Preface:
I'm sure you're all getting sick of me asking things that are probably common knowledge to you all, but I have another question. I searched the web, and went through Rendering with Radiance, but still have not been able to find the answer to my problem. I have a presentation tomorrow morning, and it seems like while I will have nice looking images, the quantitative evaluation may be lacking.
Problem:
When I generate a falsecolor image of an office space I rendered, I get a good image with a dynamic range. The command I use is:
falsecolor -s 400 -i image.pic > image2.pic
The image I rendered was a standard Radiance image that calculates the radiance/luminance.
The when I try to get luminance reading from the falsecolor image, the values I get are confounding. The brightest areas (the bright red regions) always get at reading of 47.8L. Surrounding areas typically have higher values.
To try to test what was going on, I clicked on the color scale on the falsecolor image and got readings which were not even close to what the scale indicates. The L values are maximum around the yellow/green section (reading at about 120L), and decrease going toward both the blue and section of the scale.
If anyone is reading this message on a Sunday evening, I would REALLY appreciate it if you could fill me in on how to correct this problem.
Again, apologies if I should have RTFM, but I tried, and have had no success thus far.
John
Hi!
Problem:
When I generate a falsecolor image of an office space I rendered, I get
a good image with a dynamic range. The command I use is:
falsecolor -s 400 -i image.pic > image2.pic
The image I rendered was a standard Radiance image that calculates the
radiance/luminance.
Do you want luminance or illuminance? Usually you need illuminance-pictures
rendered with rpict's -i parameter (if you want to know the amount of light
that hits the surface).
Good luck, CU, Lars.
Problem:
When I generate a falsecolor image of an office space I rendered, I get a good image with a dynamic range. The command I use is:
falsecolor -s 400 -i image.pic > image2.pic
The image I rendered was a standard Radiance image that calculates the radiance/luminance.
The when I try to get luminance reading from the falsecolor image, the values I get are confounding. The brightest areas (the bright red regions) always get at reading of 47.8L. Surrounding areas typically have higher values.
Odd that areas in the image that are not red would give higher values. Here's what Greg told me to do though:
Run "pextrem -o" on the image. (Read the manpage for details) The second line of the output is the highest value pixel in the image. Take the green value from that second line, multiply it by 179 and use that number for the scaling (-s) factor in your falsecolor renderings. That should make the red areas in your falsecolor rendering be the brightest, and allow the color range to depict the rest of the luminance best. I usually use a logarithmic scale to boot, usually 2 decades or three. Hope this helps.
To try to test what was going on, I clicked on the color scale on the falsecolor image and got readings which were not even close to what the scale indicates. The L values are maximum around the yellow/green section (reading at about 120L), and decrease going toward both the blue and section of the scale.
I would seriously doubt the pixels in the scale have any basis in reality.
If anyone is reading this message on a Sunday evening, I would REALLY appreciate it if you could fill me in on how to correct this problem.
Pathetic, but true, I am reading this. =8-) Happy Sunday everyone.
Rob Guglielmetti
[email protected]
www.rumblestrip.org
···
On Sunday, May 18, 2003, at 05:30 PM, John An wrote:
Hi John,
I hope you aren't doing what I think you might be doing -- trying to pick luminance values off the falsecolor output using ximage or suchlike. The falsecolor image has already converted the luminances to a "false color" display, and the original luminances are lost in the process -- converted to this other (visual) representation. If you want to get point luminance values, you'll need to run ximage on the original image, not on the falsecolor output. (You can have falsecolor add point values for the extrema if you like using the -e option.)
-Greg
···
From: John An <[email protected]>
Date: Sun May 18, 2003 2:30:43 PM US/Pacific
To: [email protected]
Subject: [Radiance-general] falsecolor weirdness
Reply-To: [email protected]
Preface:
I'm sure you're all getting sick of me asking things that are probably common knowledge to you all, but I have another question. I searched the web, and went through Rendering with Radiance, but still have not been able to find the answer to my problem. I have a presentation tomorrow morning, and it seems like while I will have nice looking images, the quantitative evaluation may be lacking.
