[Radiance-general] IES files and MGF geometry

[Axel Jacobs]

Hello again,

As promised, here are some more images of my box bollard.  Definitions 
first:
'closed' ... actual 6-sided box, as generated by ie2rad
'open'  ...  the above, with the top and bottom removed
'open_fixed' ... the above, with a modified source.cal

The light sources in a test room:
http://www.jaloxa.eu/pickup/closed_box_f.jpg
http://www.jaloxa.eu/pickup/open_box_f.jpg
http://www.jaloxa.eu/pickup/open_box_fixed_f.jpg
There is no noticeable difference between 'open' and 'open_fixed'.
Both are slightly different to 'closed'.

Diff images:
$ pcomb -f picdiff.cal closed_box.hdr open_box_fixed.hdr \
 > closed_v_fixed.hdr
etc

http://www.jaloxa.eu/pickup/closed_v_fixed_f.jpg
http://www.jaloxa.eu/pickup/closed_v_open_f.jpg
http://www.jaloxa.eu/pickup/open_v_fixed_f.jpg

'closed_v_open' and 'closed_v_fixed' are identical.  The reason for
this might be that practically all of the light is emitted side-ways,
so that removing the top and bottom of the box makes no difference at
all. Equally, fiddling with source.cal as per Greg's advice

"Regarding the local box correction, you could take out the third term
"noneg(abs(Pz-Dz*Ts)-A4/2)*A2*A3" from the formula:
lboxprojection = (      noneg(abs(Px-Dx*Ts)-A2/2)*A3*A4 +
                        noneg(abs(Py-Dy*Ts)-A3/2)*A2*A4 +
                        noneg(abs(Pz-Dz*Ts)-A4/2)*A2*A3 ) / Ts;"

seemed to give identical results.

Not sure if this provides a general answer to my question, but in my
particular case, it appears to be safe to reduce the box to its four
sides, which probably only works because of the very special
photometry.

Best
Axel

On 06/11/14 16:42, Axel Jacobs wrote:
> Hi Greg,
>
> the -dj is set by the rad command, which is behind ltview.
>
> render= -av 0 0 0
> INDIRECT= 0
> QUALITY= Med
> DETAIL= Low
> VARIABILITY= Med
>
> http://radiance-online.org/cgi-bin/viewcvs.cgi/ray/src/util/ltview.pl?view=markup
>
> I was also thinking whether or not the intensity of the actual 'light'
> material, as created by ies2rad, should be modified since the total
> surface area is smaller with the top and bottom of the box removed.
> But then, I guess you would have mentioned this.
>
> I will re-run the images with rpict proper and do a diff, but won't be
> able to do this until early next week.
>
> Thank you for your help
>
> Best
> Axel
>
>
>
> On 6 November 2014 15:52, Greg Ward <gregoryjward at gmail.com> wrote:
>> Hi Axel,
>>
>> Interesting results.  The correction of removing the third term for the local box correction would make the most difference in the upwards and downwards direction, so the sparse output there is why you don't see anything in your falsecolor comparison.  You could try computing a difference image using src/cal/cal/picdiff.cal.
>>
>> Also, the differences between using a large sphere and the lboxcorr won't be noticeable until your surfaces are up close.  What did you set -dj to in your runs?  This will also make a difference, and could show up artifacts in some of the simpler methods.
>>
>> Cheers,
>> -Greg
>>
>>> From: Axel Jacobs <jacobs.axel at gmail.com>
>>> Subject: Re: [Radiance-general] IES files and MGF geometry
>>> Date: November 6, 2014 3:02:26 AM PST
>>>
>>> Hello again,
>>>
>>> I've run some quick tests on with my bollard photometry. falsecolor
>>> images below:
>>>
>>> Unaltered IES file as it came from the manufacturer:
>>> http://www.jaloxa.eu/pickup/mr84_li34_fc.jpg
>>> This uses a short, capped cylinder.  rvu complains about
>>> "source aspect too small for cylinder xxx
>>> aiming failure for light source xxx"
>>> Looks weird.
>>>
>>> sphere, generated with ies2rad's -i option:
>>> http://www.jaloxa.eu/pickup/mr84_li34_sphere_fc.jpg
>>> This is a little darker than the previous image.
>>>
>>> box geometry with all 6 sides (edited IES file):
>>> http://www.jaloxa.eu/pickup/mr84_li34_closed_box_fc.jpg
>>> some 'streaky' artifacts that didn't show up in the sphere image.
>>>
>>> box with top and bottom removed. Unaltered source.cal
>>> http://www.jaloxa.eu/pickup/mr84_li34_open_box_fc.jpg
>>> a little darker than the full box
>>>
>>> box with top and bottom removed.  source.cal edited as per Greg's instructions
>>> http://www.jaloxa.eu/pickup/mr84_li34_open_box_fixed_fc.jpg
>>> Looks identical to the previous picture.
>>>
>>> I used the new ltview command for the previews, so the lower side of
>>> the enclosing plane is much further away than the actual ground would
>>> be.  This might be the reason why the 'corrected' and 'uncorrected'
>>> open boxes produce the same results.  Equally, there is hardly any
>>> downward or upward component.  Virtually all of the emitted light is
>>> projected into a narrow strip.  It still seems as if the open box does
>>> emit a little less light than its closed brother, even with the
>>> correction in source.cal
>>>
>>> Best
>>> Axel
>>
>> _______________________________________________
>> Radiance-general mailing list
>> Radiance-general at radiance-online.org
>> http://www.radiance-online.org/mailman/listinfo/radiance-general