Perhaps I'm not understanding the media sampling parameters quite right, 
but here goes...

I have a block like this:

interior
{
   media
   {
     scattering {1, 1}
     density {Fn6(x, y, z)}
   }
}

This is supposed to be the clouds for my sky (which takes up about 50% 
of the image with the current camera angle). It renders at about 3,000 
pixels/second.

Now, if I add "intervals 10" (or indeed, any other setting) the render 
slows down to 1,800 pixels/second. And yet, the documentation claims 
that 10 is the default value... Does POV-Ray do something special if no 
explicit value is given or something? (The images produced are markedly 
different as well by the way.)

Post a reply to this message

Orchid XP v2 <voi### [at] devnull> wrote:
> And yet, the documentation claims 
> that 10 is the default value...

  The documentation is old. It's the default value when using method 1, but
since the default method was changed to 3 those other defaults changed too,
but nobody remembered to update the documentation.

  I have seen the defaults for method 3 but I don't remember them. It might
have been 1 for intervals.

-- 
                                                          - Warp

Post a reply to this message

>>And yet, the documentation claims 
>>that 10 is the default value...
> 
> 
>   The documentation is old. It's the default value when using method 1, but
> since the default method was changed to 3 those other defaults changed too,
> but nobody remembered to update the documentation.
> 
>   I have seen the defaults for method 3 but I don't remember them. It might
> have been 1 for intervals.

Ah yes - isn't method 3 the adaptive subdivide one? Would make sense to 
have a single interval for that...

Just tried rendering with intervals=1, and it does indeed produce the 
same results as with no specification. (Same speed also.) Any idea what 
the default samples setting is? (If not, I'll play tomorrow and see if I 
can find it. It'll be a lowish number...)

Damn... somebody needs to update the docs! LOL. (Ah, but that would 
require a *person*. And not just any person - but a person who has the 
necessary information to do it. And the time... I imagine nobody is even 
going to look at such a thing until the current beta goes release.)

Post a reply to this message

Orchid XP v2 nous apporta ses lumieres en ce 14/04/2006 17:13:
>>
>>
>>
>>   The documentation is old. It's the default value when using method 
>> 1, but
>> since the default method was changed to 3 those other defaults changed 
>> too,
>> but nobody remembered to update the documentation.
>>
>>   I have seen the defaults for method 3 but I don't remember them. It 
>> might
>> have been 1 for intervals.
> 
> 
> Ah yes - isn't method 3 the adaptive subdivide one? Would make sense to 
> have a single interval for that...
> 
> Just tried rendering with intervals=1, and it does indeed produce the 
> same results as with no specification. (Same speed also.) Any idea what 
> the default samples setting is? (If not, I'll play tomorrow and see if I 
> can find it. It'll be a lowish number...)
> 
> Damn... somebody needs to update the docs! LOL. (Ah, but that would 
> require a *person*. And not just any person - but a person who has the 
> necessary information to do it. And the time... I imagine nobody is even 
> going to look at such a thing until the current beta goes release.)
intervals =1, no need to change it. Using a to large value can cause artefacts.
samples =3, minimum value, can be set higher as needed.
You don't need to use a spread of values like samples 10,100, just use a single value.

-- 
Alain
-------------------------------------------------
You are always doing something marginal when the boss drops by your desk.

Post a reply to this message

> intervals =1, no need to change it. Using a to large value can cause 
> artefacts.
> samples =3, minimum value, can be set higher as needed.
> You don't need to use a spread of values like samples 10,100, just use a 
> single value.

Thanks.

Post a reply to this message

> intervals =1, no need to change it. Using a to large value can cause 
> artefacts.
> samples =3, minimum value, can be set higher as needed.
> You don't need to use a spread of values like samples 10,100, just use a 
> single value.

For reasons I don't really understand, if I don't specify the number of 
intervals, no other settings seem to have any effect at all. However, if 
I set intervals=2, now changing samples has an effect. (And yes, low 
intervals and high samples does look a lot better - and renders faster too.)

Post a reply to this message