Gamma transfer function – my second try

R
Posted By
ronviers
Aug 5, 2006
Views
302
Replies
4
Status
Closed
I want to try once again to see if my understanding of the gamma transfer function is correct. Thanks for your patience.

Light intensity vs. signal input/output results in a curve. The ‘gamma transfer function’ is used to flatten this curve so the relationship of encoded image data to intensity will be a simple ratio. In other words, after the gamma transfer function has been applied to the image data, then 0=black, 255=white and, and this is the significant part, 128=middle gray. The very same gray as the 18% gray card everyone keeps talking about.

The ‘Gamma’ is an environment (device/color space) specific value used to generate the gamma transfer function. Gamma has three components – a gamma value (explicitly assigned) and black level and noise level each implicitly derived from the environment.

In a well corrected system, humans will see luminance as lightness. Edit and save in a gamma correct environment but do not tag your image.

Thanks,
Ron
Related Tags
function gamma transfer try

MacBook Pro 16” Mockups 🔥

– in 4 materials (clay versions included)

– 12 scenes

– 48 MacBook Pro 16″ mockups

– 6000 x 4500 px

K
KatWoman
Aug 5, 2006
wrote in message
I want to try once again to see if my understanding of the gamma transfer function is correct. Thanks for your patience.
Light intensity vs. signal input/output results in a curve. The ‘gamma transfer function’ is used to flatten this curve so the relationship of encoded image data to intensity will be a simple ratio. In other words, after the gamma transfer function has been applied to the image data, then 0=black, 255=white and, and this is the significant part, 128=middle gray. The very same gray as the 18% gray card everyone keeps talking about.

The ‘Gamma’ is an environment (device/color space) specific value used to generate the gamma transfer function. Gamma has three components – a gamma value (explicitly assigned) and black level and noise level each implicitly derived from the environment.
In a well corrected system, humans will see luminance as lightness. Edit and save in a gamma correct environment but do not tag your image.
Thanks,
Ron

it takes the black point in and out of the image
what else you need to know
it’s visual
LOOK AT IT
R
ronviers
Aug 5, 2006
LOOK AT IT

By "look" at it are you suggesting that I should watch what it does, then pay attention to the effects? I will consider that. If in a few days I cannot think of a reason not to do it will run it by the other members in this group to see what they say. If they agree that I should look at it I will run it by the members of my photography group to see…;)

Ron
TA
Timo Autiokari
Aug 6, 2006
wrote:

Light intensity vs. signal input/output results in a curve.

No. Light behaves absolutely linearly. Also CCD and CMOS raw data are detecting light linearly.

The ‘gamma transfer function’ is used to flatten this curve so the relationship of encoded image data to intensity will be a simple ratio.

No. The gamma function is the property of cathode ray tubes. Their native tonal reproduction curve is about gamma 2.5. So input voltages and the respective screen luminances follow the gamma function at gamma = 2.5 (a power law function with exponent of 2.5).

In a well corrected system, humans will see luminance as lightness.

No! The lightness is a totally different thing than the various surface luminances that we see in the scene. The lightness is what illuminates the scene surfaces.

The vision behaves about logarithmicly when it is assessing changes in the lightness. So e.g. every time you double the lightness of a room (by doubling the number of lamps that are illuminating the room) you will experience an increase in the room lightness that appear to be "similar" in "strength" compared to all the previous changes. This is how the vision adapts to the changes in lightness, this property of the vision is called the light adaptation.

When the lightness of the scene does not change (rapidly) and the vision is assessing the luminances of the various surfaces in the scene, the perception is baasically linear, however some local adaptation (in case of very large dark surfaces) will sometimes happen so the average perception is slightly non-linear but it does not follow the gamma function. Linear transfer function is however a very good approximation for luminance sensitivity of the vision.

Timo Autiokari
R
ronviers
Aug 6, 2006
Timo Autiokari wrote:
wrote:

Light intensity vs. signal input/output results in a curve.

No. Light behaves absolutely linearly. Also CCD and CMOS raw data are detecting light linearly.

The ‘gamma transfer function’ is used to flatten this curve so the relationship of encoded image data to intensity will be a simple ratio.

No. The gamma function is the property of cathode ray tubes. Their native tonal reproduction curve is about gamma 2.5. So input voltages and the respective screen luminances follow the gamma function at gamma = 2.5 (a power law function with exponent of 2.5).

In a well corrected system, humans will see luminance as lightness.

No! The lightness is a totally different thing than the various surface luminances that we see in the scene. The lightness is what illuminates the scene surfaces.

I will not respond point by point because I am going to start over, this time with different sources. I used reference material from TV electronics not color theory to develop my definitions. Luckily there is tons of material online.
My thinking on the luminance/lightness statement was an attempt to make the luminance definition derived from RGB color space and the L* definition derived from the CIELAB color model meet in reality. I thought it was a stretch when I wrote it.

Thank you for your time,
Ron

Must-have mockup pack for every graphic designer 🔥🔥🔥

Easy-to-use drag-n-drop Photoshop scene creator with more than 2800 items.

Related Discussion Topics

Nice and short text about related topics in discussion sections