Understand, we are on the same page. Thank you!
···
From: Greg Ward [mailto:[email protected]]
Sent: Monday, January 05, 2015 12:30 PM
To: Radiance general discussion
Subject: Re: [Radiance-general] False color and luminance values
Hi Brian,
I don't think you can read the values in that corner accurately using "-s
auto", which gets the wrong maximum due to the issue with pextrem and sets
the scale too low. If you use "-s 400" instead, you will see what I mean.
Cheers,
-Greg
From: "Brian Karr" <[email protected]>
Subject: Re: [Radiance-general] False color and luminance values
Date: January 5, 2015 4:25:52 AM PST
Gotcha. I selected a similar small area in a linear version of the false
color output (131 nits) as was selected in the rpict version (302 nits).
Here is the lin ver.
https://dl.dropboxusercontent.com/u/91366188/shelf_false_lin_norm.hdr
I hope Im understanding your question correctly.
Thank you,
Brian
From: Greg Ward [mailto:[email protected]]
Sent: Sunday, January 04, 2015 8:38 PM
To: Radiance general discussion
Subject: Re: [Radiance-general] False color and luminance values
Hi Brian,
I guess I wasn't being clear. I assumed you were selecting a box to average
in Photosphere, which does indeed work as expected. I was asking how you
read the comparison value (131 vs. 302 nits) from your falsecolor output.
Cheers,
-Greg
From: "Brian Karr" <[email protected]>
Subject: Re: [Radiance-general] False color and luminance values
Date: January 4, 2015 3:36:19 PM PST
For the small area, in Photosphere I dragged and dropped a selection box,
which updated the histogram. I wanted to perform a sanity check to see if
values in my Radiance model where matching the real world model. Guess I'm
used to that type of selection in other programs and it appeared to work.
maybe need to check further.
Point well taken on the strong primary color, its federal paint standard
Insignia Red, aka Red Safety No 11105, aka Fire Suppression System red
paint. Its used on all the launch pad sprinkler system pipes. The grey is
Grey Standard No 16473, which is launch pad grey.
Thank you,
Brian
From: Greg Ward [mailto:[email protected]]
Sent: Sunday, January 04, 2015 3:45 PM
To: Radiance general discussion
Subject: Re: [Radiance-general] False color and luminance values
Hi Brian,
How did you determine the "small area mean" of 131 nits using falsecolor?
The discrepancy between the two programs has to do with a bug in the
"pextrem" program, which uses the intensity rather than the luminance to
determine the brightest pixel. In your case, the red shelves have higher
intensity than the top shelf, despite the fact that the top of the bookcase
has a higher luminance value. Therefore, pextrem (used by falsecolor)
thinks the brightest point is on a red shelf rather than on top, and both
the luminance scale (using "auto") and the max value are incorrectly
determined.
It is possible to correct this flaw in falsecolor without too much
difficulty by replacing some lines of code, at a small expense in
computation time. This hasn't come up before, to my knowledge, as people
tend not to use such strong primary colors.
Best,
-Greg
From: "Brian Karr" <[email protected]>
Subject: [Radiance-general] False color and luminance values
Date: January 4, 2015 9:00:31 AM PST
Hello all,
A question regarding false color and luminance values. Starting with an
image created via rpict, and using the photosphere histogram, the max
luminance value is 351 nits. Selecting a small area around the top outer
corner of the shelf results in a mean around 302 nits. This agrees fairly
close to a measurement with a spot meter on the outer corner of a physical
version of the scene, so far so good.
Command: rpict -vtv -vp 0.75 .75 5.5 -vd 0.3 1.0 -0.6 -vu 0 0 1 -vh 45 -vv
45 -vo 3 -va 0 -vs 0 -vl 0 -av .1 .1 .1 -x 2000 -y 2000 -ab 2 -aa .1
shelf.oct > shelf_rpict.hdr
File: https://dl.dropboxusercontent.com/u/91366188/shelf_rpict.hdr
With a false color version, the histogram max luminance however is 177 nits,
the small area mean is 131 nits. Im not sure why the luminance values
between the two versions are different?
Command: falsecolor -ip shelf_rpict.hdr -e -s auto -n 16 -lh 1000 -lw 225
-log 4 -l > shelf_false_log_norm.hdr
File: https://dl.dropboxusercontent.com/u/91366188/shelf_false_log_norm.hdr
Thoughts?
Thank you!
Brian