Thomas de Groot wrote:
> "stbenge" <UN### [at] hotmailcom> schreef in bericht 
> news:4b283e59@news.povray.org...
>> I really need to make a small functioning example to provide along with a 
>> bug report.
> 
> I may be talking nonsense here, but those same features are often also 
> created by WM or GeoControl. I don't remember them from Leveller, but it has 
> been a while since I last used that app. Do you think they have the same 
> origin? Are they the same kind of bug?

I doubt they are the same bug.

Check this out:

height_field{
  png"mountains.png"
  ...
}

It works just fine, always has.

Compare it to this:

height_field{
  function 513, 513{
   pigment{
    image_map{
     png"mountains.png"
     interpolate 2 once
    }
   }
  }
  ...
}

This is when edge artifacts show up. The dimensions are correct. The 
problem happens regardless of image_map attributes such as interpolation 
and repetition.

I would say it's a height_field issue, but I know it's not. I have 
encountered this same problem when using image_maps at other times, like 
for my luminous bloom code.

> In GeoControl I can correct manually with a brush, but in hf's from WM I 
> just cut off a thin slice around from it by using an intersection in POV. 
> Not very elegant as a solution but it works.

Whatever works :)

Post a reply to this message

Kenneth wrote:
> Where was this years ago??!!  You've created some real magic here. It must be
> the coffee--drink lots more, so you can come up with more cool ideas!

Drink more coffee? Okay :)

> I'm really looking forward to trying this out (and looking over your code, to
> glean some ideas from 'the Master.')

Whatever you do, have fun!

Post a reply to this message

"stbenge" <UN### [at] hotmailcom> schreef in bericht 
news:4b2931c8@news.povray.org...
> Check this out:
>
> height_field{
>  png"mountains.png"
>  ...
> }
>
> It works just fine, always has.
>
> Compare it to this:
>
> height_field{
>  function 513, 513{
>   pigment{
>    image_map{
>     png"mountains.png"
>     interpolate 2 once
>    }
>   }
>  }
>  ...
> }
>
> This is when edge artifacts show up. The dimensions are correct. The 
> problem happens regardless of image_map attributes such as interpolation 
> and repetition.
>
> I would say it's a height_field issue, but I know it's not. I have 
> encountered this same problem when using image_maps at other times, like 
> for my luminous bloom code.

Ah, right. I see your point. One way or another I have never noticed this 
difference. Strange indeed. Thanks for drawing my attention to it.

Thomas

Post a reply to this message

stbenge <UN### [at] hotmailcom> wrote:

> I doubt they are the same bug.
>
> Check this out:
>
> height_field{
>   png"mountains.png"
>   ...
> }
>
> It works just fine, always has.
>
> Compare it to this:
>
> height_field{
>   function 513, 513{
>    pigment{
>     image_map{
>      png"mountains.png"
>      interpolate 2 once
>     }
>    }
>   }
>   ...
> }

Yeah, there are some strange goings-on in HFs made from functions.  I think I've
noticed the edge artifacts you mentioned.  But there's also the odd fact that
the applied pattern (or image_map) ends up being flipped upside down(?) like a
mirror-image, AND that the 'origin point' of the pattern/image is out of place.
The usual origin that we all know and love is at <0,0,0>; with a HF function,
the 'origin' is mysteriously at <0,0,1>, and the applied pattern/image
scales out from there. I've come up with simple workarounds to these particular
situations--but, along with the edge artifacts, it's apparent that HF functions
are not working correctly. I suppose, though, that re-engineering the
problematic code (whatever it is!) would cause lots of backward-compatibility
problems with previously-made scenes.

Ken

Post a reply to this message

"Kenneth" <kdw### [at] earthlinknet> wrote:
> ...But there's also the odd fact that
> the applied pattern (or image_map) ends up being flipped upside down(?) like a
> mirror-image, AND that the 'origin point' of the pattern/image is out of
> place...

BY 'image_map,' what I meant to say was "an image_map used as [or used as part
of] the HF function"--not an image_map *texture* applied to the HF afterward
(which is applied as it should be.) Sorry if that confused anyone.

Ken

Post a reply to this message

Kenneth wrote:
> "Kenneth" <kdw### [at] earthlinknet> wrote:
>> ...But there's also the odd fact that
>> the applied pattern (or image_map) ends up being flipped upside down(?) like a
>> mirror-image, AND that the 'origin point' of the pattern/image is out of
>> place...
> 
> BY 'image_map,' what I meant to say was "an image_map used as [or used as part
> of] the HF function"--not an image_map *texture* applied to the HF afterward
> (which is applied as it should be.) Sorry if that confused anyone.

It didn't confuse me, Kenneth :) It's an annoying problem, but easy 
enough to work around. It probably should be fixed, but you're right 
when saying that by doing so, many older scenes will be broken.

Post a reply to this message