Mental Ray = Good , Radiosity = Bad

[Ninja]

After a year of playing around with Radiosity in Max 5 and 6 I've FINALLY realised that Mental Ray is a far more efficient / predictable renderer.

 

The light tracer renderer is excellent for small scenes...especially exteriors..but its ludicrously slow

 

What I find particularly frustrating with Radiosity is the amount of hurdles it places before you can get a decent result...ie ultra clean geometry, annoying "exposure" settings, accurate flux / light data etc etc.

 

It seems I can get far better results in a fraction of the time with MR GI

 

Am I alone in thinking this, or do you guys have a similar tale to tell ?

[abicalho]

I've been working with Radiosity for almost 3 years, with mental ray GI for 4-5, and with Light Tracer for 2.5.

 

I find Radiosity the most predictable and easier to use of the 3.

 

Alexander

[Fran]

I also prefer radiosity, though I certainly feel there is room for improvement. Alex probably already knows what my "whine" list is.

[jucaro]

I for one, find MR more predictable in achieving better results in producing GI renders. Quality-wise, (personally) I even find it better than VRay, although VRay really rocks in the speed department.

[William Alexander]

Re: Mental Ray = Good , Radiosity = Bad

 

Correction both good.

 

Mental Ray can lite a scene with decent GI in short period of time, at the cost of accuracy, however impercievable. It is the absolute best of the three hands down, quality output. My feelings are that MR acutually requires a greater understanding of the true physical properties of light & materialsto get outstanding results. The learning curve is steep.

 

Radiosity, even with photometric lights, is the quickest way to a very good lighting solution. The meshing is simple, it's like a physical screen sifting rocks and sand. The size lets fine sand through or rocks. The regathering options really can reduce the calc times to a good output, by breaking the rocks down. Still uses the scanline renderer and it can be difficult to get the sampled accuracy of MR.

 

As far as clean modeling, it's valuable for many reasons excluding rendering and GI. Even with MR if you ignore clean modeling and go for very high quality results you find that a clean models can count imho.

 

WDA

[bobmaxer]

Good Artist = scanline render = good, mental ray = good, radiosity = good