Processing RAW images for accurate color measurements in scientific applications

[Manfred M]

Just a query - why would the on-screen colour be sRGB? Surely it would be a conversion of the working colour space to the monitor colour space (as specified in the monitor profile)? Or are you saying that by way of simplification, as most monitors have a colour space of approximately sRGB?

First of all most standard monitors are 6-bit and emulate 8-bit per channel (how well they do the emulation is very much up for debate), so at best they can display sRGB, which is 8-bits per channel / 24-bit. This makes up most of the monitors out there that people are using.

If you happen to have an expensive high gamut monitor, these are true 8-bit per channel displays and emulate 10-bit (30-bit), so they can natively display sRGB and hit most or all of AdobeRGB only through emulation.

Both of these are moot points, to some extent, as our eyes can only see around 10 million different tones, so well below the 16 million that sRGB can output.

[Simon Garrett]

If you happen to have an expensive high gamut monitor, these are true 8-bit per channel displays and emulate 10-bit (30-bit), so they can natively display sRGB and hit most or all of AdobeRGB only through emulation.

Forgive me, but in terms of colour space I think it's the other way round. Wide gamut monitors can natively display a gamut of typically around Adobe RGB, and through emulation some can display sRGB. Emulation of sRGB by modern wide gamut monitors is quite good, and pretty awful in older ones like my HP LP2475w. My monitor offers only an 8-bit interface to an 8-bit panel, but its native colour space is slightly larger than Adobe RGB.

What you see from a colour-managed application to a profiled and calibrated monitor (of whatever colour space) is a conversion of the image from application working space to monitor colour space. That's what colour management to a monitor does.

However, as you say most monitors out there are 6 or 8 bit with a colour space that approximates to sRGB.

[Manfred M]

Forgive me, but in terms of colour space I think it's the other way round. Wide gamut monitors can natively display a gamut of typically around Adobe RGB, and through emulation some can display sRGB. Emulation of sRGB by modern wide gamut monitors is quite good, and pretty awful in older ones like my HP LP2475w. My monitor offers only an 8-bit interface to an 8-bit panel, but its native colour space is slightly larger than Adobe RGB.

What you see from a colour-managed application to a profiled and calibrated monitor (of whatever colour space) is a conversion of the image from application working space to monitor colour space. That's what colour management to a monitor does.

However, as you say most monitors out there are 6 or 8 bit with a colour space that approximates to sRGB.

Simon - so far as I understand they work, what I have written is correct. The wide gamut monitors are native 8-bit, so by definition they display 24-bit Truecolor, which is 16.2 million coloures = sRGB. I double checked some of the EISO monitor specs just to make sure that the really super high end ones are more or less the same as the merely expensive wide gamut ones.

The way that AdobeRGB is done is through a technique called Frame Rate Control (FRC). If your high gamut monitor is using AdobeRGB and you are running at 60Hz, you will get one frame showing one set of colours (i.e. 30 HZ) and the second frame showing another set of colours (also at 30 HZ). The two flickering colours set emulate the AdobeRGB gamut.

Some of the articles I've read suggest that to call this true AdobeRGB is wishful thinking, as not everyones eyes process this quite the same way. The nearest analogy I can think of is interlaced standard definition television where only half of the lines are displayed at any one time.

[Colin Southern]

Both of these are moot points, to some extent, as our eyes can only see around 10 million different tones, so well below the 16 million that sRGB can output.

Not really a moot point because the NUMBER of colours doesn't equate to the RANGE of colours. With monitors using additive colour -v- printers using subtractive colour, monitors with an sRGB gamut excel at displaying images with red and blue content, but struggle to reproduce strong cyans and magentas that printers excel at (and in contrast, printers can struggle to reproduce strong reds & blues).

In "ballpark" terms, sRGB is a common colour gamut between printers and monitors whereas Adobe RGB on a monitor is roughly equivalent to expanding the monitor to include typical printer gamuts.

