> Alain<aze### [at] qwertyorg>  wrote:

>>> HEllo group, and pov ray gurus i am currently stugggling at making a realistic
>>> smoke, but whenever i' scale it up, smoke appears completely black and white and
>>> let's no light throught it, i played with scaling options of emissions but my
>>> objective it to provide some real transluscent smoke like in realist chimney
>>> that pump quite low emissions that when we've seen in the 19 century industrial
>>> chimenys !
>>>
>>> here is the code that is not working when scaling thanhs to lowmuller pages !
>>> when i cale it to about 150 all i get is completly black and white smoke that
>>> not let light through... it now that emission directive shoud de scalled down
>>> (divided by the bontainer scale, but no way)
>>>
>>
>> The aspect of media depends heavily on the lenght of the ray as it
>> travel through it. Scaling up means that the travel lenght increases by
>> the scaling amount. To compensate, you must reduce the density by the
>> same value.
>>
>> If you scale the container up by x150, then you must divide the density
>> of the media by the same amount:
>> media{emission 0.6/150...
>> scattering{1, 1*3/150...
>> Altrernatively, you can divide the values in the color_map by that value
>> instead of dividing the main density of the media:
>>       [0.00 rgb 0]
>>       [0.05 rgb 0]
>>       [0.20 rgb 0.2/150]
>>       [0.30 rgb 0.6/150]
>>       [0.40 rgb 1/150]
>>       [1.00 rgb 1/150]
>>
>>
>> Normaly, extinction should remain at 1, it's default value.
>> Also, normaly, smoke don't emit any light unless it's overheated to some
>> ridiculous temperature. I'd replace the emission by some actual fading
>> light_source near the top of the chemeny if you realy need some glow up
>> there.
>>
>> Some complementary notes about media sampling:
>>
>> The default method is method 3
>>
>> samples only take a single integer value. You can add a second value,
>> but it will be ignored. method 1 and 2 DO use 2 values.
>>
>> confidence is not needed and not used. Only usefull when using method 1
>> ou 2.
>>
>> You use samples 1. It will take only one sample ether as the ray enter
>> the container or at the middle point. Both will give you the wrong
>> result. The minimum value for samples should be 3 and defaults to 10.
>>
>> NEVER EVER use intervals! It defaults to 1 and should be left at that
>> value. It must only be used with method 1 or method 2.
>> "samples 30" is faster than "samples 5 intervals 3" for twice the total
>> samples count. I know, I've done some testing and found that using ONLY
>> samples you may have about 10 times as many samples as you get using
>> samples and intervals for about the same render time.
>> A media with samples 300 rendered in about the same time as the same
>> media with samples 5 intervals 5
>> intervals can also cause artefacts to appear.
>>
>>
>>
>> Alain
>
> Hello Alain, thanks for the reply, it solved most of my issues. there is only
> little problem left, when i scaale up, the bounding sphere appears in a light
> shadowy grey even if a specify correct transmit value, i do not know if it is
> linked to the fog, sky sphere above plan cloud etc...
> if you could shed some light on the issue a would appreciate
>
>

The use of fog is the most porbable cause of what you describe.

With fog, you start evaluating the fog between the camera and the 
container sphere, this gives you a value for the fog effect. You then 
evaluate the fog inside the hollow sphere giving you another value. And 
finaly, you evaluate the fog for the rest of the ray after the sphere up 
untill you hit something else or the background, giving you a third 
value. All those values are added toggether and don't summ up to the 
same as the fog alone when you never encounter the container. This 
especialy true when you reatch the background or a background feature as 
a sky_sphere.

The cure is usualy to use actual media instead of fog... With the 
obvious effect on the render time.


Alain

Post a reply to this message