Subsurface Scattering

[Roy Hirshkowitz]

Coming soon-- path tracer only.  Architectural applications are probably limited-- but it's interesting new territory anyway.

 

[Rich Rosemann]

So, this could be the answer for people who want to show scattered evening light streaming in thru a window, or a port hole, or a graveyard gate, etc. to simulate an eerie atmosphere?  Not that I would be interested in such a project, of course.

 

Rich Rosemann

 

[Roy Hirshkowitz]

Yep.

[Ros. Riha]

There is an old story regarding Maxwell. I do not know, whether it is trustworthy, but anyway - it goes like this:

At the beginning, Maxwell engineers did create an engine, that was not 100% accurate/realistic - it was "only" 95% realistic, but it was very fast. Engineers were not satisfied, because their goal was to achieve 100%... so they changed an engine in a way, that allows to achieve 99% realism, but it is 10times slower than the original one.

Be it fairy tale or not:

1. In our office, we are more than happy with what nXt+ACA combo allows us to achieve in terms of realism - easy to set up, easy to use (regardless small hurdles in the UI here and there - we believe these will be fixed up along the way).
2. In our office, we value the pathtracer very highly - big improvement over the packet tracer... once pathtracer was implemented, the time, required to set a rendering up did decrease substantially - pathtracer is able to give an overview of the final result much faster than the packet tracer.
3. In our office, we miss all these "glossy" and "ambient" switches - in old packet tracer times these switches did allow us to trade realism for speed - which is sometimes very important... so we would welcome if such switches were re-introduced into pathtracer as well - would it be possible, Roy?
4. What will be the new engine, like, Roy? Would it allow to choose different level of realism and to trade them for speed?

[Roy Hirshkowitz]

It's a fairy tale.

 

Nope-- no switches like glossy and ambient for the path tracer.  The big one that I go back and forth on is daylight portals.  This would make most daylit interiors go quite a bit faster-- at the expense of simplicity.  I haven't completely made up my mind on where I stand on this issue...

 

The new engine isn't even coded yet-- it's pure vapor.  But it will be designed to  move even more in the direction of simplicity-- so no junk-- no.  The code is also simpler.  (I'm trying to fit the core code on a postcard.)

[Kevin Lockwood]

Will that be included as a standard upgrade or a paid upgrade? Just curious.

[Roy Hirshkowitz]

Kevin-- the AutoCAD product is my lab product-- you get everything I'm working on for free.  (If I don't like it I won't tell you how to access it-- IOW I have no clue yet on whether the newly imagined engine will even turn out to be practical.)

[Peter Milner]

The nice thing with Accurender is that you have the choice.

For speed use the the packet tracer with all the different shaders to simplify materials.

For accuracy use the Pathtracer.

It may be worth keeping 2 separate libraries of textures optimised for each rendering engine.

[Chas / Aug Posekany]

I have noticed the two engines treat transparant materials quite a bit different. The path tracer tends to render with more transparency than the packet tracer using the same material and set-up.

[Ros. Riha]

Our experience is quite different - we usually spend plenty of time playing with lighting setup etc. - pathtracer is much faster there, IOW it gives you an idea how the final resolved image would look like much faster/earlier than packettracer.

Also - generally speaking - pathtraced images are usable sooner than packettraced ones...

All in all, pathtracer proves to be faster - it gives you usable result sooner. That is why I am asking for these switches to speed pathtracer up even more.