To Fractal, or not to Fractal an M8 file

[portocar]

Hello

I would like to know, from those that use Genuine Fractals to UpRez their M8 files prior to printing, which is better:

1) From an M8 file, on an Epson printer, I can make a 16x20 print if I set the output resolution to 180 DPI.

2) Or, I can make a similar sized print, if I use Genuine Fractals to UpRez the image and then print with an output resolution of 360 DPI.

My question to those who have performed these two printing options, is, which do you think is better when making a 16x20 print from an M8 file?

1) Print the file at 180 DPI

2) UpRez the image with Genuine Fractals and print the file at 360 DPI

Thanks

 

[ruslan]

Yes, Genuine Fractals is a great add, I'm using GF all the time and large prints looks good.

 

[eleskin]

I print 17"x22" without Uprezing

I print 17"x22" without Uprezing

Honestly, I do not need to up rez and I print 17"x22" on my Epson 3800. For 40" prints and larger, yes. Some on this forum have increased pixel resolution with great success. It seems the M8 files are so good, that you can't even tell if they went through Alien Skin Blowup or similar software.

 

[MCTuomey]

why not see whether a GF trial is available and do the test yourself on the subject matter you intend to shoot? or do the up-rez in PS (which is quite competent imho) and then do the tests at 180dpi vs 360 dpi. in the latter case, you could assume that GF would improve the up-rez outcome to some degree so you'd have a basis to consider whether GF would be worth the trip.

in either case, i think you could get a good idea of the results by prepping the two files as you laid out and then comparing them on a good monitor at 50%. 50% gives a reasonable approximation of print quality in my experience.

i don't think generalities are useful here (all due respect to the response from *eleskin*). your answer will likely vary with subject matter (e.g. a portrait versus an architecture shot), intended viewing distance, etc.

 

[RichC]

I agree with Eleskin.

The short answer is that if you have a technically good image file - sharp, well exposed, low noise, no camera shake, etc., you won't notice any difference between 180 ppi and 360 ppi by eye. You might be able to see minor differences with a loupe - perhaps a tad more detail in some areas.

I typically make A3+ (18 inch) prints for exhibitions and sale, and usually don't bother to resample because the differences are so small.

Much below 180 ppi and I do resample.

Also, there is a penalty to pay with GF and other interpolation methods: they can add artefacts, especially to patterned or textured areas. I recall trying out GF on a photo that included visible sand grains - GF turned the grains into angular, geometric clumps! There's no such thing as a free lunch!

I spent a day testing various interpolation methods last year. It was noticeable that the content of a photo affected the performance of the methods: some types of image suited certain methods, and conversely some methods worked less well on certain images. The main problem was the production of artefacts. I can't recall the details now, unfortunately, but I do recall being unimpressed with GF. Two methods shone above the rest: (1) SizeFixer - usually gave the best results with the least artefacts; (2) a Photoshop technique - this often rivalled, sometimes equalled, and occasionally bettered (fewer artefacts) dedicated software. The Photoshop technique is as follows: resize a finished, capture-sharpened image to 120% the target size with bicubic smoother, sharpen further if needed, then resize to the smaller, final target size using bicubic sharper. More info here: http://www.outbackphoto.com/workflow/wf_60/essay.html

For general resizing I use the Photoshop method, but if I need to make exhibition-quality prints or very large prints, I use both SizeFixer and the Photoshop method, and pick the one that gives the best results. Sometimes I'll combine the two results in Photoshop as layers, and select the best bits of both.

 

[sojournerphoto]

A couple of points - upressing will not and cannot add detail. It can improve the look of a print over the internal upres that the print driver would do and it can also add artefacts that may be mistaked for detail or 'sharpness'.

Having got that out of the way🙂 I almost always print through Qimage (windows only) whcih I use to resize the input file to the native input resolution of my print driver (600dpi in my case, but the epsons are mostly 360 or 720 now). Qimage has a range of upressing algorithms available and also applies a bit of usm sharpening to the output file, which I turn down a bit as I find the default setting a bit crunchy. IT also does the colour management from working sapce to printer profile.

Just another option.

Mike

 

[thompsonks]

Like Mike, I always res up to 600 ppi, the native resolution for Canon wide-format printers & the correct resolution for their Photoshop printer plug-in. But I use Genuine Fractals for large prints. When I had time to test them, GF was the best interpolation engine for Mac users. (As Mike said, QImage is Win only.)

I use Photoshop's Bicubic Smoother for prints up to 10x15, & GF for larger prints.

Note of caution: methods of res-ing up & methods of sharpening interact in some pretty serious ways. When I tested them, GF worked well with the PhotoKit Sharpener suite, but GF + Nik produced artifacts that looked like crinkled tinfoil. So be sure to try interpolation engines & sharpeners in pairs.

Kirk

 

[portocar]

Kirk

Do you sharpen before or after you resize the image with GF?

 

[kbg32]

Are you processing RAW files? Everything can be done through Photoshop with no loss of quality.

 

[Tom Niblick]

I agree with Keith. Small bumps in size (up to 20x30 at 240 or 220) from RAW files work fine from Photoshop. I save my Alien Skin Blowup for when I need a miracle rescue. I'm not wild about the artifacts the ASBU program induces if I use its sharpening or grain functions.

 

[thompsonks]

Hi Portocar,

Photokit Sharpener does 'before-during-after' shaarpening: Capture sharpening for the DNG, Creative for areas, & Output for ... output.

M8 files can use light capture sharpening (opacity reduced to 66%); M9 files look worse if you do any.

Creative sharpener: used with discretion.

Output sharpener: Always, but at lower opacities for small prints.

Kirk

 

[Paul Roark]

I've posted an example of what Genuine Fractals does for M9 images at http://www.paulroark.com/BW-Info/GF.jpg

Note that these are small sections of files, not scans of the final prints.

My tentative conclusions are that GF is very good at preserving and sharpening edges, and it appears to have a better "signal to noise" ratio than using the Photoshop methods. That is, GF seems to not sharpen the noise, whereas the other methods do. However, the color images, as opposed to the B&W ones I work up, seemed to have some color changes that were not very pleasing. I've also found that on most images it makes no visible difference in apparent final print sharpness.

So, bottom line, GF looks like an interesting and sometimes useful tool, but it's no panacea. I'm glad I have it in the tool box.

Paul
www.PaulRoark.com