Welcome aboard photo enthusiasts. In the last few articles I introduced you to understanding the new camera light meter of the digital age – the histogram.
I mentioned in these histogram discussions that it is best to place exposures slightly right of center. In other words, unlike film (remember film?) for which it was better to underexpose for the shadows and over develop for the highlights, for digital it is better to slightly over expose…I said slightly.
I can simply tell you this is how it is, but I think it’s interesting and worthwhile to give you an abbreviated explanation as to why. Here we go…
To begin, all computer based processing begins with the basic binary digit, or bit. Digital sensors and images are made from bits. Each bit can have two values, either a “0” or a “1”. All cameras can use a jpg format (some also have RAW and TIFF formats available) to process and capture images. This format has a bit depth of 8 bits that can address 256 levels of luminosity. This is arrived at by taking the 2 values (“0”&”1”) to the power of 8, or 28 meaning 2x2x2x2x2x2x2x2, which multiplies out to 256. This luminosity (tonal) range is represented by the chart shown above.
A good exposure at its very best may capture 5 stops of light compared to the human eye, which can capture far more. When we are shooting jpg (not RAW) as all pocket cameras can, we know jpg’s contain 256 bits or tone levels. You can see by the diagram above, which is made up of 5 f/stops that the lightest f/stop contains half of the 256 bits or 128 bit levels available in a jpg. The next f/stop has half of those 128 remaining levels or 64bits, and so forth, with the last (darkest) f/stop having only the 8 bit levels remaining.
This means when you bring your digital photo into your computer to manipulate, retouch and enhance with image processing software (darkroom like) the lightest areas provide the most bits (128) to work with. Making adjustments with programs such as PhotoShop can be destructive to a photograph. Therefore, in summary, if corrections need to be made, there is more bit depth/information in the light areas to play with and potentially sacrifice, as opposed to the dark areas, with the darkest having little information with only 8 bit levels.
What has been explained so far offers a perfect lead in as to why RAW (and TIFF) formats are more powerful than the JPG format. The latter has a bit depth of 8. RAW files have a bit depths of 12 (and 14 working up to 16). Let’s expand the formula of 212 or 2x2x2x2x2x2x2x2x2x2x2x2 = 4096 levels of luminosity. This far exceeds the 256 levels offered by the JPG format and providing a lot more information to work with.
You may have seen in your camera or computer monitor manuals that the images or screen are capable of displaying up to 16,777,216 colors. Wow that’s a lot. We know from this figure it is a JPG file. How?
All color photograph, your color computer monitor (your TV screen) images are all made from the colors red, blue, and green. Multiply their individual luminosity levels together, 256x256x256, and you will come to the figure of 16 million colors.
If we calculate the red, blue, green luminosity levels of an image(s) with a depth of 12 bits (4096x4096x4096) we come up with billions of colors (and levels of color). Of course, we can’t discern all these colors, but the more information retained the more bit levels and colors we are able to sacrifice in processing our images without losing control of the overall quality. More is better.
The same goes for pixel resolution of our image sensors… bigger bits and more bits are better. No doubt there is a point when increasing these numbers becomes negligible and will serve only inflate the size of the files, increase the memory requirements to store them, and add to the extra time it takes to save and upload them, without their being any extra benefits derived.
If and when this threshold will be reached is beyond me. It’s summertime, the living is easy, and it time for fun in the sun with permission to go ashore.