Camera response functions as a tonemapping option

[Lord Crc]

Could this have something to do with the gamma compensation you were working on committing?

Most likely it's due to the parser changes I made at the same time. Could you please post the agfapan-apx-400.crf file (PM is ok)? Just zip it up and attach it.

[Lord Crc]

I believe I've identified the issue. I assumed the monochromatic CRF's would be encoded in the "red" channel and thus have "red" in the name of the function (in the file). I'm fixing it now

[ralvision]

I believe I've identified the issue. I assumed the monochromatic CRF's would be encoded in the "red" channel and thus have "red" in the name of the function (in the file). I'm fixing it now

Thank You! I'm glad to see the acceleration of various facets of Lux converging towards 0.8's final release. I can only imagine where we will be by the time a 1.0 benchmark is reached!

Kudos.

[niverik2k]

The CRF funcion is really great... Is there any documentation on making your own camera response functions. We need CRFs for modern cameras, such as the canon eos line, the red camera, modern Kodak cinema film. I own a Mac, so I can't use windows software. Is it possible to input information such as the sensitometric curves. Shouldn't the CRFs affect the white balance? A tungsten film in blue when used in daylight conditions. Shouldn't the CRF be related to the linear work flow, because the film types specify their ISO, such as 250T is designed to be recorded in those conditions.

Here is a link to Kodak 2647 film specs
http://motion.kodak.com/motion/uploadedFiles/TI2647.pdf

How does all of this, including the tone mapping features relate to a LUT(lookup table?)
http://en.m.wikipedia.org/wiki/3D_LUT

Thankyoufor all of tour hard work, for such an amazing program
Justin W. K.

[jeanphi]

Hi,

The CRF will convert the raw (linear) RGB values to adapted RGB values matching the film properties. So you need to correctly choose your colorspace so that the raw RGB values are in the CRF color space. That's because this kind of data is more universally available. The data provided in you Kodak film spec is much more detailed and could obviously be used to compute a CRF file.

Jeanphi

[niverik2k]

That's great to know. Do you think you could talk a little more about how one could go about creating CRF file. Rokszycka started a post asking that question, viewtopic.php?f=16&t=6005. The idea sounds fascinating, and I imagine having addittional CRFs would be useful for the community. Thank you for your help.

[dougal2]

Luminance HDR can calculate a CRF when merging bracketed exposures. I've attached an example of the data it outputs, however I have already spent several hours this evening trying to adjust it to work in LuxRender (by applying various normalisations) but I cannot get a decent exposure out of it.

[Rom1]

Hi,

I just own a Canon EOS 5D mkII.

I want to create my CRF too, let me explain what I understand :

I just need to take the same picture with 5 different exposure (+/- 2Ev) and let's luminance HDR compute "a curve"

The part that is unknown is : how to convert "luminance HDR curve" to "CRF curve understood by luxrender"

Am I true ?

[Rom1]

I found another website talking about CRF :

http://www.jaloxa.eu/webhdr/calibrate.shtml

And for my own purpose, there is a "CRF" file
http://www.jaloxa.eu/webhdr/cameras/Can ... k_II.shtml

Would it be possible to do some conversion ?

[dougal2]

I tried several different methods to compute the response from images, but I didn't have any success. the most advanced/complete/modern tools I could find are the ones included with hugin, but the response estimations it came up with were way off the mark.