Strange behaviour with mixed material

Dear all,

I'm trying to define a mixfunc material to modify a mirror definition, so I can create holes (horizontal bands in this case) in the mirror. This was explained in an old tutorial by G. Antonutto, which now seems lost (at least I couldn't find it online... pillo?).

The complete definition is like this:

## metal

void metal metal_01

0

0

5 .8 .8 .8 .05 0

## mix the metal with void to create holes

void mixfunc strip

4 metal_01 void 'if(sin(Pz/10),0,1)' .

0

0

## alter the reflectance of mirror, reflectance is 0% for holes

void brightfunc band

2 'if(sin(Pz/10),0,1)' .

0

0

## mirror attached to the metal surface with holes

band mirror mix_01

1 strip

0

3 .9 .9 .9

If I create a cube with the "mix_01" material, it looks as expected. Horizontal bands of void in the mirror let the direct light through and the reflected light is also in matching stripes.

However, with the diffuse calculation, light doesn't seem to get through the void part. For example, if I measure irradiance inside the cube with only a glow source outside, I get zero.

The strangest part is that if I look at the model in rvu, it looks fine and there is diffuse light inside... unless I set -aa 0!!!

All other settings seem to make no difference (with ab > 0, of course).

Rpict behaves like rtrace, no diffuse light through.

I've tried setting all rtrace parameters to the rvu defaults (including dr 0), but the result is still nothing.

I'm sure there is an explanation, which is possibly that the material is just too weird, but I'm still confused by the strange difference in behaviour between rtrace/rpict and rvu...

Any ideas?

Thanks!

Santiago

···

____________________________________________________________
Electronic mail messages entering and leaving Arup business
systems are scanned for acceptability of content and viruses

Hi Santiago,

Giulio kindly shared his tutorials and I haven't gotten around to putting them onto radiance-online.org, mostly because I don't know how to modify the website anymore since it was updated. However, I have uploaded this particular tutorial for you to:

  http://www.anyhere.com/gward/pickup/giulio_mirror.tar.gz

However, this is really a situation that didn't occur to me, where you have indirect rays coming from the opposite side. I see now that I should modify a test in the m_mirror.c routine to account for this case. I will send you a modified version of this file in hopes that you can rebuild and test it for me.

Cheers,
-Greg

···

From: Santiago Torres <[email protected]>
Date: June 3, 2016 1:10:50 PM PDT

Dear all,

I’m trying to define a mixfunc material to modify a mirror definition, so I can create holes (horizontal bands in this case) in the mirror. This was explained in an old tutorial by G. Antonutto, which now seems lost (at least I couldn’t find it online… pillo?).

The complete definition is like this:

## metal
void metal metal_01
0
0
5 .8 .8 .8 .05 0

## mix the metal with void to create holes
void mixfunc strip
4 metal_01 void 'if(sin(Pz/10),0,1)' .
0
0

## alter the reflectance of mirror, reflectance is 0% for holes
void brightfunc band
2 'if(sin(Pz/10),0,1)' .
0
0

## mirror attached to the metal surface with holes
band mirror mix_01
1 strip
0
3 .9 .9 .9

If I create a cube with the “mix_01” material, it looks as expected. Horizontal bands of void in the mirror let the direct light through and the reflected light is also in matching stripes.

However, with the diffuse calculation, light doesn’t seem to get through the void part. For example, if I measure irradiance inside the cube with only a glow source outside, I get zero.

The strangest part is that if I look at the model in rvu, it looks fine and there is diffuse light inside… unless I set –aa 0!!!

All other settings seem to make no difference (with ab > 0, of course).

Rpict behaves like rtrace, no diffuse light through.

I’ve tried setting all rtrace parameters to the rvu defaults (including dr 0), but the result is still nothing.

I’m sure there is an explanation, which is possibly that the material is just too weird, but I’m still confused by the strange difference in behaviour between rtrace/rpict and rvu…

Any ideas?
Thanks!
Santiago

____________________________________________________________
Electronic mail messages entering and leaving Arup business
systems are scanned for acceptability of content and viruses

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general

Just a quick follow-up. This was indeed a bug in the m_mirror routine, which is now sorted out. The next distribution should fix it.

Thanks, Santiago!

