Luminous Landscape Forum

Raw & Post Processing, Printing => Digital Image Processing => Topic started by: ttrask on May 05, 2007, 05:23:42 pm

Title: genuine fractals
Post by: ttrask on May 05, 2007, 05:23:42 pm
I was wondering how the genuine fractal plugin compares to bicubic smoother for uprezzing images for large prints. TWT
Title: genuine fractals
Post by: DiaAzul on May 05, 2007, 07:00:10 pm
Quote
I was wondering how the genuine fractal plugin compares to bicubic smoother for uprezzing images for large prints. TWT
[a href=\"index.php?act=findpost&pid=115892\"][{POST_SNAPBACK}][/a]

Best way that I can describe is that bicubic smoother the final up-ressed image looks as if a gaussian blur has been applied to the image, whereas with genuine fractals it can look as if a median filter has been applied to the image. You end up with more pixels, just depends on which look you prefer - soft and fuzzy, or defined edges but painterly.

There is a new genuine fractals which may give a better defined image, but haven't tried it yet.
Title: genuine fractals
Post by: Ray on May 05, 2007, 07:08:57 pm
In my experience, the differences are very subtle unless the uprezzing is huge. I always considered Genuine Fractals to be a useful program for making poster size prints from small negatives and low rez sensors. The effects then are more noticeable. Certain edges in the image, around leaves, window frames, building outlines, rocks etc, appear sharp, although the detail within the objects is no better than bicubic can produce.

When the uprezzing with GF is taken to extremes, the image takes on a 'painterly' effect and you eventually arrive at something that begins to resemble an impressionistic painting.
Title: genuine fractals
Post by: thompsonkirk on May 05, 2007, 09:10:39 pm
The answer depends mostly on what you mean by large prints - what range did you have in mind?  

If you're starting with hi-res files & going to 16x24 or 20x30, GF will yield clearer results than Bi Smoother & will not generate noticeable artifacts such as mentioned above.  

MEthods of res-ing up interact with methods of sharpening.  Artifacts will be minimized if you use both capture sharpening & output sharpening.  Otherwise you get the problem mentioned above, for example well-outlined trees with mushy leaves.  

GF just released a Universal Binary version (ver 5).  It's still a slow program - lots of computation - so you need generous processor speed & RAM to keep it from boring you.  It works well in conjunction with PhotoKit Sharpener, but not with Nik.  GF & Nik Sharpener together can create artifacts that look like crinkled tinfoil.  

QImage is a faster alternative, but is PC-only. GF is  a good choice for Mac.  

Kirk
Title: genuine fractals
Post by: dkeyes on May 06, 2007, 01:02:13 am
Quote
I was wondering how the genuine fractal plugin compares to bicubic smoother for uprezzing images for large prints. TWT
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=115892\")
Here are some articles with test comparisons between programs for doing uprezzed files.
It basically said that 200% or under enlargements, you can use any program (bicubic smoother in Photoshop), and they all look similar. Over 200% and especially at 400%, Genuine Fractals really starts to stand out as the best program.
- Doug

Ron Bigelow's articles on interpolation:
[a href=\"http://www.ronbigelow.com/articles/interpo...terpolation.htm]http://www.ronbigelow.com/articles/interpo...terpolation.htm[/url]
http://www.ronbigelow.com/articles/interpo...erpolation2.htm (http://www.ronbigelow.com/articles/interpo...erpolation2.htm)
http://www.ronbigelow.com/articles/interpo...erpolation3.htm (http://www.ronbigelow.com/articles/interpo...erpolation3.htm)
Title: genuine fractals
Post by: francois on May 06, 2007, 02:55:20 am
Quote
Here are some articles with test comparisons between programs for doing uprezzed files.
It basically said that 200% or under enlargements, you can use any program (bicubic smoother in Photoshop), and they all look similar. Over 200% and especially at 400%, Genuine Fractals really starts to stand out as the best program.
- Doug

Ron Bigelow's articles on interpolation:
http://www.ronbigelow.com/articles/interpo...terpolation.htm (http://www.ronbigelow.com/articles/interpo...terpolation.htm)
http://www.ronbigelow.com/articles/interpo...erpolation2.htm (http://www.ronbigelow.com/articles/interpo...erpolation2.htm)
http://www.ronbigelow.com/articles/interpo...erpolation3.htm (http://www.ronbigelow.com/articles/interpo...erpolation3.htm)
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=115926\")
Correct links:
[a href=\"http://www.ronbigelow.com/articles/interpolation/interpolation.htm]http://www.ronbigelow.com/articles/interpo...terpolation.htm[/url]
http://www.ronbigelow.com/articles/interpo...erpolation2.htm (http://www.ronbigelow.com/articles/interpolation2/interpolation2.htm)
http://www.ronbigelow.com/articles/interpo...erpolation3.htm (http://www.ronbigelow.com/articles/interpolation3/interpolation3.htm)