: I mean VRay's raw acescg >> lin srgb >> CIE (-1 exp, log lut)

246

[Max] General Discussion / Re: Tonemapping - Plz Halp

« on: 2021-12-07, 09:27:34 »

Quote from: cjwidd on 2021-12-07, 02:51:56

@piotrus3333 do you have LUTs for these, I'd like to test them out(?)

these are as plain as possible. just raw srgb and acescg with tone mapping. ampas curve for srgb and default srgb transform for acescg. so pretty much the same curves. fancy lens to make them look a bit less synthetic.

247

[Max] General Discussion / Re: Tonemapping - Plz Halp

« on: 2021-12-06, 14:41:42 »

Quote from: lupaz on 2021-11-30, 17:22:11

My only expectation is that colors don't get so saturated on hotspots. I believe that was part of the conversation here on this thread. For example, when the sun hits wood you get a yellowish halo around the white area. Yes, you can lower the overall saturation, but you're affecting the entire image. Reinhardt or crunching the whites don't seem to work as if you were working with an HDR. I may be wrong and I'm far from an expert in the subject.

hello again.
regarding the burning need for ACES in Corona - at least for those not in film/vfx - please spare a moment and have a look at these two. the difference is internal color space: ACEScg vs sRGB. so wide gamut vs, well, narrower gamut. Corona's internal is wide from day one (remember this one from 2014: "wow, Corona has such a nice natural lighting").

so how do these two look?

248

[Max] General Discussion / Re: Tonemapping - Plz Halp

« on: 2021-12-03, 00:23:28 »

Quote from: cjwidd on 2021-12-02, 23:56:03

@piotrus3333 are you sure that the VFB doesn't render .cube vs .3dl differently? It looks like there is some color burn, or something, using the .3dl in VFB LUT, compared to .cube(?) Testing mP_Reinhard2FstormDefaults_v1.3dl.

this one is quite saturated. format shouldn't matter but I've never checked that in Corona to be honest. first luts I posted were linear in 3dl format and 2.2 in cube. recent ones are all 2.2 in 3dl. corona does backflips with gamma in vfb.

249

[Max] General Discussion / Re: Tonemapping - Plz Halp

« on: 2021-12-03, 00:09:59 »

one more.
since ampas seems to be the winner - this is more precise version of VRayAMPAS lut. check input LUT is in Log color space and leave HC at default (or you can actually use it to move the white point up or down).

250

[Max] General Discussion / Re: Tonemapping - Plz Halp

« on: 2021-12-02, 20:46:08 »

Quote from: dj_buckley on 2021-12-02, 20:03:00

Awesome, is this the same as the hardcoded LUT that FStorm uses as a base?

heh, no. not the same. Fstorm’s curve has a few adjustable parameters and is not very straightforward. this is fairly decent aproximation of the defaults. you can send low res raw exr from corona and I can show you the difference. this lut relies on reinhard starting point so images with lower dynamic range get more noticable difference.

251