[ajohnw]

I believe the general impression is that 10bit per colour channel is needed to match the eye's performance and some feel that it really needs to be more. It's also been possible to have a pc set up that will provide this for some time but it belongs in the high end graphics work station world. On the other hand Windows 7 has been designed to support 30 and 48bit colour already so who know what will happen in the next few years. Might be worth reading about deep colour here http://en.wikipedia.org/wiki/Color_depth There is nothing new about it at all other than it has been for a rather restricted market but note Adobe offered extensions to cope with a 30bit work flow. I saw a portrait photographer using a deep colour display plugged into a medium format digital camera some years ago working in a local shopping precinct. Enviable tethered shooting but the costs must be astronomic. Pointless unless the work is going to be printed.

As I understand it current 24bit colour is handled at 32bit per pixel and the spare 8 bits are used for transparency. I'd guess microsoft have decided to switch to 64bit as that allows a similar arrangement with either 30bit or 48 bit colour. This area is likely to be the next one to get us to change our gear even out TV's at some point.

I also believe many current monitors are truly 8bit per channel and have been so for some time. LCD displays have come a long long way since they started and continue to improve largely driven by the TV market.

There are numerous debayering algorithms about. You can see the artefacts that may occur on some here along with an explanation why this happens - interpolation and assumptions about colour - the image changes as the mouse is moved across each algorithm.

http://www.pages.drexel.edu/~par24/r...Artifacts.html

Oddly I use vng 4 colour maybe Nikon do too or the other 4 colour method. The AHD algorithm is more popular. The 4 colour ones are really intended for sensors that have a particular problem. There is another one ACC. There are some enlarged crops of results here with adobe acr included.

http://www.my-spot.com/rhc/RHC_Demosiac.htm

-

[Simon Garrett]

The wide gamut monitors are native 8-bit, so by definition they display 24-bit Truecolor, which is 16.2 million coloures = sRGB.

The number of bits is not the same thing a colour space, and does not define the colour space. An 8-bit per pixel monitor can be any colour space.

The native colour space of a monitor is not dependent on the number of bits, but on the choice of primary colours used to create Red, Green and Blue colours on the screen.

My HP monitor is 8 bits per pixel, but has a native colour space a bit wider than Adobe RGB.

I double checked some of the EISO monitor specs just to make sure that the really super high end ones are more or less the same as the merely expensive wide gamut ones.

The way that AdobeRGB is done is through a technique called Frame Rate Control (FRC). If your high gamut monitor is using AdobeRGB and you are running at 60Hz, you will get one frame showing one set of colours (i.e. 30 HZ) and the second frame showing another set of colours (also at 30 HZ). The two flickering colours set emulate the AdobeRGB gamut.

Some of the articles I've read suggest that to call this true AdobeRGB is wishful thinking, as not everyones eyes process this quite the same way. The nearest analogy I can think of is interlaced standard definition television where only half of the lines are displayed at any one time.

That may be how EISO monitors work, but it's not how monitors in general work. Mine certainly doesn't emulate Adobe RGB in the way you say.

Adobe RGB is a wider colour space than sRGB, and it contains colours that are outside the sRGB colour space. It is possible to emulate sRGB on an Adobe RGB monitor, as the sRGB colour space is a subset of Adobe RGB. All sRGB colours are also Adobe RGB colours. But the other way isn't true. Some Adobe RGB colours are not sRGB colours, and can't be emulated by a monitor whose native colour space is sRGB.

For give me if I'm incorrectly interpreting what you are saying.

[xpatUSA]

Just a query - why would the on-screen colour be sRGB? Surely it would be a conversion of the working colour space to the monitor colour space (as specified in the monitor profile)? Or are you saying that by way of simplification, as most monitors have a colour space of approximately sRGB?

Yes, it was by way of simplification, because Erich seemed more interested in accurate measurement of color than how it looks on-screen. Introducing the monitor color space into the discussion, as seen in most posts subsequent to the OP, has so clouded the issue that my guess is that Erich is now thoroughly confused . . . to the extent of not re-joining us after setting the Cat amongst the Pigeons ;-)