 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hi there, I noticed a couple problems with the transmit feature. I made a simple
scene demonstrating the problems. I also made two demonstration images. One
image is made with radiosity and one without. The scene consists of test objects
(cylinders)with varying amounts of transmit. Three objects are placed slightly
in front of a white plane and three cylinders are placed with the ends extending
slightlly into the plane.
The image without radiosity,
http://mysite.verizon.net/moronicon/Transmit-NoRadiosity.jpg
shows that the cylinder with transmit set to .8 and placed slightly in front of
the plane (right middle object) doesn't appear tinted enough from the sides, but
does appear tinted correctly from the top end.
The image with radiosity,
http://mysite.verizon.net/moronicon/Transmit-Radiosity.jpg
shows that trasmit values don't appear to be taken into account properly when
calculating radiosity.
Here is the scene used to generate the images:
#local UseRadiosity = no; //Change to yes to see the radiosity problem
global_settings {
assumed_gamma 2.2
#if (UseRadiosity)
radiosity {
count 300
error_bound .02
pretrace_start .08
pretrace_end .02
recursion_limit 1
}
#end
}
camera {
location <0, 0, -4>
look_at <0, 0, 0>
}
#local White = color rgb <1, 1, 1>;
#local TestColor = color rgb <160, 255, 90> / 255;
#if (UseRadiosity)
difference {
sphere {0, 2001}
sphere {0, 2000}
pigment {color White * 16}
}
#else
light_source {<10, -15, -50> color White * 1.3 shadowless}
#end
#local TestObject = cylinder {<0, 0, -1> <0, 0, 0>, .3}
plane {z, 0 inverse texture {pigment {color White}}}
/* Left column is three cylinders placed slightly in front of a white plane, */
/* with transmit values (from top down) of 0, .8, and 1. */
object {
TestObject
texture {pigment {color TestColor transmit 0}}
translate <-1, 1, -.01>
}
object {
TestObject
texture {pigment {color TestColor transmit .8}}
translate <-1, 0, -.01>
}
object {
TestObject
texture {pigment {color TestColor transmit 1}}
translate <-1, -1, -.01>
}
/* Right column is three cylinders extending slightly into a white plane, */
/* with transmit values (from top down) of 0, .8, and 1. */
object {
TestObject
texture {pigment {color TestColor transmit 0}}
translate <1, 1, .01>
}
object {
TestObject
texture {pigment {color TestColor transmit .8}}
translate <1, 0, .01>
}
object {
TestObject
texture {pigment {color TestColor transmit 1}}
translate <1, -1, .01>
}
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
How come every time I post something, as soon as I post it I immediately realize
I forgot something important, no matter how carefully I proof-read? Darn it! I'm
using version 3.6.2. Sorry.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> Hi there, I noticed a couple problems with the transmit feature. I made a simple
> scene demonstrating the problems. I also made two demonstration images. One
> image is made with radiosity and one without. The scene consists of test objects
> (cylinders)with varying amounts of transmit. Three objects are placed slightly
> in front of a white plane and three cylinders are placed with the ends extending
> slightlly into the plane.
>
> The image without radiosity,
>
> http://mysite.verizon.net/moronicon/Transmit-NoRadiosity.jpg
>
> shows that the cylinder with transmit set to .8 and placed slightly in front of
> the plane (right middle object) doesn't appear tinted enough from the sides, but
> does appear tinted correctly from the top end.
>
> The image with radiosity,
>
> http://mysite.verizon.net/moronicon/Transmit-Radiosity.jpg
>
> shows that trasmit values don't appear to be taken into account properly when
> calculating radiosity.
>
> Here is the scene used to generate the images:
>
> #local UseRadiosity = no; //Change to yes to see the radiosity problem
>
> global_settings {
> assumed_gamma 2.2
> #if (UseRadiosity)
> radiosity {
> count 300
> error_bound .02
> pretrace_start .08
> pretrace_end .02
> recursion_limit 1
> }
> #end
> }
>
> camera {
> location <0, 0, -4>
> look_at <0, 0, 0>
> }
>
> #local White = color rgb <1, 1, 1>;
> #local TestColor = color rgb <160, 255, 90> / 255;
>
> #if (UseRadiosity)
> difference {
> sphere {0, 2001}
> sphere {0, 2000}
> pigment {color White * 16}
> }
> #else
> light_source {<10, -15, -50> color White * 1.3 shadowless}
> #end
>
> #local TestObject = cylinder {<0, 0, -1> <0, 0, 0>, .3}
>
> plane {z, 0 inverse texture {pigment {color White}}}
>
> /* Left column is three cylinders placed slightly in front of a white plane, */
> /* with transmit values (from top down) of 0, .8, and 1. */
>
> object {
> TestObject
> texture {pigment {color TestColor transmit 0}}
> translate <-1, 1, -.01>
> }
>
> object {
> TestObject
> texture {pigment {color TestColor transmit .8}}
> translate <-1, 0, -.01>
> }
>
> object {
> TestObject
> texture {pigment {color TestColor transmit 1}}
> translate <-1, -1, -.01>
> }
>
> /* Right column is three cylinders extending slightly into a white plane, */
> /* with transmit values (from top down) of 0, .8, and 1. */
>
> object {
> TestObject
> texture {pigment {color TestColor transmit 0}}
> translate <1, 1, .01>
> }
>
> object {
> TestObject
> texture {pigment {color TestColor transmit .8}}
> translate <1, 0, .01>
> }
>
> object {
> TestObject
> texture {pigment {color TestColor transmit 1}}
> translate <1, -1, .01>
> }
>
>
>
In both cases, when the end is inside the plane, you have one less
surface to traverse. It's normal that the tint is not the same.
