AccuRender Studio for AutoCAD

[Roy Hirshkowitz]

We're up to Build 0082.  This new major release of the AccuRender and nXt product line now includes a plugin that works entirely within AutoCAD just like previous versions. It's got a ton of new features and is available for free for a while.  Take a look and let me know what you think.

See the Download->AccuRender Studio menu items for more details.

[Roy Hirshkowitz]

I'm still mulling that one over.  I took it out because it's very confusing to people about when to use mapping or when to use a local transform in the material-- and how the two interact.  I can put a rotation back in pretty easily, a lot of other rendering packages include this ability.

BTW-- how many passes (Samples Per Pixel) were reported in your farm rendering?  I ended up with 5351 on my 1 hr x 32 core cloud rendering.

[Peter Milner]

7302 SPP - 40 minutes on 80 cores.

[Peter Milner]

I got the decals working, but if I then edit one of the materials, the decals disappear when the model reloads. If I then close ArStudio and re-launch it, they're back again.

[Roy Hirshkowitz]

I'll give that a try later this morning.  I may ask for more details or a model next week.

[Roy Hirshkowitz]

Posted build 0085 which addresses this.

[Peter Milner]

Everything seems to be working fine now, except that spotlights don't glow. In other words, the light source is not visible.

[Peter Milner]

It seems to be connected to the beam angle.

[Roy Hirshkowitz]

Two factors will affect how bright the lights appear.  The angle that you're looking relative to the beam angle, as you've already noticed.  The smaller the beam angle the less bright the light will be at glancing angles.

The other factor is the size of the light, in particular it's area.  Decreasing the radius of disk by a factor of 2, for example, will cause the surface brightness to increase by a factor of 4.

[Peter Milner]

This is where glow would be useful.

I've noticed that the beam angle in AR Studio is double the value in nXt.

Is that because nXt is reading the angle from the vertical whereas Studio is reading the angle across the whole beam?

This would explain why the spotlights virtually disappear at a beam angle of 25.

[Roy Hirshkowitz]

The actual angles internally are the same-- nXt just expresses it as a half angle.

Glow is actually a really big cheat.  It's a light that appears when we look directly at it or see it in a mirror.  Which begs the question of what happens when we see it in a glossy surface?  In nXt I cheat by reducing the effect as the surface becomes more diffuse by some arbitrary amount.  I'm no longer willing to do this in the current engine-- anything that glows is a light source.

However, nXt spot lights had a setting called "diffuse" (I believe) which allowed a portion of the light to be transmitted diffuse (beam angle = 120).  This could accurately model a recessed fixture where there is some diffuse leakage due to reflection from the can.  I can add this back pretty easily-- it's really just a more complex distribution.

[Peter Milner]

You can see from these photos that distant spots look just as bright as the close ones:

However with ArStudio the distant ones are dimmer because of the viewing angle:

I agree that what you say is technically accurate, but in the real world it doesn't actually work that way

[Roy Hirshkowitz]

They're quite a bit dimmer in the photos-- you're also probably seeing the sides of the cans in the more distant ones.  Like I said, I can do a more complex distribution from the light that will mimic this-- I can even draw a diagram of the distribution in the interface and have the default include some of this leakage.  What I can't do it cheat.  (I will check and make sure my calcs are correct before embarking on any of this-- it's also possible that I've made a mistake here, either in the basic translation-- or in the interpretation of nXt's beam angle.)