Why 'more or less stopped using ISO'?

Started Feb 19, 2014 | Questions thread
Iliah Borg
Forum ProPosts: 16,690
Re: Maybe this will help.
In reply to bobn2, Feb 21, 2014

bobn2 wrote:

crames wrote:

jackdan wrote:

Iliah Borg wrote:

Another thing I need help with is that in RGB color space I understand brightness can be quantited.

Only if you know how to convert to XYZ, where Y is luminance. CIE states that brightness is relative luminance. Relative luminance is defined as photometric luminance normalized to 0..100 in such a way that reference white is 100.

So, even though I don't know how to do it, brightness can be quantitated as luminance using RGB values.

If I understand correctly, and I wouldn't be surprised if I don't, that in RGB color space, whatever that is, brightness is a quantity that is proportional to the magnitude of the pixel value from 0 to 255 is it, for each of the three values for R,B, andd G.

There is no need to use 255, you can set the upper limit as you wish, it is a question of normalizing. 0..1 is often used. Conversion of RGB to XYZ is often through a matrix, see http://brucelindbloom.com/Eqn_RGB_to_XYZ.html and http://brucelindbloom.com/WorkingSpaceInfo.html#Specifications The Y values for primaries R, G, and B allow to calculate the overall Y of a given RGB triplet by a simple weighted sum

Y = Yr*R+Yg*G+Yb*B

However RGB must be in a gamma=1 space. If it is some other gamma it needs to be linearized before anything.

So, brightness as luminance is proportional to RGB values in linear light space or as luma in gamma corrected RGB values in nonlinear light space.

The above however is not very relevant to RG1BG2 sensors, and especially if the transform from the sensor RGB to working RGB is given by a more complex transform rather than by a simple matrix

Thank you. I don't want to be any more of a distraction than I already have. I feel I now have at least a general idea about the relationship between image data and brightness Thanks to your generosity and patience.

There is one more thing, though. You did not comment on the 1st part of my post. Are the photometric units of measurement applicable to brightness, since they are for measurements of light in terms of perceived brightness to the human eye?

I copied that part of my post below for reference:

But I already asked to provide brightness unit. It did not help

-- hide signature --

show signature --

Please help me to sort all this out. I understand brightness is only a perception of the human eye, but isn't photometry the measurement of light in terms of perceived brightness to the human eye? And doesn't photometry weight actual radiant energy by a lumenosity function that models human brightness sensitivity? And doesn't photometric measurements have units for quantities such as luminance, luminous flux, luminous intensity, and luminous energy?

So although there are no units for brightness there are units for the measurment of the perceived brightness to the human eye.

And isn't that what matters when you are talking about perceived brightness as you and others are here.

The CIE has published the CIECAM02 Color Appearance Model, which includes a calculations to estimate brightness, "Q". You need not only the absolute luminance of the stimulus

If 'brightness' is dependent on 'absolute luminance of the stimulus' then there is no such thing as 'ISO'.

The disconnect between colour science (and Photoshop) and photography in terms of terms and definitions was always huge, and will remain that way (if not grow) because it is necessary that people are confused as much as possible We recently published a DNG simulation of Q13http://www.rawdigger.com/data/img/Q13_RGBG16.dng with embedded reference data (extract with exiftool -usercomment -b Q13_RGBG16.dng >Q13.csv) and I already received a couple of angry letters telling that "the vast majority of photographers will not be able to understand and use it properly".

Sadly, the disconnect in terms leads to different ways of calculations for the values that are supposed to be characterizing the same phenomenon - but they are not.

-- hide signature --
Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow