Is Genuine Fractals worth it? A Photoshop CS vs Genuine Fractals 3.5 showdown at 6 megapixels
At $140 the program is not cheap. The question is, is it worth it? Especially when Adobe Photoshop CS has such a good tool for upresing files with Bicubic Interpolation. So I bought a copy and decided to do a quick test. With my new Maxxum 7D I now have lots of 6 megapixel images to test. The 7D produces a 3008 x 2000 pixel image. Some stock houses want at least 5000 x 3000 pixels, so some uprezing will be necessary. The question is, do the extra steps involved in using Genuine Fractals produce a better image? Or will a quick Photoshop action do the trick?
The Test
For the test I picked a picture almost at random, then enlarged it 400% (to about 12,000 x 8,000 pixels) with both Photoshop CS and Genuine Fractals 3.5. Below are 100% crops of four different parts of the image. I was in a hurry, so the crop areas are not exactly identical, but you can easily compare the two results.
| Genuine Fractals 3.5 | Photoshop CS Bicubic Interpolation |
|---|---|
 |
 |
 |
 |
 |
 |
 |
 |
Observations
At first glance the results are very close. Unless you go to 100%, you won’t tell a difference between the two. At 100% there are slight differences. Genuine Fractals renders the details smoother, in what to my eye appears a more artificial way. The image takes on a hint of computer generated "fakeness". Look at the highlight on the glasses frame in example 1, or the highlights in the sweater knit in example 3. The one place that this tendency to smooth the highlights works well is in the close-up of the hair in example 2. In that crop, Genuine Fractals 3.5 has a slight edge over Photoshop CS. But the tendency to smoothing serves only to eliminate the subtlest detail in fairly uniform areas, such as example 4.
Conclusions
This test is hardly scientific or detailed. It shows the results of manipulation at only one size (400% enlargement) on only one image (chosen more or less at random). And it uses only one starting file size – 3008 x 2000 pixels. Yet this random sample is likely close to what your average dSLR shooter will be using the program on. I have already read other reviews that state that Genuine Fractals does better the more detail it has, and thus would do less well with smaller images such as are produced by a 3 megapixel camera.
At best Genuine Fractals 3.5 does no worse than Photoshop CS. But for $140 that is not enough to recommend the program. Worse, it introduces a smoothing that starts to look fake on close examination. Photoshop CS renders a much more natural looking result that to my eye is preferable overall.
Save your money.