PDA

View Full Version : optimization?



iekoman
07-17-02, 11:06 AM
Can I optimize my models in ZBrush with any way after modelling them? what i have tried my model has always changed back to where it started, ball, box, cylinder or something.

Thanks.

Vikki
07-17-02, 11:23 AM
Hi Iekoman,

In the Tools>Modifiers>Deformation section there is an option for Optimization. You can optimize your model from there using the slider. :D You need to be in edit mode on the model I think also.

Stonecutter
07-17-02, 11:56 AM
Hey I!...You should optimise in small increments, don't just move the slider...
I like to enter '10' at a time, using undo if it goes too far... ;)
You find the 'optimise' function, as Vikki has pointed out, in the Deformations palette...Click on the double-arrows at the top of the pallette to open the second page of deformers...

iekoman
07-17-02, 12:37 PM
Thanks guys.

Wow! There is many "new" features I haven't noticed, and optimization works perfectly.

Thanks again guys

Hanuman
07-17-02, 01:26 PM
Actually there is a very subtle aspect of optimisation that not many people understand. That is that the numbers in the optimisation slider do not refer to the amount of optimisation.

That is whether you optimise at 10 or 100 makes no difference at all. So if you optimise in increments of 10 you are not really optimising by increments of 10

At a setting of 50 everything gets optimised but the bit that is masked is optimised much less than the unmasked area.

This is quite counterintuitive but easy to verify.

The amount on the slider is for when you have applied a mask to limit which part of the mesh is optimised.

The numbers on the mask slider represent the percentage of "respect" that the optimisation will give to the mask. At a setting of 100 then the optimisation process will not respect the mask at all and will optimise as if there was no mask applied. At a setting of 10 the optimisation will pay a lot of respect to the mask and will not effect the mesh under the mask.

Try this, get a default plane and turn on dots so you can see the mesh. Then apply an optimisation settin of "1" three times. Then try the same thing at "100" you will get the same result.

On an umasked object each optimisation, whatever the number, will have the exact oposite effect of "divide".

Stonecutter
07-17-02, 01:50 PM
Actually I don't usually use optimisation and masking together, I just optimise when I want to simplify a model for futher work...

Here's a head, along with a 30% optimised version of the same head for comparison...

http://www2.zbrushcentral.com/zbc_uploads/user_image-1026939019jxl.jpg

Hanuman
07-17-02, 02:33 PM
If you don't use masking then you either optimise or you don't optimise. There is no such thing as 30 percent.

Check out these diagrams as you can see there is no difference between 1 or 100 optimisation.

http://www2.zbrushcentral.com/zbc_uploads/user_image-1026941556rqm.jpg

http://www2.zbrushcentral.com/zbc_uploads/user_image-1026941574qej.jpg

http://www2.zbrushcentral.com/zbc_uploads/user_image-1026941591uid.jpg

aurick
07-17-02, 03:08 PM
Hanuman is correct. Another way to verify this is to make two copies of a tool. If you hold your pointer over the tool in the Tool palette, you will see the number of polygons that it has.

Choose the first one and use Optimize with a value of 10. Now choose the second and Optimize with 100. If you check the polygon counts again, they will be identical.

Without masking applied, any optimization amount will yield the same results.

With masking applied, it is a different matter. This means that you can actually Tool>Modifiers>Selection>Mask All before optimizing to take control of the deformation. When the object is completely masked, the slider will respond like you would expect, with 100 being maximum optimization, and 10 being a very small amount.

Kathy
07-17-02, 03:10 PM
Things you learn even after a year and a half.

Grub
07-17-02, 03:17 PM
Thanks Aurick, thats some really useful info, cheers. :tu:

Stonecutter
07-17-02, 03:21 PM
Ah, Matthew, thank you...That last paragraph turned on the light bulb!! ;) :D :tu:

Mentat7
07-17-02, 03:41 PM
This deserves further investigation...hmmm a way to actually clean up the polygons... You know combined with the shaded polygon mode Pix showed us it will be quite easy to mask certain groups of polygons explicitly for the purpose of optimization!! Many possibilities here! Thank you both Han and Aurick for opening a new chapter in Z-ness...........!

Hanuman
07-17-02, 03:58 PM
<BLOCKQUOTE>quote:</font><HR> When the object is completely masked, the slider will respond like you would expect <HR></BLOCKQUOTE>

Aurick what you have said sounds like it would be so but the truth is that it is even less like what you would expect.

If you mask the object COMPLETELY, that is do not leave even one polygon umasked then the optimisation slider behaves as if it weren't masked at all. That is a default sphere of 8064 polys will go to 2048 polys if it is fully masked whether you give it one or 100.

But here's the strange bit. If you have even a single polygon unmasked then it will behave again very differently. In this case with 99 percent of the object masked you will find that there is very little difference between 1 and 99 optimisation, that is the result will be about 7800 polys. But if you go to 100 then it will optimise fully down to 2048

Strange but true.

I'm beginning to wonder whether this is a little buggy and if it will be different in V 1.5?

Hanuman
07-22-02, 01:10 AM
After further investigation of the optimize with mask applied I have to say that the result is inconsistent. The above information about respecting the mask does not always apply. I would conclude then that this is definately a bit buggy.