Problem:
When I generate a falsecolor image of an office space I rendered, I get a good image with a dynamic range. The command I use is:
falsecolor -s 400 -i image.pic > image2.pic
The image I rendered was a standard Radiance image that calculates the radiance/luminance.
The when I try to get luminance reading from the falsecolor image, the values I get are confounding. The brightest areas (the bright red regions) always get at reading of 47.8L. Surrounding areas typically have higher values.
To try to test what was going on, I clicked on the color scale on the falsecolor image and got readings which were not even close to what the scale indicates. The L values are maximum around the yellow/green section (reading at about 120L), and decrease going toward both the blue and section of the scale.
If anyone is reading this message on a Sunday evening, I would REALLY appreciate it if you could fill me in on how to correct this problem.
Again, apologies if I should have RTFM, but I tried, and have had no success thus far.
John
_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general
John, my previous email was probably misleading. Listen To Greg, which is always good advice. Another way to look at things is this: think of a raw, unfiltered image fresh from rpict as a collection of "intelligent pixels". That is, the value of each pixel is the radiance of that point, as you know. But any *filtered* version of that image, while still containing that depth of information about each pixel, has been filtered, so the info no longer has any basis in reality. I learned this rather quickly when I tried doing a falsecolor image of an image that had already been filtered with pcond -h. Oops. It's a good argument for always saving the original image, as you never know when you might want to extract more info from the image in the form of a falsecolor, or whatever.
Rob Guglielmetti
[email protected]
www.rumblestrip.org
···
On Monday, May 19, 2003, at 01:50 AM, Greg Ward wrote:
Hi John,
I hope you aren't doing what I think you might be doing -- trying to pick luminance values off the falsecolor output using ximage or suchlike. The falsecolor image has already converted the luminances to a "false color" display, and the original luminances are lost in the process -- converted to this other (visual) representation. If you want to get point luminance values, you'll need to run ximage on the original image, not on the falsecolor output. (You can have falsecolor add point values for the extrema if you like using the -e option.)
Thanks all for your quick response.
Cc: [email protected]
To: [email protected]
From: Greg Ward <[email protected]>
Subject: [Radiance-general] Re: falsecolor weirdness
Reply-To: [email protected]
Hi John,
I hope you aren't doing what I think you might be doing -- trying to
pick luminance values off the falsecolor output using ximage or
suchlike. The falsecolor image has already converted the luminances to
a "false color" display, and the original luminances are lost in the
process -- converted to this other (visual) representation. If you
want to get point luminance values, you'll need to run ximage on the
original image, not on the falsecolor output. (You can have falsecolor
add point values for the extrema if you like using the -e option.)
-Greg
Well, being the idiot that I am, I was trying to get luminance values from the falsecolor output. I guess that explains why the red always read 47.8L. Evaluating the original image gave me the values I needed. Thanks.
From: "Lars O. Grobe" <[email protected]>
To: [email protected]
Subject: Re: [Radiance-general] falsecolor weirdness
Do you want luminance or illuminance? Usually you need illuminance-pictures
rendered with rpict's -i parameter (if you want to know the amount of light
that hits the surface).
Good luck, CU, Lars.
Though I had purposely done a luminance rendering for today's presentation, I will have to do illuminance analysis, so the -i parameter is next.
Subject: Re: [Radiance-general] falsecolor weirdness
From: Rob Guglielmetti <[email protected]>
To: [email protected]
Reply-To: [email protected]
Odd that areas in the image that are not red would give higher values.
Here's what Greg told me to do though:
Run "pextrem -o" on the image. (Read the manpage for details) The
second line of the output is the highest value pixel in the image.
Take the green value from that second line, multiply it by 179 and use
that number for the scaling (-s) factor in your falsecolor renderings.
That should make the red areas in your falsecolor rendering be the
brightest, and allow the color range to depict the rest of the
luminance best. I usually use a logarithmic scale to boot, usually 2
decades or three. Hope this helps.
I will try that next time around. Seems like this command will give me better control over the dynamic range in the falsecolor image.
John