«

when brightening my photos which were not taken with flash, when should
i modify (i use photoeditor) brightness and when do i use gamma? they
seem similar but there must be a good reason for two seperate controls.

thanks

beerismygas <> wrote:
: when brightening my photos which were not taken with flash, when should
: i modify (i use photoeditor) brightness and when do i use gamma? they
: seem similar but there must be a good reason for two seperate controls.

: thanks

As I understand the terms Brightness is the adjustment of how intense
White is. This would be like dimming a light bulb. White would still be
white but the intensity of that white is higher. Gamma is more a measure
of how intense the colors are in relation to eachother. So if you want
reds brighter than greens you would adjust more to the red end of the
adjustment (tho the scale may not be labeled as red/green but as a +/-).
So the way I do such adjustments is to first adjust the brightness and
contrast (the range between the brightest and darkest parts of the image),
and then adjust the gamma to tweek for the most asthetic balance of colors
for the particular image. The gamma is also where you will adjust skin
tones so they don't look too red or green (which can occur from varying
lighting or other such conditions at the time of image capture). One other
use for the gamma correction is to make up for the fact that some sensors
are more sensitive to certain colors of light than others. So if you find
that almost all your photos have a reddish cast (even after properly
setting the white balance) you would compensate with the gamma.

If you have experience with older TVs you may have had to play with the
controls and had to adjust the "color" knob. That is Gamma.

Randy

==========
Randy Berbaum
Champaign, IL

On Sat, 29 Jul 2006 06:09:45 +0000 (UTC), Randy Berbaum wrote:

> As I understand the terms Brightness is the adjustment of how intense
> White is. This would be like dimming a light bulb. White would still be
> white but the intensity of that white is higher. Gamma is more a measure
> of how intense the colors are in relation to eachother.
> . . .
> If you have experience with older TVs you may have had to play with the
> controls and had to adjust the "color" knob. That is Gamma.

I don't know about that. Based on very limited experience with
only one program, Gamma seems more of a complex brightness control
that doesn't affect the color relationship. At least with IrfanView
the Gamma slider doesn't seem to be a color control. It operates
more like a cross between the Brightness and Contrast sliders. As
if when the Brightness is increased, each intensity level in the
image is increased by the same absolute amount, resulting in both
light and dark areas becoming overly bright with small increases of
the slider's position. When the Gamma value is increased, it's as
if the intensity level of each pixel is increased by an amount
that's a function of the pixel's intensity level, irrespective of
its color. So by increasing the Gamma, overall brightness can be
increased, but the darker shadow areas, while brightening somewhat,
don't brighten nearly as rapidly.

The effect of this is that by pushing the Brightness slider to the
extreme left or right positions, the image will appear as either
pure black or pure white, and no visible detail is left in either
case. Do the same with the Gamma correction slider and you get a
different result. At one end the image gets extremely dark, but
some very small highlights can usually still be seen. Push the
slider to the other extreme and the image becomes not pure white,
but *very* light, with a faded, washed out appearance, and the image
and colors don't vanish, but can still be seen. Reds don't become
more relatively more intense than greens. To do that, there are
three (R, G and B) color balance sliders. In whatever program you
use (PS?) I assume that these two controls (Brightness and Gamma
Correction) are available and have been implemented similarly. Do
they not work the same way?

beerismygas wrote:
> when brightening my photos which were not taken with flash, when should
> i modify (i use photoeditor) brightness and when do i use gamma? they
> seem similar but there must be a good reason for two seperate controls.

Personally I use gamma correction almost exclusively. I almost never
use contrast or brightness.

Here's some graphs (only if you love graphs) that explain the
input/output correction that occurs with each type of correction.
Basically, gamma changes the "bend" of the curve. Contrast changes the
slope. Brightness shifts the curve up or down.

http://www.astro-imaging.de/astro/gamma.html
http://www.poynton.com/notes/brightness_and_contrast/

ASAAR <> wrote:
>On Sat, 29 Jul 2006 06:09:45 +0000 (UTC), Randy Berbaum wrote:
>
>> As I understand the terms Brightness is the adjustment of how intense
>> White is. This would be like dimming a light bulb. White would still be
>> white but the intensity of that white is higher. Gamma is more a measure
>> of how intense the colors are in relation to eachother.

Yes.

>> . . .
>> If you have experience with older TVs you may have had to play with the
>> controls and had to adjust the "color" knob. That is Gamma.