-Greg

···

From: Greg Ward <[email protected]>
Date: June 3, 2016 1:38:53 PM PDT

Hi Santiago,

Giulio kindly shared his tutorials and I haven't gotten around to putting them onto radiance-online.org, mostly because I don't know how to modify the website anymore since it was updated. However, I have uploaded this particular tutorial for you to:

  http://www.anyhere.com/gward/pickup/giulio_mirror.tar.gz

However, this is really a situation that didn't occur to me, where you have indirect rays coming from the opposite side. I see now that I should modify a test in the m_mirror.c routine to account for this case. I will send you a modified version of this file in hopes that you can rebuild and test it for me.

Cheers,
-Greg

From: Santiago Torres <[email protected]>
Date: June 3, 2016 1:10:50 PM PDT

Dear all,

I’m trying to define a mixfunc material to modify a mirror definition, so I can create holes (horizontal bands in this case) in the mirror. This was explained in an old tutorial by G. Antonutto, which now seems lost (at least I couldn’t find it online… pillo?).

The complete definition is like this:

## metal
void metal metal_01
0
0
5 .8 .8 .8 .05 0

## mix the metal with void to create holes
void mixfunc strip
4 metal_01 void 'if(sin(Pz/10),0,1)' .
0
0

## alter the reflectance of mirror, reflectance is 0% for holes
void brightfunc band
2 'if(sin(Pz/10),0,1)' .
0
0

## mirror attached to the metal surface with holes
band mirror mix_01
1 strip
0
3 .9 .9 .9

If I create a cube with the “mix_01” material, it looks as expected. Horizontal bands of void in the mirror let the direct light through and the reflected light is also in matching stripes.

However, with the diffuse calculation, light doesn’t seem to get through the void part. For example, if I measure irradiance inside the cube with only a glow source outside, I get zero.

The strangest part is that if I look at the model in rvu, it looks fine and there is diffuse light inside… unless I set –aa 0!!!

All other settings seem to make no difference (with ab > 0, of course).

Rpict behaves like rtrace, no diffuse light through.

I’ve tried setting all rtrace parameters to the rvu defaults (including dr 0), but the result is still nothing.

I’m sure there is an explanation, which is possibly that the material is just too weird, but I’m still confused by the strange difference in behaviour between rtrace/rpict and rvu…

Any ideas?
Thanks!
Santiago

Thanks due to Kanae Hagiwara, from our Tokyo office, who found the problem!

And to Greg, of course, for finding the solution!

S.

···

From: Gregory J. Ward [mailto:[email protected]]
Sent: 06 June 2016 17:25
To: Greg Ward
Cc: Radiance general discussion
Subject: Re: [Radiance-general] Strange behaviour with mixed material

Just a quick follow-up. This was indeed a bug in the m_mirror routine, which is now sorted out. The next distribution should fix it.

Thanks, Santiago!

-Greg

From: Greg Ward <[email protected]<mailto:[email protected]>>

Date: June 3, 2016 1:38:53 PM PDT

Hi Santiago,

Giulio kindly shared his tutorials and I haven't gotten around to putting them onto radiance-online.org<http://radiance-online.org>, mostly because I don't know how to modify the website anymore since it was updated. However, I have uploaded this particular tutorial for you to:

                http://www.anyhere.com/gward/pickup/giulio_mirror.tar.gz

However, this is really a situation that didn't occur to me, where you have indirect rays coming from the opposite side. I see now that I should modify a test in the m_mirror.c routine to account for this case. I will send you a modified version of this file in hopes that you can rebuild and test it for me.

Cheers,
-Greg

From: Santiago Torres <[email protected]<mailto:[email protected]>>

Date: June 3, 2016 1:10:50 PM PDT

Dear all,

I'm trying to define a mixfunc material to modify a mirror definition, so I can create holes (horizontal bands in this case) in the mirror. This was explained in an old tutorial by G. Antonutto, which now seems lost (at least I couldn't find it online... pillo?).

The complete definition is like this:

## metal

void metal metal_01

0

0

5 .8 .8 .8 .05 0

## mix the metal with void to create holes

void mixfunc strip

4 metal_01 void 'if(sin(Pz/10),0,1)' .

0

0

## alter the reflectance of mirror, reflectance is 0% for holes

void brightfunc band

2 'if(sin(Pz/10),0,1)' .

