Comments

Chienworks wrote on 4/17/2004, 10:00 AM
Contrast is a linear thing, it affects the entire brightness range evenly. It's downfall is that if you raise the contrast too much then highlights and shadows fall off the ends of the brighness curve and are lost. If you look at the typical brightness contrast graph with source values across the bottom and output values up the side, contrast controls how steep the line is.

Brightness is also a linear thing. On the graph it controls how far up or down the line is. As with contrast, if you increase brightness the highlights are lost and if you decrease it shadows are lost.

Gamma introduces a curve into the brightness graph. You can use it to change the brightness of some levels more than others. You could use a gamma curve to darken the highlights without losing the shadows, yet still have white be white and black be black, etc. On the graph you can "tug" the line up or down in different areas and still leave the endpoints tacked at black/black and white/white.
riredale wrote on 4/17/2004, 10:10 AM
I'll give a layman's description. I'm sure others on this forum can provide a more accurate description, but this is probably not far off the mark.

Do you know what a "histogram" is? If not, imagine you take each pixel in a monochrome picture and plot its value on a graph that runs from 0 to 255. In other words, a pixel that is completely black would go in the "0" slot, while a pixel that is completely white would go in the 255 slot. A pixel that was medium gray would go in the 128 slot. When you've finished with all 345,600 pixels (480x720 in the NTSC world), you now have a graph showing the relative brightness levels of all the pixels in your image.

Now suppose you shot the image on a very cloudy day, and the blackest pixels only range down to, say, 50, and the whitest pixels only range up to 200. To give your image more "snap" just increase the contrast. The computer will, in effect, "stretch" your histogram so that the blackest pixels will reach down to the 0 slot, and your whitest pixels will now reach up to the upper end of the chart (255). Your image will now have blackest blacks, whitest whites, and all greys in-between.

Gamma is a slightly different concept. Rather that stretching out your histogram evenly, as in the previous example, you now selectively stretch mostly the lowest part of the histogram. In other words, while "black" pixels will still remain black, pixels that are "almost black" will now appear to be dark grey. So suppose you take a flash photo of a party in a darkened room. Your subject is well-lit, but everything behind her is just a bunch of blackness. By applying gamma, suddenly that blackness becomes much more detailed; you can now make out some furniture, the walls, and so forth. You've been able to recover "shadow detail."

Hope this helps. You can really see these effects with the Histogram Video scope.
johnmeyer wrote on 4/17/2004, 7:01 PM
Here's another layman's view:

Brightness increases or decreases the brightness of every single pixel by the same amount. If you make the picture brighter, a pixel that is almost as bright as it can get will become completely bright, and any detail will be lost. Same thing happens in the other direction if you make the picture darker.

Contrast is like brightness, except it makes all the pixels that are brigher than 50% brigther, and all pixels that are darker than 50% darker. Again, details are lost.

Gamma only affects pixels in the midrange, but as you get closer to the brighter or darker parts of the image, it makes les and less of a change. A pixel that is almost completely dark or bright will not be changed at all. Thus, the highlights and shadows are preserved.

You need to use the Histogram Videoscope when using any of these controls. It will give you very accurate feedback as to whether you are losing any highlights or shadows.
kentwolf wrote on 4/17/2004, 9:11 PM
Thank you all very much!

You all taught me a lot!
24Peter wrote on 4/18/2004, 12:47 PM
Very helpful description guys. Thanks!