The "color" knob on an old TV was saturation, not gamma. It
adjusted gain for the chroma signal while the luminance level
remained constant. The "contrast" control adjusted both
luminance and chroma levels at the same time.

> I don't know about that. Based on very limited experience with
>only one program, Gamma seems more of a complex brightness control
>that doesn't affect the color relationship.

True.

>if when the Brightness is increased, each intensity level in the
>image is increased by the same absolute amount, resulting in both
>light and dark areas becoming overly bright with small increases of
>the slider's position. When the Gamma value is increased, it's as
>if the intensity level of each pixel is increased by an amount
>that's a function of the pixel's intensity level, irrespective of
>its color. So by increasing the Gamma, overall brightness can be
>increased, but the darker shadow areas, while brightening somewhat,
>don't brighten nearly as rapidly.

Gamma applies an exponential adjustment to levels.

One problem is that you'll see various different scales for
"gamma", and it is hard to relate them to each other. The
monitor screen is supposedly adjusted for "gamma 2.2". In some
programs whatever "normal" is, is 1 (or maybe 0). And in some
programs "normal" is 0.45 instead. Heres the reason:

gamma_adjusted_value = l ^ (1 / g)

Where l is the uncorrected value, and g is gamma (where the
"normal" value for a monitor screen is 2.2 (or perhaps 1.8 for
an Apple system).

When 1 (or 0) is "normal", that is just normalized to provide a
default that is an unknown value, but where you can scale
increases or decreases by the number.

The "(1 / g)" part of the expression just happens to be 0.454545
if g is equal to 2.2, hence when you see a "normal" value of
0.45, the number is actually the exponent part of the
expression.

How does it work? Well consider an image that has been
converted to digital, where we have a "value" for each intensity
level. In an "uncorrected" or "linear" file it takes 12 bits to
provide a useful range of at least 8 fstops. When "gamma
correction" is applied (the "l" in the above formula is
transformed to the "gamma_adjusted_value"), an 8 bit value can be
used to provide 8 fstops of range.

Here is a chart, that shows the values.

Levels Normalized Pixel Levels
Fstop | 12 Bit Levels 8 Bit 8 Bit JPEG
Range | Linear Linear Gam.Cor. Gam.Cor. Gamma Corr
------|---------------------------------------------------------
1 | 2048 1.0 1.0 255 69
2 | 1024 0.5 0.72974 186 50
3 | 512 0.25 0.53252 136 37
4 | 256 0.125 0.38860 99 27
5 | 128 0.0625 0.28358 72 20
6 | 64 0.03125 0.20694 53 14
7 | 32 0.015625 0.15101 38 10
8 | 16 0.007812 0.11020 28 8
9 | 8 0.003906 0.08042 21 6
10 | 4 0.001953 0.05868 15 4
11 | 2 0.0009765 0.04282 11 3
12 | 1 0.0004883 0.03125 8 2
13 | 0.0002415 0.02269 6 2

"Levels" is the number of values within that f/stop range. I
am arbitrarily designating 8 levels as the least number that is
useful. By that definition the dynamic range of a 12 bit linear
data set is 9 fstops, while a 2.2 gamma corrected data set is
8 fstops.

Note that graphing the "Normalized Pixel Levels" would show what
"gamma" changes, but you can sort of see it just from the data.
At the brightest level, normalized 1.0, gamma *makes* *no*
*difference*.

But the brightness level falls off more slowly as the gamma
correction is increased. With 2.2 gamma (or an exponent of
0.45) the brightest level in the actual scene is 1.0 and a level
two fstops lower would be 0.25, but the corrected levels are 1.0
and 0.53 (which means tones that were 2 fstops down (1/4 as
bright) are going to be displayed at 1 fstop down (1/2 as
bright). The same basic relationship (two fstop drop in actual
brightness is displayed at 1 fstop lower) is true for the entire
range. (It would actually be exactly 2:1 if the gamma was 2.0,
but with a 2.2 gamma it is close.)

Gamma changes the contrast! Which is confusing, because most
monitors have a control labeled "contrast", but it actually
controls "brightness", not contrast. The control labeled
"brightness" actually controls the black level, hence it takes
adjusting both of them to control actual contrast. That is
a holdover from the days of analog monitors. With digital
monitors it is just as easy to actually have an individual
control for brightness, contrast, and black level, but changing
the labels would confuse everyone.

--
Floyd L. Davidson <http://www.apaflo.com/floyd_davidson>
Ukpeagvik (Barrow, Alaska)

> If you have experience with older TVs you may have had to play with the
> controls and had to adjust the "color" knob. That is Gamma.

