How many times have you sweated bullets over a new render – maybe for a Weekend Challenge – and just gave up because “the picture just didn’t look good?”
Maybe the problem wasn’t you, wasn’t Blender, wasn’t the idea … maybe the problem was the picture.Maybe you, basically, asked Blender to make an image that could not, by its design and the lighting and the limits of CRT display-technology, couldn’t “look good.”
Your eye is sometimes a much better “intuitive” judge of what’s wrong with a picture than your mind, without practice, can articulate as a set of actual improvements to be made to the picture. Your eye tells you that “it’s wrong,” but you dont’ know how to improve it.
Let me strongly encourage you to vacuum your local library’s sections on formal studio photography and lighting. Let me encourage you to read every one of Ansel Adams’ “textbooks.” Surf the Internet and read about Ansel’s “Zone System.” Dig deeper to discover how it works.
Basically, what you’re going to discover are a few subtle but very important rules-of-thumb about the design of “a good picture.” To avoid repeating these sources, let me put up a handful of bullet-point highlights …
The brightest and most contrasty part of any picture is going to draw your eye like a magnet. From that point the eye wants to “trace a closed path through” the picture, winding up where it started. If that brightest-spot isn’t where you want it to be, the viewer is not going to “see at a glance” what you wanted him to see.
There are extreme limits to the actual range of so-called “f-stops” (from pure-white to pure-black) that a video screen, or film, can actually capture in a single picture. The so-called “tonal range” is extremely reduced. The eye actually perceives the image by judging the relative contrast of nearby areas. For instance, the screen area you’re looking at appears to be more “black and white” … blacker, and whiter … than it would appear if you turned the monitor off right now. The eye doesn’t like “featureless white” nor “featureless black.” (Your eyes can “see into” real-world shadows, and therefore reject as unreal any detail-free areas.)
Under-exposure can kill a computer-generated picture just as surely as it can a “real” photograph. Look at your photo with the “histogram” tool of GIMP or Photoshop. A broad normal-curve centered on medium-tones is what you want to see; no spikes, no gaps, no all-zeroes or all-ones.
Choice of lens is also important, especially in CG. If the object is detailed and small, put the camera close and shoot wide (low f-stop). Distinct pixel-values on the simulated surface of the object need to get the chance to contribute to the pixels on the image plane.
First, light for “good exposure.” Remember what I said about a limited tonal-range. First you must ensure that all of the areas of your subject are properly illuminated. Then, light for highlights. The specularity and the shadows which help to define edges, depth, and the nature of the object itself actually consist of very small additions or subtractions to the basic light-level of the pixel. The additions are smaller than they will appear to your eye. Always light in (RGB) color, using the blending of the various colors to your artistic advantage.
Light for a bad monitor, not the great one that you have. When the Beatles were recording their stuff at Abbey Road Studios, they had a tinny little speaker which simulated how the song would sound on a transistor or car radio. Be conservative and direct: subtle nuances will typically be lost anyhow.