In the non-radiosity case, the light is to one side, so that the
illumination is not the same. What you see is as expected.
In the radiosity case, you use "recursion_limit 1" whitch is not enough.
If you use recursion_limit 2, the black areas disapears. You realy need
recursion_limit 3 to have good result. You need to change your "sky
with: pigment {color White }finish{ambient 1 diffuse 0}
Otherwise, it gets REALY to bright as the plane illuminate the sky.
If you expect the colour to depend on the thickness of the object, you
need an interior where you define a fade_color, fade_distance and
fade_power. Without that, only the surfaces affect the resulting coloration.
Unrelated: assumed_gamma in the global_settings is been phased out and
will not be supported in future versions. It's beter to NOT use it and
only rely on the effective illumination level. Instead, you should rely
on the gamma settings from the main ini file, and NOT change it for a
given scene.
Alain
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Alain <aze### [at] qwerty org> wrote:
>
> In both cases, when the end is inside the plane, you have one less
> surface to traverse. It's normal that the tint is not the same.
>
> In the non-radiosity case, the light is to one side, so that the
> illumination is not the same. What you see is as expected.
>
> In the radiosity case, you use "recursion_limit 1" whitch is not enough.
> If you use recursion_limit 2, the black areas disapears. You realy need
> recursion_limit 3 to have good result. You need to change your "sky
> with: pigment {color White }finish{ambient 1 diffuse 0}
> Otherwise, it gets REALY to bright as the plane illuminate the sky.
>
> If you expect the colour to depend on the thickness of the object, you
> need an interior where you define a fade_color, fade_distance and
> fade_power. Without that, only the surfaces affect the resulting coloration.
>
> Unrelated: assumed_gamma in the global_settings is been phased out and
> will not be supported in future versions. It's beter to NOT use it and
> only rely on the effective illumination level. Instead, you should rely
> on the gamma settings from the main ini file, and NOT change it for a
> given scene.
>
>
>
> Alain
Thank you, I appreciate the helpful information. I re-rendered the test scene
using recursion_limit 2 and 3. The dark areas on the plane AROUND the cylinders
did indeed disappear, but the dark area causing the "bottoms" of the cylinders
to appear black remained. I still think there is a problem with the way
radiosity and transmit (also filter) interact. I'm reasonably sure that if one
had a real cylinder made of a clear material and placed it near a white object,
there would not be any dark ring like we see in these images.
I tried using interior. Without radiosity a mostly-transparent object looked
great, but adding the radiosity caused the dark ring to appear again. I'm not as
experienced at using interior so I'm probably doing something wrong.
Regards,
Dave Blandston
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Alain <azerty [at] qwertyorg> wrote:
>
> In both cases, when the end is inside the plane, you have one less
> surface to traverse. It's normal that the tint is not the same.
>
> In the non-radiosity case, the light is to one side, so that the
> illumination is not the same. What you see is as expected.
>
> In the radiosity case, you use "recursion_limit 1" whitch is not enough.
> If you use recursion_limit 2, the black areas disapears. You realy need
> recursion_limit 3 to have good result. You need to change your "sky
> with: pigment {color White }finish{ambient 1 diffuse 0}
> Otherwise, it gets REALY to bright as the plane illuminate the sky.
>
> If you expect the colour to depend on the thickness of the object, you
> need an interior where you define a fade_color, fade_distance and
> fade_power. Without that, only the surfaces affect the resulting coloration.
>
> Unrelated: assumed_gamma in the global_settings is been phased out and
> will not be supported in future versions. It's beter to NOT use it and
> only rely on the effective illumination level. Instead, you should rely
> on the gamma settings from the main ini file, and NOT change it for a
> given scene.