Not true: turning gamma completely down doesnt result in a b&w picture.
Saturation is the same to the color-knob on the TV.

Rutger

I second Bucky's comments.

*Step away from the brightness control!!*

Use Gamma, and once it looks about right, maybe a tiny tweak of
contrast. Brightness should be avoided except on very dark images
where Gamma just doesn't quite get there... Brightness adjustments can
too easily cause clipping of highlights (ie 'blown' whites).

Play with the two controls, and run them up to extremes to get a good
idea of what they can do.. You'll see a dramatic difference, I think..

On Sat, 29 Jul 2006 02:46:06 -0800, Floyd L. Davidson wrote:

> How does it work? Well consider an image that has been
> converted to digital, where we have a "value" for each intensity
> level. In an "uncorrected" or "linear" file it takes 12 bits to
> provide a useful range of at least 8 fstops. When "gamma
> correction" is applied (the "l" in the above formula is
> transformed to the "gamma_adjusted_value"), an 8 bit value can be
> used to provide 8 fstops of range.
>
> Here is a chart, that shows the values.

Interesting. Thanks for that.

wrote:
>I second Bucky's comments.
>
>*Step away from the brightness control!!*
>
>Use Gamma, and once it looks about right, maybe a tiny tweak of
>contrast. Brightness should be avoided except on very dark images
>where Gamma just doesn't quite get there... Brightness adjustments can
>too easily cause clipping of highlights (ie 'blown' whites).
>
>Play with the two controls, and run them up to extremes to get a good
>idea of what they can do.. You'll see a dramatic difference, I think..

They do *different* things, and each should be used appropriately.
I cannot imagine using only gamma adjustment, and not "brightness".

You trimmed the significant (and very well stated) part of Bucky's
article:

Basically, gamma changes the "bend" of the curve. Contrast
changes the slope. Brightness shifts the curve up or down.

If either the slope or the vertical position of the curve is not
correct, adjusting gamma is simply *never* going to be the right
adjustment to make.

Generally it requires, or at least is much easier with, a
histogram for reference when making the adjustments. It is also
very nice to have a display indicating which image pixels are at
maximum or minimum values (for example, flashing to black at
maximum and flashing to white a minimum). That allows you to
know exactly which areas of the image are being clipped.

Basically, the *first* adjustment should be the "brightness",
which commonly would be used to set highlights to be just barely
below the maximum brightness value. Then one would probably
want to set the slope (contrast) to have the lowest values at
"black level" (what that is depends very much on the contrast of
the display device). Then the "bend" of the curve can be
adjusted, though it might be simply "gamma" which compresses
tonal range by applying an exponantial curve, or it might also
be almost any function(x,y), or it might be a manually set curve
that does not follow any specific function.

Ideally, when done in that order there is less influence on the
first (brightness) adjustment as the last ("bend") adjustment is
made. In fact though, there is often interaction, and it may be
necessary to repeat the process multiple times.

--
Floyd L. Davidson <http://www.apaflo.com/floyd_davidson>
Ukpeagvik (Barrow, Alaska)

Floyd L. Davidson wrote:
> You trimmed the significant (and very well stated) part of Bucky's
> article:

Actually, I trimmed the whole thing... (O:

> If either the slope or the vertical position of the curve is not
> correct, adjusting gamma is simply *never* going to be the right
> adjustment to make.

True. But I find that very few of my images have a brightness level
that is very far wrong - perhaps this is a reflection of very good
exposure metering from my cameras - it can't possibly be my great
technique..!

I did shoot from the hip a little, but I have encountered *so* many
people who try to fix their images without even trying gamma, and who
think that brightness is the only way to lighten the dull areas of an
image.

> Generally it requires, or at least is much easier with, a
> histogram for reference when making the adjustments. It is also
> very nice to have a display indicating which image pixels are at
> maximum or minimum values (for example, flashing to black at
> maximum and flashing to white a minimum). That allows you to
> know exactly which areas of the image are being clipped.

Yes, but most folks who know how to use a histogram also know all about
gamma/curves!

> Basically, the *first* adjustment should be the "brightness",
> which commonly would be used to set highlights to be just barely
> below the maximum brightness value.

I find most decent digitals do pretty well at ensuring that at least
few highlights are up there near the clipping point, or unfortunately,
over it.. hence my simplistic and possibly flawed attitude!

Essentially, I agree with you!