"frankg" wrote in message
…
any thoughts on whether it’s better to use Bicubic Smoother or Genuine Fractals (PrintPro) to enlarge files from a digital camera ?
Depends on how much you resample (which digicam to what output size), and what type of images you shoot.
I’m talking about approx 150% up
Then it would be hard to see much of a difference with adequate sharpening after interpolation, assuming you are sure about the 150%.
To make sure, as an example let’s take an image from an EOS-350D (3456×2304 pixels, 8MP, physical sensor size = 22.2×14.8mm). Printing such a file at the native inkjet printer resolution (@600 or @720 PPI) on an 8x10in would require 4800×6000 or 5760×7200 pixels so upto 250% for an uncropped image
file. For that amount it will become necessary to have very good interpolation.
If you also consider that such an 8 inch output size results in a magnification of the on-sensor resolution (theoretical maximum = 77.8 lp/mm) of 203.2 / 14.8 = 13.7x, the factual resolution will have dropped to 77.8 / 13.7 = 5.7 lp/mm, which is considered to be on the lower side of the pro quality range at normal viewing distance. 8 lp/mm is considered about as good as photochemical projection print and close to human visual acuity, inkjet printers can do better (11.8 to 14.2 lp/mm).
A poorer interpolator will degrade the quality faster than the better interpolators will.
I have since also heard about PhotoZoom Pro.
Yes that’s one of the better interpolators, although I prefer Qimage for printing because it offers a much better workflow, and at a lower price.
I understand that they use different "up-rez" techniques but when I look at a side by side comparison of the same file enlarged by the same amount, I cant visually see a superior result. Are there instances where this may be more obvious?
First make sure there *is* enough image detail to make a sharp print. Resampling a low resolution image will most likely not make much difference, no matter what method is used.
Also, if you leave part of the resampling up to the printer driver, it becomes a black box without control over resampling method and subsequent sharpening. If the image has lots of critical detail in it,
e.g. nature photography or some kinds of fashion photography, you’ll
need a realistic transition going from in-focus to out-of-focus. If images have large smooth gradients, you’ll need a method that avoids posterization. A lot depends on the quality of the source material.
What do you look for in an image to decide which program should be used to "up-res" it ?
I do all my (automatic) resampling for print with Qimage. It produces good results from most types of images with it’s "Pyramid" method.
Bart