Genuine Fractals 4.1; Resampling With GF Might Make The Megapixel Race Moot Page 2

sorcadmin's picture
How did Genuine Fractals work? To test I chose a number of different image formats and file sizes, mostly from digital camera files. My motive in doing this was to check out the megapixel horse race attitude that has pervaded digicam and digital SLR marketing. The mantra of "more sensor resolution/bigger prints" is, I believe, something that has driven some people to higher megapixel cameras; from what I saw in these tests it might not always be necessary.

Image files were sourced from a Kodak EasyShare DX7590 (5-megapixel) digicam, a Canon EOS-1D Mark II N digital SLR, and from a 4-megapixel Kyocera (RIP) digicam that resided in my CD files. I worked with an HP 8750 printer (1200dpi black, 4800dpi optimized color) and both HP and Konica Minolta 13x19" paper. All printing was done on virtually Automatic mode and no sharpening was applied to any images after or before resampling.

The simplest way I found to work with Genuine Fractals was to type in the desired print size, here going from 6x7 in output at 72 ppi to 11x12 at 240 ppi. All you need do is click Apply and the program does its thing.

I chose a number of files in the 7-14MB size range, those you'd get either from 3-6-megapixel cameras or from smaller resolution choices in digital SLRs. These included JPEGs in Medium range of 6 megapixels (usually 1-8 compression as well). Ordinarily I'd print these maximum 5x7 to 8x10 in size. Once I opened and did the work on an image I accessed the Genuine Fractals work space in the File>Automate area of Photoshop. (This works in CS as well as in my Elements 4.)

The prints I made startled me. Any flaws that enlargement would ordinarily accentuate, such as camera blur or weak edges, were of course exaggerated by the degree of enlargement, as you might expect. But those artifacts and jaggies that I thought would occur due to over-enlargement, limitations assumed by file size, were simply not there.

Here are some of the resamplings I did that resulted in very good to excellent prints: 7MB to 72MB; 8.3MB to 22.2MB; and 14.2MB to 227MB (!)--I of course re-sized (not resampled) the pictures to fit my 13x19 paper.

What I did not do on this test was work with a large format printer, which I think will be the next stage of my investigations. Indeed, having a large format printer might be the main reason to invest in Genuine Fractals. This would allow you to go fairly hog wild with sizing. But that's not what I was seeking; more, I wanted to see just how well the typical file size from a typical digital camera would respond, and to what limits one could stretch the envelope. I came to a few conclusions, some of which I expected and others that surprised me.

1) The megapixel race is a bit of a myth, given, and this is a big given, that the camera/sensor/lens is optimized, and the quality of the image produced is the best you can get. Will you get better images and larger resampling potential from a better quality camera/lens/image processor? You bet. But for medium megapixel digital SLRs and digicams, the stricter side of the physical print size/megapixel myth should be dispelled. So, when folks say it ain't just megapixels they ain't kidding. (And I like my crow well done.)

2) When you make prints the printer is important. This might sound dimwitted or at least self-evident, but I have found that the more modern the printer, with better dpi ratings and finer alignment, the better the chance that you will get optimal prints. Earlier tech was fine, but printers and nozzles and ink laydown have made quantum leaps of late, and it is reflected in the results I got here. I often wonder if the higher dpi/smaller nozzle printers are as important as megapixels in the camera that made the image in the first place in determining print quality.

3) Know the limitations and keep your expectations "reasonable." Once I began working with Genuine Fractals I thought I could crop away and get the same quality in very large print sizes. There is, naturally, a limit on how big you can go. Yes, it depends on viewing distance, but I don't ever judge prints from across a room, but from arm's length. If it looks lousy at that distance it is lousy, regardless of those resolution/viewing distance charts that may have been waved in front of your eyes. I cropped into one 9MB image to make it about a 6MB image, then re-sized it to 98MB and printed it out. From across the room it was OK, but at reasonable viewing distance the sky came somewhat asunder, although I must say it was not bad, not bad at all in some areas of the print. And of course if you're going really big, then original file size is important.

4) Garbage in, garbage out. I know, that's a cliché. But just as you shouldn't try to cure a poor exposure with Photoshop rather than make a good exposure and carry on from there, you can't expect to have Genuine Fractals sharpen poor focus or fix an image that just doesn't work, regardless the size. And, making something bigger doesn't make it better, despite the Chelsea chic these days.

But when you want to make bigger enlargements than you'd been taught to expect from your more humble megapixel count camera, and don't want to dump your 5- or 6-megapixel digicam to do so, Genuine Fractals just might be the clue. At least it will show you how good your camera and printer work. All of us aim for the best image quality we can get. And yes, megapixels can make a difference in some cases, and will always give you a leg up on super-sized prints. But for those who want to print 8x10 or 13x19 (and I think even 16x20), it seems that you don't have to jump every time the megapixel race comes out of the gate. Using Genuine Fractals is not a cure-all, but it sure is a revelation.

Genuine Fractals 4.1 costs $159.95; an upgrade from previous versions costs $69.95. A trial version is available online.

For more information, contact onOne Software, 15350 SW Sequoia Pkwy, Ste. 190, Portland, OR 97224; (503) 968-1468;

Article Contents
Share | |