Isotropix Forums

3.5 Scatter Variance Strangeness

You think you may have found a bug? Please post it here.

3.5 Scatter Variance Strangeness

Unread postby tom_tm » Fri Jun 02, 2017 12:31 am

Hi,
I played with the scatter varience feature. I think, the randomness doesn't random that much, especially
if you have a smaller number of variation samples.
Let's say I have a point cloud with 6 points, and I want to have all different, I can't get this to work with the number
of variation samples 6.
Only if I put a very high number of variation samples, I get this effect, but this makes the interface very laggy.

You can download my sample scene with a 250 frames long alembic counter here:
http://www.ayashmi.com/ScatterVarianceTest.zip

Another "issue" I discovered, Clarisse always crashes if I try to scatter with variance an object with fur generator or interpolator applied.
Probably that's not possible right now?

Greets Tom
tom_tm
 
Posts: 189
Joined: Mon Nov 18, 2013 9:20 pm
Location: Switzerland

Re: 3.5 Scatter Variance Strangeness

Unread postby adomon » Fri Jun 02, 2017 10:15 am

Hi Tom

Actually, this is a normal behavior of the time variation feature. When you set the time samples attribute to a value of 6, it just means that the time interval defined by time variance will be sampled into 6 values, and the scatterer will pick a time among these values at random for each point. So indeed, if you have only 6 points and 6 samples, you would be very lucky to have 6 time offset variation on your scatterer. But if you increase the number of points to say 6000, you would nearly have 6 time offsets uniformly and randomly distributed over your set.

As the tool was designed especially to handle crowd generation, it is expecting the scatterer to have significantly more points than time samples when using time variance. But if you want to generate random time variation on a small set of points, you can for example plug a Random texture to the time offset attribute, and also use a Quantize texture if you want to have a total control over the maximum number of time samples allowed.

About the fur generator issue, thanks for reporting it, we will fix the crash for the official release, and also try to support it. For the moment, imported animated alembic fur supports the time variation feature, but unfortunately not fur generated in Clarisse.

Thanks for your feedback

Antoine
User avatar
adomon
 
Posts: 29
Joined: Fri Jul 01, 2016 11:22 pm

Re: 3.5 Scatter Variance Strangeness

Unread postby tom_tm » Fri Jun 02, 2017 12:18 pm

Hi Antoine,
thank you for explaining. I will try with the random node to achieve more variation on smaller crowds.
Variance scatter support for the fur generation/interpolate would be really nice, with the PBR fur I get very nice results.

Best Regards
Tom
tom_tm
 
Posts: 189
Joined: Mon Nov 18, 2013 9:20 pm
Location: Switzerland

Re: 3.5 Scatter Variance Strangeness

Unread postby adomon » Fri Jun 02, 2017 1:06 pm

Sure, it would be nice, we will try to add Clarisse fur support if it is possible before the final release.

Could you please share with us the crashing project with fur generator and time variation ? We are not able to reproduce it for now.

Thanks

Antoine
User avatar
adomon
 
Posts: 29
Joined: Fri Jul 01, 2016 11:22 pm

Re: 3.5 Scatter Variance Strangeness

Unread postby tom_tm » Fri Jun 02, 2017 2:20 pm

Hi Antoine,

here the link to an Alembic with animated Guides (25 frames)
If you change Scatter Time Offset or Scatter Time Variance it will crash instantly, so I let them on Zero.

If I use the alembic guides as hair, it works with variance. Fur Generator and Interpolater let Clarisse crash.

Using Win10 64.

http://www.ayashmi.com/ScatterVariation ... polate.zip

Salut
Tom
tom_tm
 
Posts: 189
Joined: Mon Nov 18, 2013 9:20 pm
Location: Switzerland

Re: 3.5 Scatter Variance Strangeness

Unread postby jandersunstar » Fri Jun 02, 2017 4:11 pm

I am having the same problem with hair_interpolate. So what would be a work around? to have all the hair inside the Alembic?

Hope you can fix this very soon. Cheers!

-Luis
User avatar
jandersunstar
 
Posts: 175
Joined: Fri Jun 06, 2014 9:21 pm
Location: Mexico City

Re: 3.5 Scatter Variance Strangeness

Unread postby adomon » Fri Jun 02, 2017 6:42 pm

Hi

Just to tell you that we have changed a bit the strategy to generate time samples per points to make the feature work with small scatterer. For example, if your support geometry have 6 points and you set time samples to 6, now you will be sure to have 6 different time offset.
This improvement will be available in the next 3.5 release candidate coming very soon.
You will be aware soon if we manage to add the support of fur generator/interpolator with time variation.

Regards

Antoine
User avatar
adomon
 
Posts: 29
Joined: Fri Jul 01, 2016 11:22 pm

Re: 3.5 Scatter Variance Strangeness

Unread postby tom_tm » Fri Jun 02, 2017 11:57 pm

Great, this will make a setup more predictable, the feature itself is very powerful.
Thanks
Tom
tom_tm
 
Posts: 189
Joined: Mon Nov 18, 2013 9:20 pm
Location: Switzerland

Re: 3.5 Scatter Variance Strangeness

Unread postby adomon » Wed Jun 07, 2017 10:34 am

Hi guys

Good news, we have added support of fur interpolate/generator for scatterer time variation feature. It will be available for the next release candidate coming very soon.

Regards

Antoine
User avatar
adomon
 
Posts: 29
Joined: Fri Jul 01, 2016 11:22 pm

Re: 3.5 Scatter Variance Strangeness

Unread postby tom_tm » Wed Jun 07, 2017 12:13 pm

great! my mass is waiting for individual hair :-)
tom_tm
 
Posts: 189
Joined: Mon Nov 18, 2013 9:20 pm
Location: Switzerland

Next

Return to Bugs