7 réponses à ce sujet

[s e n]

This may sound stupid but im have issues there, not sure I ever used it in the past but if I did, well now Im having troubles hehe:

simple square mesh with a texture uvxmap applied, animating the rotation gizmo and exporting as animesh: rotates BUT also adds some weirdness, in the shape of some texture "zoom" (scale? UV map value? I tried to block them giving fixed animation keys but still comes the wild weirdness meh!) clues?

note: uvxform rotation just adds the weirdness and doesnt rotate the texture blehs

[OldTimeRadio]

Try TXI with downsamplemin and max set to 0?

That work?

[s e n]

It doesn't

[s e n]

I tried to increase the amount of samples in the animesh and this slowly reduces the amount of "zooming" since, at about 41 samples (1 every 20 frames, I need 800 frams to match the speed of water with a 360 degrees rotation) the issue is almost unpercievable or so. Seems that the animation confuses the rotation and the scale keys, somehow and so, it needs more samples to correctly export the rotation. atm i was using 3 samples animesh and id rather continues use fairly low sampling, not sure if 41 samples for a couple or three dozen tiles for a big narrow river is too much for the engine. using rotation allows me to avoid texture stretching and gives a more smooth look, i think, while the bend method allows me to stay with 3 samples, but the texture stretches towards the closing corner and, to animate the wwww tile with a straight motion, I will need an additional water plane with planar mapping since the water plane with the bending animation isnt planar, so I need another uvmap, hence another mesh

[OldTimeRadio]

Hrm...I'm doing this with the UV XForm modifier and I assume you are too.  Is that right?

The keys that are at the heart of this are just those on that modifier.  Modifier keys.  I see the issue and it's odd because it looks like it's doing one thing (UV clamping to 1) with a small number of samples, but with more samples that (as you said) goes away.

The only thing I can really say at this point, aside from some quirk in how NWMax is processing those keys, is it's how they're being tweened.

It looks like the issue is the result of two things: You're not picking up all the keys you laid down, you're picking up where the values are from those keys when each of your samples are taken- and it's the tweening between those sampled data points that's causing the odd behavior.

(?)

Modifié par OldTimeRadio, 12 novembre 2012 - 10:00 .

[s e n]

ATM I have animated the texture rotation with the rotation gizmo of uvmap, but for linear animation I do use the uvxform modifier, yessir

I think you're right about the tweening thing. right now I reanimated the texture: the loop goes from 11 to 811 frame, and I added 5 keys, (beginning 0 degrees, then 211 90degress; 411 180degrees; 611 270degrees; ends at 811 360degrees) so, with 5 samples(sampling interval is exactly 200), should take exactly the keys I did mark. sadly the zomming stays, so I dont know what to think ehhe

[OldTimeRadio]

s e n wrote...
ATM I have animated the texture rotation with the rotation gizmo of uvmap, but for linear animation I do use the uvxform modifier, yessir

Just an anecdote, but I don't recall having all that great a luck with gizmos like this.  Just tossing that in.

I think you're right about the tweening thing. right now I reanimated the texture: the loop goes from 11 to 811 frame, and I added 5 keys, (beginning 0 degrees, then 211 90degress; 411 180degrees; 611 270degrees; ends at 811 360degrees) so, with 5 samples(sampling interval is exactly 200), should take exactly the keys I did mark. sadly the zomming stays, so I dont know what to think ehhe

I just want to be sure- are you getting that information by selecting your animation name in the Aurora Animesh modifier, clicking the Print Sample Frames button and then checking int he listener?  If not, you'll want to do that to confirm what you think is going to happen and what NWMax thinks is going to happen are the same thing.

[s e n]

yes the listener gives me the same results (i didnt use it to be sincere but same result btw)