DPChallenge: A Digital Photography Contest You are not logged in. (log in or register) 
 

DPChallenge Forums >> Hardware and Software >> Genuine Fractals worth it? A comparison
Pages:  
Showing posts 1 - 14 of 14, (reverse)
AuthorThread
 03/21/2007 11:38:51 AM · #1
I have been keenly interested in getting the best resizing program available for some gallery shots. I've been using 'Genuine' 'Fractals' but my trial 20 runs are almost up. While working a picture up I decided to compare the GF results with a simple, one-shot resize with CS2's bicubic. I'm fairly surprised at the results. You can judge for yourself, but I think the GF is barely better. $159 worth of better? Not likely. And this is with a simple one step resize on just bicubic.

Here are two 720 pixel crops from the shot. Do you think I just found a picture that didn't represent well?

The original is 4032x2912 pixels(the shot is cropped to a near 1.25:1 aspect ratio) while the full resize is 7443x5605 pixels. (that's a 184% linear enlargement or a 356% enlargement of total pixels). I have not added any sharpening.

Bicubic:


'Genuine' 'Fractals':


Looking for opinions before I drop a load of money on this.
 03/21/2007 11:40:43 AM · #2
the 'fractals' looks slightly sharper, but there's no way I'd pay for that tiny difference.
 03/21/2007 12:07:25 PM · #3
It *could* be that there's not enough detail right at the resolution limit to show up a difference, but I really doubt that. It's a pretty smooth image, but there's plenty of high-frequency transitions in the rocks and I just don't see a significant difference there. Surely nothing that would be visible on a print.
I do remember seeing some other tests namy moons ago that showed similar results. I've never considered buying GF for that reason. Certainly not at that price :-P
 03/21/2007 12:17:38 PM · #4
Before I got CS2, when I was still using PS7, this was a matter of some concern to me. I did my research diligently, and it seemed clear to me that Fred Miranda's SI Pro action was as good as 'Genuine' 'Fractals', and it only cost $19.95, so that's what I went with. Since I got CS2, I have tested and determined to my satisfaction that the native resize algorithm in CS2 is as good as the Fred Miranda action.

So, based on that, I'd say 'Genuine' 'Fractals' is a waste of money, Doc.

Robt.
 03/21/2007 12:24:03 PM · #5
You should make additional comparisons at 300%, 400%, 500%, and 800% enlargements. As I understand it, the place GF may have an advantage is in these "supersizing" enlargements. If you're not upsizing more than 200%, there's probably no advantage to GF over the PS CS BCubic Smoother algorithm.
 03/21/2007 01:01:53 PM · #6
If you look at the edge in the top right the GF is sharper. Try the test with a building with sharp lines, shadows and bright highlights. I think then you will have a better idea. Flowing water over dark rocks is kinda hard to really see.
 03/21/2007 01:53:28 PM · #7
I'll do the same comparison on my next shot I work up. Frankly I don't think the comparisons at 300-800% are worth it as I already think the image degradation is at its limit at the current 350% for my purposes. (I think GF's claim of 800% is via total pixels and not a linear resize).

Another question on CS2 is whether you want to use Bicubic smoother or sharper for an enlargement. I have conflicting information about that.
 03/21/2007 02:18:03 PM · #8
Originally posted by DrAchoo:
Another question on CS2 is whether you want to use Bicubic smoother or sharper for an enlargement. I have conflicting information about that.


The CS3 Image Size dialog has bicubic sharper parenthetically marked as "best for reduction" and bicubic smoother as "best for enlargement. That would square with what I've experienced, though I have not done a tremendous amount of enlargement.
 03/21/2007 02:35:27 PM · #9
I find GF better on some textures, and bicubic on others. So what I do (to get good results on a 24 x 36 print with a 5 megapixel file) is do both conversions, put one on top of the other and erase away the parts I don't like. Time consuming of course, but worth it for a gallery
 03/21/2007 02:41:31 PM · #10
I found this article on Ron Bigelow Photography which compares Bicubic, Bicubic Smoother, a 10% stairstep, 'Genuine' 'Fractals', and something called S-spline at 200%, 300% and 400% for 8-12 MP files.

The results were basically that at 200% there is no difference, at 300-400% it seemed to depend on the image with GF gaining a small edge when an edge was present.

In the end I think this makes me feel better about just using Bicubic (it seeemed to me that I would start with this vs. smoother) to resize for the sizes I'm looking at (12x16 at 460DPI or 24x30 at 250 DPI).
 03/21/2007 03:41:05 PM · #11
Originally posted by Bear_Music:
Before I got CS2, when I was still using PS7, this was a matter of some concern to me. I did my research diligently, and it seemed clear to me that Fred Miranda's SI Pro action was as good as 'Genuine' 'Fractals', and it only cost $19.95, so that's what I went with. Since I got CS2, I have tested and determined to my satisfaction that the native resize algorithm in CS2 is as good as the Fred Miranda action.

So, based on that, I'd say 'Genuine' 'Fractals' is a waste of money, Doc.

Agreed. I bought SI Pro when using PS7 and have the same experience as Robert. I further agree that SI Pro gives as good of results as GF.

Lastly, like Robert I extensively compared CS2's "Bicubic Smoother" option for upscaling images for gallery prints and found not only that it gave results as good as SI Pro, but it is a LOT faster doing it. It is one of the few areas where native CS2 performs as well as third party add-ons.

Don't bother purchasing 'Genuine' 'Fractals'.
 03/21/2007 03:52:44 PM · #12
Originally posted by DrAchoo:
In the end I think this makes me feel better about just using Bicubic (it seeemed to me that I would start with this vs. smoother) to resize for the sizes I'm looking at (12x16 at 460DPI or 24x30 at 250 DPI).

I think it might be a mistake, you should use "Bicubic Smoother" for your enlargements. Most of my gallery prints have been at 12 X 18 and 16 X 20 at 300dpi and my experience and extensive comparisons is that there is a noticeable difference between "Bicubic" and "Bicubic Smoother" in overall print quality, at least with the images I took with the Sony F717. SI Pro did a better job than "Bicubic" which is why I purchased it in the first place but not better than "Bicubic Smoother" which is why I stopped using SI Pro with CS2. I like my images crisp.
03/21/2007 04:22:21 PM · #13
Here was a picture I processed with PSE4 for a challenge:



Then when I found GF I got this (I think it still needs a sharpen and contrast):

03/26/2007 07:04:23 PM · #14
Originally posted by DrAchoo:
I found this article on Ron Bigelow Photography which compares Bicubic, Bicubic Smoother, a 10% stairstep, 'Genuine' 'Fractals', and something called S-spline at 200%, 300% and 400% for 8-12 MP files.


Random interjection. CS2, from the information I've been given, automatically does the 10% stairstep process internally when adjusting image sizes. So, while that may have been best practice with earlier versions, it's no longer necessary from CS2 forward.
Pages:  
Current Server Time: 08/30/2025 03:05:50 PM

Please log in or register to post to the forums.


Home - Challenges - Community - League - Photos - Cameras - Lenses - Learn - Help - Terms of Use - Privacy - Top ^
DPChallenge, and website content and design, Copyright © 2001-2025 Challenging Technologies, LLC.
All digital photo copyrights belong to the photographers and may not be used without permission.
Current Server Time: 08/30/2025 03:05:50 PM EDT.