>
>
>
> Alain
/* This might be a duplicate posting - I don't see the first one even though the
message counter incremented. Sorry if this gets double-posted! /*
Thank you, I appreciate the helpful information. I re-rendered the test scene
using recursion_limit 2 and 3. The dark areas on the plane AROUND the cylinders
did indeed disappear, but the dark area causing the "bottoms" of the cylinders
to appear black remained. I still think there is a problem with the way
radiosity and transmit (also filter) interact. I'm reasonably sure that if one
had a real cylinder made of a clear material and placed it near a white object,
there would not be any dark ring like we see in these images.
I tried using interior. Without radiosity a mostly-transparent object looked
great, but adding the radiosity caused the dark ring to appear again. I'm not as
experienced at using interior so I'm probably doing something wrong.
Regards,
Dave Blandston
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> Alain <aze### [at] qwertyorg> wrote:
>> In both cases, when the end is inside the plane, you have one less
>> surface to traverse. It's normal that the tint is not the same.
>>
>> In the non-radiosity case, the light is to one side, so that the
>> illumination is not the same. What you see is as expected.
>>
>> In the radiosity case, you use "recursion_limit 1" whitch is not enough.
>> If you use recursion_limit 2, the black areas disapears. You realy need
>> recursion_limit 3 to have good result. You need to change your "sky
>> with: pigment {color White }finish{ambient 1 diffuse 0}
>> Otherwise, it gets REALY to bright as the plane illuminate the sky.
>>
>> If you expect the colour to depend on the thickness of the object, you
>> need an interior where you define a fade_color, fade_distance and
>> fade_power. Without that, only the surfaces affect the resulting coloration.
>>
>> Unrelated: assumed_gamma in the global_settings is been phased out and
>> will not be supported in future versions. It's beter to NOT use it and
>> only rely on the effective illumination level. Instead, you should rely
>> on the gamma settings from the main ini file, and NOT change it for a
>> given scene.
>>
>>
>>
>> Alain
>
> Thank you, I appreciate the helpful information. I re-rendered the test scene
> using recursion_limit 2 and 3. The dark areas on the plane AROUND the cylinders
> did indeed disappear, but the dark area causing the "bottoms" of the cylinders
> to appear black remained. I still think there is a problem with the way
> radiosity and transmit (also filter) interact. I'm reasonably sure that if one
> had a real cylinder made of a clear material and placed it near a white object,
> there would not be any dark ring like we see in these images.
>
> I tried using interior. Without radiosity a mostly-transparent object looked
> great, but adding the radiosity caused the dark ring to appear again. I'm not as
> experienced at using interior so I'm probably doing something wrong.
>
> Regards,
> Dave Blandston
>
>
I tested on my side and with recursion_limit 3, I got prety good result
and no black area. Only difference: I used version 3.7 beta 34.
Some times, increasing max_trace_level can help. With 3.6, it sometimes,
can help aleviate /some/ radiosity problems.
Alain
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Alain <aze### [at] qwertyorg> wrote:
> I tested on my side and with recursion_limit 3, I got prety good result
> and no black area. Only difference: I used version 3.7 beta 34.
>
> Some times, increasing max_trace_level can help. With 3.6, it sometimes,
> can help aleviate /some/ radiosity problems.
>
>
>
> Alain
That's very interesting. Yet one more reason to look forward to version 3.7! (I
haven't been this excited about a new version since support for True Type fonts
was added.)
I've spent a lot of time working on this over the last couple days since this
issue was causing problems with my IRTC "Space" entry. Your help is much
appreciated.
Regards,
Dave Blandston
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I notice that your objects have no finish{} block. That means they are using the
default finish, which has a non zero ambient, IIRC. This will make the objects
glow when using radiosity & that will have a noticeable effect on the appearance
of translucent objects.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"PM 2Ring" <nomail@nomail> wrote:
> I notice that your objects have no finish{} block. That means they are using the
> default finish, which has a non zero ambient, IIRC. This will make the objects
> glow when using radiosity & that will have a noticeable effect on the appearance
> of translucent objects.
Oh, yes - I hadn't thought of that. The object I'm working on for my scene
actually has ambient .6! I may adjust it, but that's probably close to where it
will end up. But that dark area between the object and the background plane is
driving me nuts...
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> Alain <aze### [at] qwertyorg> wrote:
>> I tested on my side and with recursion_limit 3, I got prety good result
>> and no black area. Only difference: I used version 3.7 beta 34.
>>
>> Some times, increasing max_trace_level can help. With 3.6, it sometimes,
>> can help aleviate /some/ radiosity problems.
>>
>>
>>
>> Alain
>
> That's very interesting. Yet one more reason to look forward to version 3.7! (I
> haven't been this excited about a new version since support for True Type fonts
> was added.)
>
> I've spent a lot of time working on this over the last couple days since this
> issue was causing problems with my IRTC "Space" entry. Your help is much
> appreciated.
>
> Regards,
> Dave Blandston
>
>
One thing that you can do:
Increase the gap as much as you can, or have no gap at all.
Alain
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
|
|
