On 4/18/2024 1:35 PM, Alain Martel wrote:
> Le 2024-04-18 à 08:56, RATH, CHRISTOPHER A a écrit :
>> On 4/17/2024 12:17 PM, Cousin Ricky wrote:
>>> On 2024-04-17 08:28 (-4), RATH, CHRISTOPHER A wrote:
>>>>>
>>>> This is a ways from being done, but I wanted to share the updated image
>>>> with the bubbles fixed.  The balloon texture is still not quite right,
>>>> even when I render with radiosity (which this version does not).
>>>
>>> I doubt radiosity makes any difference here.  I would give Alain's
>>> suggestion of blurred reflection a try.  If you haven't seen it yet, the
>>> technique is at:
>>>
>>> https://wiki.povray.org/content/Knowledgebase:Language_Questions_and_Tips#Topic_13
>>>
>>> To avoid having this technique stall and take forever, use Alain's
>>> modification:
>>>
>>> https://news.povray.org/49b9809b%241%40news.povray.org
>>>
>>> And I would cut the reflection in half or even lower; and if you're
>>> using finish-level Fresnel, remember to use the reduced reflection as
>>> its argument:
>>>
>>> finish
>>> { fresnel 0.5
>>>    reflection { 0.5 fresnel }
>>>    conserve_energy
>>>    ...
>>> }
>>>
>>> The reflection of the water highlights on the right edge of the balloon
>>> looks suspiciously metallic.  Is this not a rubber balloon?
>>>
>> I looked at the code for doing the blurred reflections and started 
>> experimenting with it.  I was able to get a nice blurring mirror using 
>> CSG objects and this technique where I can play with the variables.
>>
>> However, the balloon is an isosurface with a pretty small accuracy 
>> value to remove artifacts, so hitting it with 40X more rays drops the 
>> rendering to a halt.  I can get it to go with a small blur count, but 
>> I don't think that will give me the blur effect.
>>
>> I'm still playing with it.  If I can find a good way to create the 
>> balloon using CSG instead of isosurfaces that may be the better 
>> solution, since I don't need a bumpy surface.
>>
>> -- Chris R
> Then, you need to use the micro-normal technique. Add a normal statement 
> scaled very small.
> 
> normal{granite scale 0.01}
> 
> Then use a reasonable amount of antialiasing. Without antialiasing, the 
> micro-normal will result in a grainy image.
> 
> Optionally : You may add some averaged blurring using a small number of 
> textures. Using averaged textures and micro-normals, you can use about 4 
> to 6 averaged textures and get very good results.

I noticed, (duh), that my complex shape function for the balloons, (A 2D 
polar function I found that I then turn into an SOR in the isosurface), 
looks remarkably like the built-in ovus shape with a slightly stretched 
distance value.  So, I recreated the balloons using that and was able to 
get a reasonable rendering time with a small scale blur and 10 blur 
samples that looks pretty good "in the lab", which included a wavy 
reflective pool to mimic the reflections in the fountain.  It isn't as 
fast as a simple box or sphere, but better than the isosurface.

However, now that I've inserted it into the scene my rendering is 
getting stuck on the balloon again, even with pretty bad anti-aliasing, 
so I will go back to the drawing board with your suggestion above using 
the micro-normals.  I think dropping the fresnel and reflection values 
to 0.3 or lower is at least removing the unexpected shininess, but maybe 
the micro-normals will give enough of the blurring to make it realistic 
enough.

-- 
-- Chris R

Post a reply to this message