Expression out of range error

TangoAlpha

This little blighter has started happening during a render. It only seems to happen at high quality settings (and at 4 hours through a 6 hour render, "words were said", I can tell you! :snake:) Render the same scene at low settings and there's no error.

Of course there's the usual Carrara lack of any useful information about the error.

Through a process of elimination (over many hours) I've tracked it down to one particular replicator. There are three vertex objects in the replicator, each of which renders perfectly happily outside that replicator. They do have fairly high poly counts - 9000-15000 - since they're essentially flowering weeds.

Anyone come across this problem before, or know how to get around it (other than disabling the replicator - I didn't build these models just to throw them away!)?

Thanks :)

TangoAlpha

No one?

This is driving me potty. I think I've found the replicator that causes the problem, so I take it out. Then it happens on a different one. So I take it out. Then it's a different one again and the cycle repeats. I get about 1 successful render in 10 (and each failed render takes between 90 minutes & 4 hours to fail)

I've even tried rebuilding the project from scratch, & rendering on a different computer, for all the good it did.

Maybe I should just throw out Carrara as a waste of time and money and take up freefall - it's much less stressful! There's simply no point in me trying to get this damn thing into the store if the piece of crap program won't even render it. Excuse the language, but I feel like I've wasted the last 6 months on this when I could have been doing something useful. Hell I could have written a new novel - at least books don't crash.

DUDU

Four hours for a render, I suppose that's an animation.
This problem arrives to me sometimes but, no panic !
I always make my renders in images sequences and all starts again normally after a stop of the software.
I think that's the disk cache memory who is full and a re-start of Carrara solve this issue.

Design Acrobat

Since it's an 'out of range' in an animation, I would shorten the number of frames rendered in the render room (frames 0 to 49, eg.) and then do the next 50-100.

That way, I suppose one could see in what sequence the error is occurring. If it's an animation error, then perhaps try changing then play with the tangents (linear, flat, unlink, etc.)

evilproducer

Tim_A said:

No one?

This is driving me potty. I think I've found the replicator that causes the problem, so I take it out. Then it happens on a different one. So I take it out. Then it's a different one again and the cycle repeats. I get about 1 successful render in 10 (and each failed render takes between 90 minutes & 4 hours to fail)

I've even tried rebuilding the project from scratch, & rendering on a different computer, for all the good it did.

Maybe I should just throw out Carrara as a waste of time and money and take up freefall - it's much less stressful! There's simply no point in me trying to get this damn thing into the store if the piece of crap program won't even render it. Excuse the language, but I feel like I've wasted the last 6 months on this when I could have been doing something useful. Hell I could have written a new novel - at least books don't crash.

Sorry, been busy. Are you replicating replicators? I had that once. I used replicators to drive an effect on a plant, and then stuck the plant in a replicator. I did this in C7.2 and had the error. I assumed it was because the math/calculations exceeded the capabilities of the 32 bit Carrara.

Also, how are you setting up your replicators? If you look at my Fantasy Village Terrain over at ShareCG, I broke my replicators up for a couple reasons. The first was to allow the user to "turn off" replicators that weren't going to be in-frame, but also to limit the number of replications per replicator.

TangoAlpha

Thanks for your replies, and sorry I got grumpy - this thing blew my self-imposed deadline of submitting to the store by the end of May completely out of the water.

It's a single frame not an animation. It takes so long because there's a lot of trees & plants (& water) & translucency & transparency & it's on high settings & I only have a humble i5.

My replicators are split both by zone and plant type (trees north, small weeds north, short grass north, etc) - about 20 in all. I did start off with nested replicators for clumps of stuff, but the results were too random and I ended up baking the clumps into real instances which I tweaked. The plant replicators generally have thousands of replicants, but it's the trees that are giving the trouble, and they only have 15 or so replicants in each.

I suppose it could be said that treed are extremely complex (I did an export to vertex model once that came out - eventually - at over 2 million polys), so need to be subdivided into even more replicators. Worth a try, although with so few trees per replicator I could almost just group duplicates (which would also overcome the problem of position clashes)

TangoAlpha

I've now identified 3 separate tree models that cause the problem (they have one thing in common - they use the "hybrid" tree setting). On their own they render fine, and if there's just one visible in the scene that'll render fine too. But more than one . . . kabloowie! It doesn't matter if they are in a replicator with other trees, in their own replicator, or just real instances in the scene, same result.

The expedient thing right now I think, is just to take them out. Maybe I can figure out a solution later and use them in another scene.

Diomede

Wow, important caution. I'd hate to think how long it took you to isolate that particular problem. Thanks for the explanation.