Genuine Fractals (Perfect Resize) vs PhotoZoom Pro 5

[Paris1968]

I have been a happy user of Genuine Fractals, now called Perfect Resize, for more than four years.  Only lately, I ran into an image that did not upscale with what I considered acceptble results. It was a painting about 1200 x 1800 pixels, eight inches wide, and 72 dpi.  I wanted to increase it to 22 inches in width at 300/360 dpi.  I admit this was a bit of a challenge, but the painting was a good candidate for it with broad areas of solid color, and I was starting with a reasonably clean jpeg. When Perfect Resize did not produce what I wanted, no matter what adjustments I made, I downloaded a trial version of PhotoZoom 5 from BenVista and gave that a try.  I then magnified both images in Photshop CS5 to where the defects in Perfect Resize were apparent and compared the two.  The PhotoZoom version was much smoother with fewer artifacts.   Convinced it was worthwhile, I purchased to software and tried it on other samples with good results.

I am curious about what others have experienced working with these two products, and what anyone could tell me about the settings in PhotoZoom.

[bill t.]

By coincidence I spent a lot of today looking PhotoZoom5 and Resize7, among other things.

The attached file shows what would be part of an image more than 5 feet tall.  The lineup is...

Photozoom5 using SSplineMax__original corrected only for contrast__Resize7

It's tempting to give PZ5 the prize, but on actual large prints I think Resize7 might give the impression of a little more detail being present.

This is another classic case of not enough examples to make intelligent decisions, but more than enough information to jump to conclusions!  Whatever else, as others have pointed out you simply must make test prints to decide what look you prefer.  As with sharpening, the screen is a poor indicator of what you will perceive on the print.

Bottom line for me is that there is not much visible difference if you start with 20mp+ images and make normal sized prints, say up to 24" wide.  And that would include upsizing with the built in CS5 or CS6 upsizing stuff, and even a properly sharpened print without any special blowup processing.

But when you get up to mural sized prints, I think either PZ5 or Resize7 are worthwhile, if only because they remove grungy looking stair-stepping and pixel clumps.  They save the day for giant prints that will be seen close up.

If you are blowing up small jpgs, then the Alien Skin blowup software is probably an especially good choice for its almost complete elimination of jpg artifacts.  PZ5 is almost as good at jpg artifact suppression, and Resize7 much less so.  Persons doing service printing from smartphones and consumer point&shoots would definitely want to consider Alien Skin since it cleans up those kinds of images quite nicely, but I don't think it compares well for already high resolution files.

Those who have used PhotoZoom4 will be happy to know version 5 is much faster!  First thing after install it does some sort of analysis of your video system to take maximum advantage of whatever GPU resources are present.  Even with my geriatric 8600 Nvidia chipset the speed increase was pretty dramatic.

[Bart_van_der_Wolf]

I am curious about what others have experienced working with these two products, and what anyone could tell me about the settings in PhotoZoom.

Hi,

Your findings are similar to mine, Photozoom gives the best results for significantly upsampled output. I've shared some findings in this thread. Also relevant is the discussion about which PPI should be the goal for enlarged output. And here is another related discussion.

As for the (S-Spline Max) settings, there are only a few in PhotoZoom Pro. My recommendation is to be careful with the sharpness slider when doing extreme magnifications. The edge detail can become too sharp looking compared to the rest of the image detail, in which case you can reduce the sharpness amount all the way to zero if needed. The result will still look good.

Having said that, do try and print a crop. You'll lose some resolution in the print process, so some additional/excess sharpness can help, especially when you enlarge to, and print at, the proper native printer resolution.

PhotoZoom does not create halo artifacts during the upsampling, so the optional additional sharpening with USM could be used to beef up some larger detail (perhaps best tuned if the output viewing distance is relatively fixed), although I'd prefer to do that with dedicated tools which offer more control and multiple ranges of detail size (e.g. TopazLabs Detail plugin). Adding a bit of grain can help to give some texture to areas that could look too smooth, and therefore I would use that especially for the huge enlargements. Although the control is called Film grain, I don't think film grain is that useful for printing, but grain is (and that is more like what PhotoZoom produces). We do not need actual new noise 'detail', but just a suggestion of more subject surface micro-detail from light and shadow.

The Artifact Reduction control is mainly useful for images that already had artifacts (e.g. stair stepping of edges and lines) to begin with, which will get accurately magnified by PhotoZoom.

Cheers,
Bart

P.S. I just tried to also do some of those resampled image crops in the referenced threads with Perfect Resize for comparison, but it crashes as a standalone application (should not be affected by its most recent Photoshop update related issues).

[Paris1968]

Thank you for the links to those threads. The discussion of what resolution works best was interesting, even if nothing I do comes out of Lightroom.  I have little, if any, control over the resolution of the image I start with.  I began putting everything into 300 dpi when Canon printers were all I used. Now, if there is a print that I know I am going to use an Epson printer on, I'll resize the image to 360, but I still strike a copy at 300 dpi.  The reason is that I want all of my files to be the same resolution.  If I am putting together a composition of any number of images, starting out with everything at the same resolution makes it so much easier.

Have you found a use for any other resizing method other than S-Spline Max in PhotoZoom?  I've been playing with it for a day or so, and  S-Spline Max works so well with anything I try it on I wonder why the other choices are there.  

I too have had stability issues with Perfect Resize.  It will lock up and crash when I use it as a Photoshop plug-in so often that I adopted the habit of making sure I had saved everthing before using it.  That was my experience with it on three different computers using XP, Vista, and now W7. PhotoZoom is much quicker, and does not seem to tax my computer's resources like Perfect Resize does.

[Bart_van_der_Wolf]

Have you found a use for any other resizing method other than S-Spline Max in PhotoZoom?  I've been playing with it for a day or so, and  S-Spline Max works so well with anything I try it on I wonder why the other choices are there.

Since they added the S-Spline Max algorithm, I've not used anything else for upsampling (usually to 600PPI, the native resolution of my Canon printers, or 720 when I outsource the printing).

Therefore the other algorithms are of limited (historical) use, except perhaps the S-Spline XL (the previously best one they had to offer) for a rare type of image content. Strangely, the downsampling quality of Photozoom is not very good. Even the traditional resampling methods like Lanczos or Mitchell produce a lower quality than a straight forward implementation like in  free applications such as Irfanview or RIOT (Radical Image Optimization Tool). But then I only use Photozoom for upsampling, S-Spline Max is basically all I use, and it does it so well that that alone justifies the price for me.  

Cheers,
Bart