0

0

## mirror attached to the metal surface with holes

band mirror mix_01

1 strip

0

3 .9 .9 .9

If I create a cube with the "mix_01" material, it looks as expected. Horizontal bands of void in the mirror let the direct light through and the reflected light is also in matching stripes.

However, with the diffuse calculation, light doesn't seem to get through the void part. For example, if I measure irradiance inside the cube with only a glow source outside, I get zero.

The strangest part is that if I look at the model in rvu, it looks fine and there is diffuse light inside... unless I set -aa 0!!!

All other settings seem to make no difference (with ab > 0, of course).

Rpict behaves like rtrace, no diffuse light through.

I've tried setting all rtrace parameters to the rvu defaults (including dr 0), but the result is still nothing.

I'm sure there is an explanation, which is possibly that the material is just too weird, but I'm still confused by the strange difference in behaviour between rtrace/rpict and rvu...

Any ideas?

Thanks!

Santiago
____________________________________________________________
Electronic mail messages entering and leaving Arup business
systems are scanned for acceptability of content and viruses

I've put up the "Asphalt" example at
http://www.radiance-online.org/learning/tips-tricks/Asphalt. Please let me
know if it looks OK.

Randolph

···

On Tue, Jun 7, 2016 at 7:35 AM, Santiago Torres <[email protected]> wrote:

Thanks due to Kanae Hagiwara, from our Tokyo office, who found the problem!

And to Greg, of course, for finding the solution!

S.

*From:* Gregory J. Ward [mailto:[email protected]]
*Sent:* 06 June 2016 17:25
*To:* Greg Ward
*Cc:* Radiance general discussion
*Subject:* Re: [Radiance-general] Strange behaviour with mixed material

Just a quick follow-up. This was indeed a bug in the m_mirror routine,
which is now sorted out. The next distribution should fix it.

Thanks, Santiago!

-Greg

*From: *Greg Ward <[email protected]>

*Date: *June 3, 2016 1:38:53 PM PDT

Hi Santiago,

Giulio kindly shared his tutorials and I haven't gotten around to putting
them onto radiance-online.org, mostly because I don't know how to modify
the website anymore since it was updated. However, I have uploaded this
particular tutorial for you to:

                http://www.anyhere.com/gward/pickup/giulio_mirror.tar.gz

However, this is really a situation that didn't occur to me, where you
have indirect rays coming from the opposite side. I see now that I should
modify a test in the m_mirror.c routine to account for this case. I will
send you a modified version of this file in hopes that you can rebuild and
test it for me.

Cheers,

-Greg

*From: *Santiago Torres <[email protected]>

*Date: *June 3, 2016 1:10:50 PM PDT

Dear all,

I’m trying to define a mixfunc material to modify a mirror definition, so
I can create holes (horizontal bands in this case) in the mirror. This was
explained in an old tutorial by G. Antonutto, which now seems lost (at
least I couldn’t find it online… pillo?).

The complete definition is like this:

## metal

void metal metal_01

0

0

5 .8 .8 .8 .05 0

## mix the metal with void to create holes

void mixfunc strip

4 metal_01 void 'if(sin(Pz/10),0,1)' .

0

0

## alter the reflectance of mirror, reflectance is 0% for holes

void brightfunc band

2 'if(sin(Pz/10),0,1)' .

0

0

## mirror attached to the metal surface with holes

band mirror mix_01

1 strip

0

3 .9 .9 .9

If I create a cube with the “mix_01” material, it looks as expected.
Horizontal bands of void in the mirror let the direct light through and the
reflected light is also in matching stripes.

However, with the diffuse calculation, light doesn’t seem to get through
the void part. For example, if I measure irradiance inside the cube with
only a glow source outside, I get zero.

The strangest part is that if I look at the model in rvu, it looks fine
and there is diffuse light inside… unless I set –aa 0!!!

All other settings seem to make no difference (with ab > 0, of course).

Rpict behaves like rtrace, no diffuse light through.

I’ve tried setting all rtrace parameters to the rvu defaults (including dr
0), but the result is still nothing.

I’m sure there is an explanation, which is possibly that the material is
just too weird, but I’m still confused by the strange difference in
behaviour between rtrace/rpict and rvu…

Any ideas?

Thanks!

Santiago

____________________________________________________________
Electronic mail messages entering and leaving Arup business
systems are scanned for acceptability of content and viruses

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general