Test with interior lighting only

[Peter Milner]

I thought I'd compare the 3 engines using only interior lighting (no daylight).

The time is in minutes and seconds.

[George Ioannidis]

Really close to the third image previously posted. (engine 2)

I would be more than happy with such result.

[Peter Milner]

Here's a similar test with daylight.

Each was run for about an hour and a half.

Engine 0

Engine 1

Engine 2

[Rich Rosemann]

Big difference between interor and exterior lighting differences.  All three are more acceptable, except that E2 appears softer, or grainer or...something. 

 

[George Ioannidis]

I've got different result. Test scene with just two IES lights. Engine 1 does nice soft shadows, but engine 2 produces some strange rectangular light pattern on the roof. Lights are custom made with IES light inserted in place of bulb. Glass cover tagged not to produce shadows. Can it be some kind of caustics? Engine 1 time near 1 hour, engine 2 time about 25 minutes.

[Roy Hirshkowitz]

It's absolutely a caustic.  You'll need to either model the light without ies data-- OR-- model the glass with IOR = 1.0;

[George Ioannidis]

I'll try that (set IOR to 1) Roy, because I like the way IES lights do light patterns. Look very realistic (to photographer's eye). Thanks.

[Roy Hirshkowitz]

Try both ways, if you have time.

[George Ioannidis]

Well, caustics go away with IOR set to 1 with IES lights, but glass cover almost disappears and lot of fireflies appear. In both machines. I will try same renders without IES lights later.

[Peter Milner]

And now here's Engine 4.

1 hour 15 minutes and 1670 passes.