FluidRay RT real-time renderer 1.0.3 released

[nicolacandussi]

We just released FluidRay RT 1.0.3, you can get it here.

In this release:

 

 

• Major speed improvements and bug fixes
  
• Much better Rhino import, especially for high complexity models
  

 

[John Dollus]

Speed improvements? How can you improve on real-time?

With everything being called real-time when it isn't, should we start calling normal time 'hyper-time'?

 

Caffeine-induced snark aside, it looks like a fine GPU engine. The marketplace is getting crowded with unbiased solutions though.

[nicolacandussi]

FluidRay RT is actually completely CPU based. Here's why: http://www.fluidray.com/why-no-gpu/

[John Dollus]

right. My point still stands, though, regardless of which processor is handling the calcs.

[Devin Johnston]

I prefer interactive renderer instead of real time, John's right it's kind of misleading although this seams to be the label the industry had decided to go with.

[numerobis]

although this seams to be the label the industry had decided to go with.

not all of them...

yes, interactive sounds better

[John Dollus]

To me, the most important aspect of any engine is the flexibility of the output. The feature list on your site is a bit vague in this regard. For example, what passes can be extracted in post production and by which apps? Not every application works the same so the same output may have to be treated differently in Nuke than Photoshop. Some screenshots showing how your output is composited in a few of the top applications would go a long way toward building confidence in how your solution fits within existing pipelines.

[nicolacandussi]

Hi John,

Thanks for the suggestion. Since you seem to care a lot about the flexibility of the output, I'd like to ask you a couple of questions on the subject.

Are you aware of Light Path Expressions? How many people do you think are able to take advantage of them?

What outputs do you think are a must-have?

[John Dollus]

Yes, but I'm never faced with adjusting individual light contributions/adjustments at the post stage. That's all pretty much balanced before final render. I can certainly see where that would be valuable in, say, product design but I am faced with changing design characteristics hours before a presentation as opposed to environmental. I would much rather have great support for deep compositing to help with adding bits and pieces in post.

Most of the post headaches in my particular area have to deal with replacing entire elements or adjusting surface materials so object passes, UV's, material ID's, normals, zdepth etc are far more valuable. Pretty much look at the render elements available through vray and you will have it covered. They go a little overboard imo but it's more valuable to have more than I need than not have the one that is critical.

 

Even using the traditional compositing process of shadows, spec, diffuse etc is often more trouble that it's worth simply because we rarely, if ever, need to quickly adjust an image based on those parameters. It's always scene based changes so having to update all of those elemental passes just to update a single piece of furniture or a column design leads to massive file sizes and needless complexity.

[nicolacandussi]

We just released FluidRay RT 1.1, you can find it here.

What's new:

 

- Improved convergence speed

- Ability to hide objects/materials

- Zoom extend selected

- Major bug fixes