Well, a company that has a website that says nothing about their product other than to contact their sales team is somewhat suspect/suspicious to me. At least explain what your product is and does, or you look like a cult.
That being said, it’s all detailed in their patent, which I just read: US9438821B1 - Method for applying multi-layered film grain and texture mapping to a digital video image - Google Patents
They take issue with the fact that digital grain is not authentic, and that other common grain methods (e.g. Cinegrain) apply a singular exposed scan onto the image, and could have a repeating pattern.
I think the repeating pattern is solved or easily avoidable. That’s a red herring.
In their process they use up to 8 separate grain assets and repeat the process at different exposures (3 are for the film curve toe, slope shoulder; the other 5 are a simple band pass as we know from many color apps, including Flame’s tone operator in Image).
Lastly they allow the operator to mask and manipulate the grain. And are advocating that the regraining process be done during color correction.
We can argue about the perfect place in the pipeline to regrain. We just had the whole thread on whether you color before or after comp. So it’s not that easy to answer.
Does up to 8 luma separated grain layer look better than a single Cinegrain plate? I’m sure there is some material where that is true. Maybe three is sufficient in most cases.
Do you need their ueber expensive tool? I think if it’s important enough for you, and you have render cycles to spare and schedule real-estate for it, it’s totally fine to set that up yourself in Nuke, Flame or whereever.
The bigger issue might be getting a handle on separate exposure scans, and solving the problem that the most frequently used blend mode (overlay) works best at 0.5. But a bit of GLSL code fixes that fast.
As @aaronneitz said - this is probably in the 99.999th percentile of things that will elevate your image further. Everything else before that has to be to perfection. Best script ever, only A-list talent in front of camera, and in a good mood. Fantastic DP with unlimited budget. Perfect weather. blahblahblah…
The end result is only as good as the weakest link. The difference between 1, 5, or 8 grain luma separations is unlikely your downfall.
Spend the money on hiring better talent, feeding the crew, get an extra set day for a few extra takes or whatever. You’ll be glad you did.
PS: The argument/value proposition is also likely derived from the shit quality grain first and second generation digital cinema cameras came with. If you look at an Alexa35 or Sony Venice, the grain they produce is quite pleasing. Easy to extract via DasGrain (or Flame’s new Regrain node) and then restore. And then you don’t have to monkey with weird grain arithmetic, the sensor applied it based on exposure. It’s real and every frame is bespoke.
