Some digital camera reviews are illustrated with "image samples" which are not what they claim, or seem, to be. These are not actual pictures of real-life objects, but just shots (reproductions) of images taken with other (possibly film) cameras. Pictures of pictures, in short.

One may argue that this may make the results easier to compare between various cameras, but the method is, unfortunately, flawed and, I believe, close to useless for the purpose it is supposed to serve.

• A real outdoors scene may have the tonal range of 1:10000, while a paper print of it — 1:100. The "sample" produced by photographing the print is useless in judging how the camera handles the scene tonal range, especially at its extremes.
• The color rendering of the final image is affected by that of the original photographic process, and by the light used in the reproduction, in addition to whatever the camera does. Again, the result has almost nothing to do with how would the camera handle the colors of the original scene.
• Possibly the toughest part of camera's color handling is how it translates a given "pure" color (i.e. light of a given wavelength) to its RGB representation (i.e., a mix of three wavelengths, resulting in the same subjective reception by a human viewer). Thus, for example, photographing a yellow flower is something entirely different than photographing its picture, where yellow might be produced by mixing of red and green. Both procedures may give very different results.

  (Really, I am not quite accurate here, but this is as far as I can go without discussing the additive and subtractive methods of representing colors with primary hues, and various implementations of those methods in print and transparency technologies.)

• Any unsharpness and grain (noise) of the print being reproduced are added to those generated by the camera used to reproduce it.
• The apparent depth of field in the reproduced photo depends exclusively on the first camera used in the process; the second one photographs just a flat object.
• When samples are supposed to illustrate flash use (yes, I've seen such), the whole approach just becomes a monstrous joke. Which flash? The original one, used to photograph the object, or the one used in taking the picture of a picture?

To make things worse, the offending parties usually do not make it clear that their "samples" were made by reproducing pictures taken with other cameras, and this is grossly misleading. No names, please, but the culprits include at least one of the generally recognized "big" digital camera review sites on the Web.

Anyway, before trusting any of sample pictures you encounter, especially on Web sites trying to follow all models on the market (admittedly, not an easy task), check images offered there for other camera models. By a quick look at some detail (for example, tree branches) you may clearly say if these are real-scene pictures or just fakes. In the latter case, the images are useful only to evaluate only how good is the camera for reproduction work, nothing else.

Someone had to tell you.