Sorry if this has been asked before, but I have a puzzling problem with an
isosurface I am using.  When I trace the scene I get a warning that the
max_gradient is about 8+ and I need to increase it or I'll get holes.  So I
do, and trace again, but then I get a message saying max_gradient is 7.1
and I should reduce it to make it trace faster.  This happens even though I
haven't changed anything except the max_gradient value.  I can chase the
max_gradient value around indefinitely but no matter what I do it always
gives me either the 'too big' or 'too small' message and keeps changing.
Why does this happen and is there anything I can do to correct it?

L
-

Post a reply to this message

"Loki" <nomail@nomail> wrote in message
news:web.41b6f7a173f8cbc1dc74bbb60@news.povray.org...
> Sorry if this has been asked before, but I have a puzzling problem with an
> isosurface I am using.  When I trace the scene I get a warning that the
> max_gradient is about 8+ and I need to increase it or I'll get holes.  So
I
> do, and trace again, but then I get a message saying max_gradient is 7.1
> and I should reduce it to make it trace faster.  This happens even though
I
> haven't changed anything except the max_gradient value.  I can chase the
> max_gradient value around indefinitely but no matter what I do it always
> gives me either the 'too big' or 'too small' message and keeps changing.
> Why does this happen and is there anything I can do to correct it?
>
> L
> -
>
>

Hi Loki,

Have you tried using the 'evaluate' keyword to 'dynamically adapt the used
max_gradient'?

Chris.

Post a reply to this message

"Chris B" <c_b### [at] btconnectcom> wrote:
>
> Hi Loki,
>
> Have you tried using the 'evaluate' keyword to 'dynamically adapt the used
> max_gradient'?
>
> Chris.

Thanks Chris, actually no I haven't tried that yet.  I'll give it a go,
thanks.

L
-

Post a reply to this message

Wasn't it Loki who wrote:
>Sorry if this has been asked before, but I have a puzzling problem with an
>isosurface I am using.  When I trace the scene I get a warning that the
>max_gradient is about 8+ and I need to increase it or I'll get holes.  So I
>do, and trace again, but then I get a message saying max_gradient is 7.1
>and I should reduce it to make it trace faster.  This happens even though I
>haven't changed anything except the max_gradient value.  I can chase the
>max_gradient value around indefinitely but no matter what I do it always
>gives me either the 'too big' or 'too small' message and keeps changing.
>Why does this happen and is there anything I can do to correct it?

Don't worry about it as long as the surface looks OK. 

What's happening is that the isosurface evaluator makes a series of
guesses as to where a ray might intersect your isosurface, evaluates the
function there, and repetitively refines its guess until it is close
enough. It only observes the gradient values at the particular points
that it happens to evaluate. Increasing your max_gradient value has
shortens the steps between the guesses, thus changing the evaluated
locations, and it now happens to not perform any evaluations in the
region where the gradient is highest.

-- 
Mike Williams
Gentleman of Leisure

Post a reply to this message

Mike Williams <nos### [at] econymdemoncouk> wrote:
> Don't worry about it as long as the surface looks OK.
>
> What's happening is that the isosurface evaluator makes a series of
> guesses as to where a ray might intersect your isosurface, evaluates the
> function there, and repetitively refines its guess until it is close
> enough. It only observes the gradient values at the particular points
> that it happens to evaluate. Increasing your max_gradient value has
> shortens the steps between the guesses, thus changing the evaluated
> locations, and it now happens to not perform any evaluations in the
> region where the gradient is highest.

Thanks Mike, for the explanation.  I'm currently rendering a mid-sized final
test, using the estimation method, and it seems to be working fine (except
for running at 300PPM! - using focal blur and lots of reflections)  I'll
post it up in images when it's finished; it's looking kinda weird, but
kinda cool - like most of my work ;)

L
-

Post a reply to this message