Understanding Gamma, what is gamma?

sblantipodi

2[H]4U
2FA
Joined
Aug 29, 2010
Messages
3,805
Is there a simple explanations without many formula to summarize what is gamma and why use gamma 2.2 instead of sRGB gamma or viceversa?

What is the best gamma to use with a Eizo S2433W for general use and photo editing?
Is it good to use gamma 2.2 for sRGB images or for web?
 
Last edited:
Gamma can be needed to be adjusted based on your environments lighting, its a flexible value. However it really shouldn't differ more than 2.0-2.4, as for a more technical explanation on what gamma IS, you'll have to wait for another poster. Google is your friend.
 
Gamma can be needed to be adjusted based on your environments lighting, its a flexible value. However it really shouldn't differ more than 2.0-2.4, as for a more technical explanation on what gamma IS, you'll have to wait for another poster. Google is your friend.

What is the difference between Gamma 2.2 and Gamma sRGB?
 
Not sure what you mean with sRGB Gamma... sRGB is a colorspace standard thingy with a temp of 6500k.

You can run a monitor in sRGB mode (or custom settings corresponding to it) but still have gamma set to whatever level (it ought to be 2.2 though).
 
We should use the term "gradation". It defines the brightness distribution of for example a display or a working color space (to remain in the area of the subject). When talking about "gamma xy" we mean a gradation that can be described through a simple power function with accordant exponent (for the initial gamma correction the accordant inverse is used).

Let's look at a LCD: It would theoretically allow a linear behaviour (gamma = 1.0) of input and output intensity. That means 50% input brightness results in 50% output brightness. We assume a whitepoint with 140 cd/m² and 8bit RGB signal. RGB(127/127/127) input => 70 cd/m² output. Do we want this behaviour? => No. The human perception is not linear. In this case perceived half brightness lies around 26 cd/m² (!). So many tonal values would have been wasted in wrong areas. That's why the L* charateristic (corresponds with human perception) is the best way to transform tonal values from a linear space to a gamma corrected space with limited tonal value range. The ECI-RGB 2.0 color space was accordingly defined and is therefore well suitable for RAW developing. Other definitions are based on (often) historical contraints like the typical CRT brightness distribution or print charateristics (that's why the UGRA recommends a gamma of 1.8 for offset proofing).

What is the best gamma to use with a Eizo S2433W for general use and photo editing?
As long as your current display gradation is noted correctly in the ICC profile you will achieve a correct representation regarding the brightness distribution in color aware software. The CMM carries out the required transformations. When using a display with extensive electronic (> 8bit LUT) it makes sense to achieve the gradation of the desired working color space via OSD controls in the calibration workflow to decrease the loss of tonal values. But in principle you could calibrate your display for example with a gamma of 1.8 and get a correct representation of AdobeRGB content (gamma 2.2) in color aware software - as long as the gamma of 1.8 is noted in the ICC profile for each channel.

Not sure what you mean with sRGB Gamma... sRGB is a colorspace standard thingy with a temp of 6500k.
sRGB gradation is a gradation that associated with the sRGB color space (btw. the primaries are defined relative to D65 but you can of course use any other color temperature to work in sRGB). It can't be descibed through a simple power function although the corresponding gamma is over long passages near a gamma of 2.2. It has a linear fraction in the lower area. Quite difficult to see in the gradation curve itself but when comparing the correspondig gamma over the luminance axis it gets quite visible (below 10% the buckle is even sharper).

sRGB gradation - corresponding gamma:

3030323062653537.jpg


Best regards

Denis
 
Last edited:
As long as your current display gradation is noted correctly in the ICC profile you will achieve a correct representation regarding the brightness distribution in color aware software. The CMM carries out the required transformations. When using a display with extensive electronic (> 8bit LUT) it makes sense to achieve the gradation of the desired working color space via OSD controls in the calibration workflow to decrease the loss of tonal values. But in principle you could calibrate your display for example with a gamma of 1.8 and get a correct representation of AdobeRGB content (gamma 2.2) in color aware software - as long as the gamma of 1.8 is noted in the ICC profile for each channel.

wow Denis, thanks but what about software that isn't color aware?
 
wow Denis, quite a complicated answer. idk if im just stupid or something but i might have to read what you wrote a few times to understand it.
 
If your color primaries and 'gradation' curve don't match the image in question, everything will look different compared to how it would when color managed. Make an effort to always use color-managed software whenever possible (Image Viewer/Editor, Web Browser, Video Renderer). The default color space for web images is sRGB, and they will look grossly incorrect on a wide-gamut monitor (without gamut emulation) no matter what 'gradation' curve you use. That is why color-managed software is so important.

sRGB = sRGB non-linear curve
AdobeRGB = 2.2 gamma power-curve
ProPhoto RGB = 1.8 gamma power-curve

LAB = L* non-linear curve
XYZ = Linear (?)
LOG color space = non-linear logarithm curve

Picking the curve which matches your working color space will yield the best results on-screen because it means less adaption is needed when color-managed.
On the other side of the coin you have the option of using L*, which is theorized (controversial) to offer more accurate transformations, since ICC color corrections are done in L* D50.
Also worth keeping in mind, 'perfect' gamma power-curves are able to be linearized more accurately than specialized curves like sRGB. (not worth worrying about unless you are troubleshooting print-matching issues)
 
Last edited:
On the other side of the coin you have the option of using L*, which is theorized (controversial) to offer more accurate transformations, since ICC color corrections are done in L* D50.
I would recommend L* as display gradation via calibration when working in ECI-RGB 2.0 and if the display can achieve this gradation in hardware (therefore avoids a loss of tonal values at this stage - which are about 25% when comming from a gamma 2.2 => quite painful in a 8bit space). But his Eizo only has some fixed gamma settings.

wow Denis, thanks but what about software that isn't color aware?
As Cyberbeing said: It's of course not only a question of gradation when talking about correct representation in unmanaged applications. However: Your display has some fixed gamma settings and when thinking of "usual" content you will be just fine with sRGB or gamma 2.2 gradation as calibration aim. With "2.2" OSD option your display rather tends in the direction of sRGB gradation (when looking at your posted UDACT) so I would probably choose that. Whitepoint in consideration of colour matching and ambient conditions.

Best regards

Denis
 
Last edited:
I would recommend L* as display gradation via calibration when working in ECI-RGB 2.0 and if the display can achieve this gradation in hardware (therefore avoids a loss of tonal values at this stage - which are about 25% when comming from a gamma 2.2 => quite painful in a 8bit space). But his Eizo only has some fixed gamma settings.


As Cyberbeing said: It's of course not only a question of gradation when talking about correct representation in unmanaged applications. However: Your display has some fixed gamma settings and when thinking of "usual" content you will be just fine with sRGB or gamma 2.2 gradation as calibration aim. With "2.2" OSD option your display rather tends in the direction of sRGB gradation (when looking at your posted UDACT) so I would probably choose that. Whitepoint in consideration of colour matching and ambient conditions.

Best regards

Denis

wow. I haven't any other questions in this thread but I have a problem with iColor when trying to calibrate to sRGB gamma.
When calibrating the White Point I can achieve easily precise 6500K with both sRGB and 2.2 Gamma.
Unfortunantly the white point measured in UDACT is always quite far from the one I measured in the
white point menu before profiling when using sRGB gamma, can't understand why, this makes really
difficult to calibrate using sRGB gamma.
 
Last edited:
Back
Top