POV-Ray : Newsgroups : povray.binaries.images : First attempts at blurred reflection... Server Time: 19 Apr 2014 17:02:44 GMT
  First attempts at blurred reflection... (Message 1 to 6 of 6)  
From: Jörg 'Yadgar' Bleimann
Subject: First attempts at blurred reflection...
Date: 28 Nov 2011 01:05:24
Message: <4ed2de54@news.povray.org>
High!

As I want to have floor tiles with slightly disturbed reflection in my 
Port Whatmough villa, I started to play around with the blurred 
reflection hack described here: 
http://wiki.povray.org/content/Knowledgebase:Language_Questions_and_Tips#Topic_13

Slightly disturbed reflection works quite well (see my test table 
below), but I wonder how to achieve that strong, smooth blur I've seen 
in so many scenes during the last years - using a large scaling factor 
for the normal (e. g. 1000) yields weird results with high turbulence 
values, while using small scaling factors like 0.001 gives something 
remotely resembling my notion of blurred reflection, albeit still quite 
grainy.

Should I try an even smaller scaling factor (0.0001 or so) and/or using 
stronger anti-aliasing?

See you in Khyberspace!

Yadgar


Post a reply to this message


Attachments:
Download '2011-11-27 die, takes 12 to 61 - blurred reflection test.png' (2702 KB)

Preview of image '2011-11-27 die, takes 12 to 61 - blurred reflection test.png'
2011-11-27 die, takes 12 to 61 - blurred reflection test.png


 

From: clipka
Subject: Re: First attempts at blurred reflection...
Date: 28 Nov 2011 03:05:14
Message: <4ed2fa6a@news.povray.org>
Am 28.11.2011 02:02, schrieb Jörg 'Yadgar' Bleimann:

> Slightly disturbed reflection works quite well (see my test table
> below), but I wonder how to achieve that strong, smooth blur I've seen
> in so many scenes during the last years - using a large scaling factor
> for the normal (e. g. 1000) yields weird results with high turbulence
> values, while using small scaling factors like 0.001 gives something
> remotely resembling my notion of blurred reflection, albeit still quite
> grainy.

It looks to me like you're using just one single bumpy texture, rather 
than multiple ones - or you've forgot to translate each of them 
randomly; you need either multiple layers (e.g. 40 as in the example), 
or - my personal preference - some slight but high-quality focal blur 
for good oversampling (anti-aliasing doesn't quite cut it), in which 
case a single small-scale bumpy texture will suffice.


Post a reply to this message

From: Jörg 'Yadgar' Bleimann
Subject: Re: First attempts at blurred reflection...
Date: 28 Nov 2011 13:10:54
Message: <4ed3885e$1@news.povray.org>
Hi(gh)!

On 28.11.2011 04:05, clipka wrote:

> It looks to me like you're using just one single bumpy texture, rather
> than multiple ones - or you've forgot to translate each of them
> randomly; you need either multiple layers (e.g. 40 as in the example),
> or - my personal preference - some slight but high-quality focal blur
> for good oversampling (anti-aliasing doesn't quite cut it), in which
> case a single small-scale bumpy texture will suffice.

In fact I did so... here's the code:

#declare N_Blurred =
normal
{
   bumps pow(10, div(clock, 5)-2)
   turbulence pow(10, mod(clock, 5))
}

#declare F_Metallic =
finish
{
   diffuse 1
   brilliance 0.7
   specular 1
   roughness 0.001
   reflection 0.5
   metallic
}

#declare T_Blurred_Gold =
texture
{
   average
   texture_map
   {
     #while (i<40)
       #declare rn=seed(0);
       [1  pigment { color rgb <0.8, 0.75, 0> }
           finish { F_Metallic }
           normal { N_Blurred translate <rand(rn),rand(rn),rand(rn)>*100 
scale 0.001  } ]

       #declare i=i+1;
     #end
   }
}

// end of code

See you in Khyberspace!

Yadgar


Post a reply to this message

From: clipka
Subject: Re: First attempts at blurred reflection...
Date: 28 Nov 2011 15:19:11
Message: <4ed3a66f$1@news.povray.org>
Am 28.11.2011 14:10, schrieb Jörg 'Yadgar' Bleimann:
> Hi(gh)!
>
> On 28.11.2011 04:05, clipka wrote:
>
>> It looks to me like you're using just one single bumpy texture, rather
>> than multiple ones - or you've forgot to translate each of them
>> randomly; you need either multiple layers (e.g. 40 as in the example),
>> or - my personal preference - some slight but high-quality focal blur
>> for good oversampling (anti-aliasing doesn't quite cut it), in which
>> case a single small-scale bumpy texture will suffice.
>
> In fact I did so... here's the code:

> #declare T_Blurred_Gold =
> texture
> {
> average
> texture_map
> {
> #while (i<40)
> #declare rn=seed(0);

Heh! You're seeding the random number generator inside the loop, giving 
you the very same sequence of "random" numbers (and hence the same 
translation vector) over and over again. Don't do that :-P


Post a reply to this message

From: Jörg 'Yadgar' Bleimann
Subject: Re: First attempts at blurred reflection...
Date: 28 Nov 2011 22:23:20
Message: <4ed409d8@news.povray.org>
Hi(gh)!

On 28.11.2011 16:19, clipka wrote:

> Heh! You're seeding the random number generator inside the loop, giving
> you the very same sequence of "random" numbers (and hence the same
> translation vector) over and over again. Don't do that :-P

Oh yes... it seems to be not a matter of stupidity, but of impatience!

Here's the correct table!

See you in Khyberspace!

Yadgar


Post a reply to this message


Attachments:
Download '2011-11-28 die, takes 62 to 111 - blurred reflection test.png' (2740 KB)

Preview of image '2011-11-28 die, takes 62 to 111 - blurred reflection test.png'
2011-11-28 die, takes 62 to 111 - blurred reflection test.png


 

From: clipka
Subject: Re: First attempts at blurred reflection...
Date: 28 Nov 2011 23:31:06
Message: <4ed419ba$1@news.povray.org>
Am 28.11.2011 23:19, schrieb Jörg 'Yadgar' Bleimann:

> Here's the correct table!

Yup, looks more like it :-)

As you may notice, turbulence doesn't add anything useful to it; neither 
does a bump value of >> 1. The latter is because it causes lots of rays 
to get reflected "into" the geometric object, which doesn't make sense, 
so POV-Ray's algorithm replaces them with rays reflected according to 
the unpertubed normal instead.


Post a reply to this message

Copyright 2003-2008 Persistence of Vision Raytracer Pty. Ltd.