|
 |
> Hello
> I was considering trying to use POV ray to compute illuminance values for a
> solar energy system. Can the povray output be used to determine illuminance,
> e.g. lumens/sqm or the like? Has anyone developed such methods?
> Thanks
> Jeremy
>
>
In a way...
Using the photons feature, you get acurate reflection of incomming light.
If you use a standard image format, part of the illumination can easily
get clipped. You can circumvent that using some high dinamic range
format like HDR or EXR.
If you place a screen at the correct place, you can get an image where
the brightness is directly proportional to the illumination.
Using the animation feature, you can have that screen move, giving you a
serie of sampling planes.
While the results may not be as acurate as you may want/hope, they will
give you at least a reasonable estimate.
Alain
Post a reply to this message
|
|
|
|
Thanks
Not being too familiar with the software, perhaps there are settings that will
allow for higher accuracy - what would be the causes of inexactitude?
Alain <kua### [at] videotron ca> wrote:
> > Hello
> > I was considering trying to use POV ray to compute illuminance values for a
> > solar energy system. Can the povray output be used to determine illuminance,
> > e.g. lumens/sqm or the like? Has anyone developed such methods?
> > Thanks
> > Jeremy
> >
> >
>
> In a way...
> Using the photons feature, you get acurate reflection of incomming light.
> If you use a standard image format, part of the illumination can easily
> get clipped. You can circumvent that using some high dinamic range
> format like HDR or EXR.
>
> If you place a screen at the correct place, you can get an image where
> the brightness is directly proportional to the illumination.
> Using the animation feature, you can have that screen move, giving you a
> serie of sampling planes.
>
> While the results may not be as acurate as you may want/hope, they will
> give you at least a reasonable estimate.
>
>
>
> Alain
Post a reply to this message
|
|
|
|
> Thanks
> Not being too familiar with the software, perhaps there are settings that will
> allow for higher accuracy - what would be the causes of inexactitude?
>
> Alain <kua### [at] videotronca> wrote:
>>> Hello
>>> I was considering trying to use POV ray to compute illuminance values for a
>>> solar energy system. Can the povray output be used to determine illuminance,
>>> e.g. lumens/sqm or the like? Has anyone developed such methods?
>>> Thanks
>>> Jeremy
>>>
>>>
>>
>> In a way...
>> Using the photons feature, you get acurate reflection of incomming light.
>> If you use a standard image format, part of the illumination can easily
>> get clipped. You can circumvent that using some high dinamic range
>> format like HDR or EXR.
>>
>> If you place a screen at the correct place, you can get an image where
>> the brightness is directly proportional to the illumination.
>> Using the animation feature, you can have that screen move, giving you a
>> serie of sampling planes.
>>
>> While the results may not be as acurate as you may want/hope, they will
>> give you at least a reasonable estimate.
>>
>>
>>
>> Alain
>
>
>
>
POV-Ray is not meant to be used for your purpose.
To get more acurate results, you MUST set assumed_gamma 1. This way,
you'll work in a linear colour space.
When unsing a regular file format, like PNG ot BMP, your dynamic range
is small, resulting in clipping in areas of highest illumination. The
version 3.6.x don't support outputing in a high dynamic range format,
so, you need version 3.7.
Note: In the current RC5, photons reflection is broken. This should be
corrected in RC6 that should get out shortly.
Using version 3.6.x, you can get an extended range using hf_gray_16 in
the global settings and reducing your light's intensity by a factor of
256. This result in a special PNG gray level image using the red and
green channels to store low and high bytes.
It may be possible to fake a higher dynamic rangs. This involve
outputing in PNG or TGA set to use 16 bits per channel.
You then use weak a light like: rgb<1, 1/pow(256,2), 1/pow(256,4)>/256,
the "normal" range been 0..1. This will fake about 40 bits integer
output. You'll get some granularity as the red then green channels get
clipped as the light's intensity increases.
Using high dynamic range image format, you prevent clipping, but you can
get a rather croase granularity in the brightest areas. To get realy
acurqte results, you need double precision floats (64 bits per channel),
and the current formats only support half precision (16 bits per
channel), maybe single precision (32 bits per channel).
Alain
Post a reply to this message
|
|
