I'd appreciate some experienced and expert views on the differences, if any, between these two resizing apps, particularly how each performs in the 300% to 600% range.
Other app recommendations are also appreciated. I'm on a Mac Intel, dual processor.
I compared BlowUp to Genuine Fractals a while ago, and I found there were more similarities than differences in the end. In either case, adding a subtle amount of grain after enlargement with Alien Skin Exposure did wonders for creating the illusion of added detail. You should really consider adding Exposure to your shopping list if you're regularly going to be enarging.
I don't think either one will do much better than using Photoshop.
Depending on the photo, Photoshop can do a good job of enlargements, but I found that both Genuine Fractals and Blow Up were more pleasing on large prints (17x22 and up). Photoshop has a particularly hard time with edges. It likes to make staircases instead of smooth lines.
Dan Martin wrote:
You should really consider adding Exposure to your shopping list if you're regularly going to be enarging.
Both Genuine Fractals and BlowUp have grain generators of their own; the price of Exposure is a little steep just to add grain to interpolated images.
Photoshop can add some decent grain of its own, as well.
Wayne Fox wrote:
I don't think either one will do much better than using Photoshop.
For smaller enlargements or much larger enlargements, I'd agree. In the 300% to 600% range, though, tools like Genuine Fractals and BlowUp can shine because they purposely detect and retain edge detail. Photoshop treats every pixel identically. Depending on what you're working with, a dedicated interpolation program can make a huge difference in perceived sharpness, perceived detail, and the like.
Which brings me to my advice: Download the demos of both tools and try them out. They work differently, neither one really "better" than the other. See which interface you like, which one has a speed you can handle, and which one works best with your images at the sizes you plan to print. And be sure to try them against Photoshop's bicubic interpolation, too. You may find for what you're trying to do, neither tool achieves what you're hoping for—in which case you save $200.
I generally think of PS as up to 2x linear enlargement. It is fine with my D200 and 16x24s @ 300ppi. Also try QImage and as a hoot Irfanview, a freebee that has a good enlargement algorithm. QImage has always scored high in testing. I don't use QImage in my own work for the reason cited above, PS is fine for what I do.
colinm wrote:
Both Genuine Fractals and BlowUp have grain generators of their own; the price of Exposure is a little steep just to add grain to interpolated images.
Photoshop can add some decent grain of its own, as well.
Yes, I bought both, and I started with just BlowUp before adding Expsure to the kit. BlowUp's grain control is very limited whereas Exposure gives you control in spades. It's an expensive kit for sure, but I had a project that required a lot of enlargements and the results were exceptional. I didn't get anywhere near the same results with Photoshop's grain controls...
Ive used GF PPro. The fractal-look is unique and works for some images very well.
Tell us about your imagery. I ask bc for landscapes, I like GF very much! If you think of a a tree with each leaf seen in detail ...THAT is very good fodder for fractal interpolation imho
I wanted to revive this thread as I'll be doing some 20x30's from my 5D and 16x24's and maybe some 20x30s from my 20D.
Will GF or Blow up make a significant difference? These are going to be fine art gallery prints, so quality is important
Next.. where do you place sharpening in your workflow if using GF or Blowup/Exposure? Before? So after doing all the processing, sharpen, then upsize/sharpen all in one?
One more vote to try Qimage. It doesn't add grain, but it interacts with the driver and thus resizes to the exact required ppi for the driver not to do the resampling once again.
Qimage will resize (on the fly upon printing) to 600ppi for Canon and HP printers and to 720ppi for Epsons as they state those are the drivers' actual working resolutions. Also, there are several advanced resizing algorithms to choose from.
Not to mention the rest of the benfits Qimage offers for printing, color management included.
agedbriar wrote:
One more vote to try Qimage. It doesn't add grain, but it interacts with the driver and thus resizes to the exact required ppi for the driver not to do the resampling once again.
Qimage will resize (on the fly upon printing) to 600ppi for Canon and HP printers and to 720ppi for Epsons as they state those are the drivers' actual working resolutions. Also, there are several advanced resizing algorithms to choose from.
Not to mention the rest of the benfits Qimage offers for printing, color management included.
That means that I can't sharpen and add grain after resizing though.. which doesn't really meet my needs.
Sharpening after resize can be done (USM or Smart & amount), although you can't control its effect on screen.
I do the normal sharpening under visual control with Focal Blade before printing, while within Qimage I add a minimal amount (my own default), just right to compensate the softening effect the resizing would have introduced.
Floris, do please satisfy my curiosity over grain-adding, as it's the first time I hear about that option - always the opposite.
Does it mean that when you resize for a very large amount, the "made-up" parts of the picture are lacking the noise present in the "true" sections of the image, and the difference can be spotted?
agedbriar wrote:
Floris, do please satisfy my curiosity over grain-adding, as it's the first time I hear about that option - always the opposite.
Does it mean that when you resize for a very large amount, the "made-up" parts of the picture are lacking the noise present in the "true" sections of the image, and the difference can be spotted?
No 'made up parts' to my images (with the exception of occasionally adding canvas to a bird shot with a very smooth background).
I find that upsizing an image (say around 16x24 for a 5D) you can tell that it's a digital image.. there's digital artifacts, it just has a pixely smooth flat feeling to it. Adding a touch of film grain will give the image a more natural texture. It basically hides the digital upsizing artifacts and softness by creating the illusion of detail. Of course, you have to be careful, too much, especially in smooth areas like sky, and it just looks like you got lots of noise. But a little bit of texture is nice So for me it's a critical stage. I think I might just stick with PS's bicubic various sharpening techniques and a simple grain layer..
Mr Mouse wrote:
Why is that Qimage can interpolate images up in size and save image files. You should able to do anything you want to the interpolated image files.
Ok, I assumed 'on the fly' meant, when you go to print it's kind of built into the process, so that it was less of an action, and more of a printing tool. I like to control it every step of the way. I suppose I'll check it out.
Oct 09, 2008 at 09:30 PM
Mark Metternich Offline • • • • • • Upload & Sell: On
veroman wrote:
I'd appreciate some experienced and expert views on the differences, if any, between these two resizing apps, particularly how each performs in the 300% to 600% range.
Other app recommendations are also appreciated. I'm on a Mac Intel, dual processor.
Thanks,
Steve
I spent months studying this issue comprehensively and did the tests and everything. Adobe Photoshop Bicubic SMOOTHER is the winner. Produces the least amount of artifacts into an image which allows better and more sharpening in the end, which produces the best print.
mark70x70 wrote:
I spent months studying this issue comprehensively and did the tests and everything. Adobe Photoshop Bicubic SMOOTHER is the winner. Produces the least amount of artifacts into an image which allows better and more sharpening in the end, which produces the best print.
How much of an enlargement do you think is possible with this method? Over 300%-400%?
mark70x70 wrote:
I spent months studying this issue comprehensively and did the tests and everything. Adobe Photoshop Bicubic SMOOTHER is the winner. Produces the least amount of artifacts into an image which allows better and more sharpening in the end, which produces the best print.
What happens when the printed area is not the same as the one input in Photoshop for resizing? In fact, the actual printed area never matches the nominal dimensions of the paper - there is either borders or an unknown width of overspray.
Will the driver be happy with the submitted pixel count, or will it do another resize on the fly, to exactly match the number of pixels to the true printed (or sprayed over) area?
I'm asking this because I assume that the printer (the printing head/paper positioning being tied to the digital printing grid) can not tollerate one single pixel mismatch with regard to the printing template, which is determined from the internally computed actual printing area in steps (accuracy) equal to the native printer resolution.
