Ad
Ad
Ad
Pages: [1] 2 »   Bottom of Page
Print
Author Topic: Genuine Fractals plug-in issues  (Read 8927 times)
andyptak
Full Member
***
Offline Offline

Posts: 240


« on: January 18, 2008, 10:19:52 AM »
ReplyReply

One of the Stock agenices I use requires a minimum of 48 megs for an 8 bit image. I shoot with Pentax K10D's and a 16 bit TIFF file is roughly 50-60 megs, half that when I drop it to 8 bits. At their suggestion, I'm using Genuine Fractals to bump up the image size and I'm getting a lot of QC rejects for "soft definition". I know the orginals are fine because I have no QC issues with the same shots from other agencies who do not require interpolation.

I'm using the Express option, because I thought it would be simpler and less of a learning curve, but I guess that's been a mistake. Under the option of "Pixel dimension" which works in terms of %, I increase it by whatever amount it takes to get to the file size I need.

Anybody have any experience with this plug in, in terms of settings etc., if I use the full, rather than Express option? Thanks in advance.
Logged
Jonathan Wienke
Sr. Member
****
Offline Offline

Posts: 5759



WWW
« Reply #1 on: January 21, 2008, 06:15:10 PM »
ReplyReply

You're expecting software to fix a problem that can only be solved by capturing more real image data instead of "faking" image information by making educated guesses with software. Some agencies are more stringent regarding the technical quality of their images than others. As you've discovered, trying to fool the more demanding by using software interpolation instead getting a higher-resolution camera has mixed results at best, and futzing with the settings is not going to change that. Capture resolution does make a real difference, or we could all use cell phone cameras and upsize in software to whatever size we needed later.
Logged

andyptak
Full Member
***
Offline Offline

Posts: 240


« Reply #2 on: January 22, 2008, 06:22:28 AM »
ReplyReply

To the contrary. The agency is recommending it's contributors  use Genuine Fractals because it takes a $25,000 Hasselblad to naturally produce 8 bit file of a 50 meg minimum. My problem is not getting caught trying to fake it, it's the learning curve of following their recommendations. I gave enough camera information that you should have known I wasn't using a piece of crap. Apology accepted.
Logged
Jonathan Wienke
Sr. Member
****
Offline Offline

Posts: 5759



WWW
« Reply #3 on: January 22, 2008, 04:24:39 PM »
ReplyReply

Quote
To the contrary. The agency is recommending it's contributors  use Genuine Fractals because it takes a $25,000 Hasselblad to naturally produce 8 bit file of a 50 meg minimum.

Bull$#@*. Both the Canon 1Ds-II and 1Ds-III meet or exceed the minimum file size (16.67MP) with no upsizing. The Mark II can be had for less than $4000 now that the Mark III is out.

All upsizing software like Genuine Fractals can do is spread out captured detail among a greater number of pixels without introducing obvious pixelization artifacts such as jagged edges. You can tweak the settings to make edges crispier, but that will run you afoul of a whole new set if image quality issues (the sharpening artifacts and probably noise, too). If using such programs gets your images rejected for softness, then you are out of luck.

If you're submitting images to an agency with a 50 MB file size minimum, you probably are competing against medium format digital shooters who can reduce their file size by half and still meet the minimum. If you expect to compete under such circumstances, then you should probably switch to a camera that doesn't require you to upsize to meet the minimum requirements. Your camera is not a piece of crap, but you do need to realize that there are DSLR's with twice your pixel count and better per-pixel image quality, and the medium-format shooters have nearly 4 times your pixel count and somewhat better per-pixel quality than DSLRs. Upsizing your images with Genuine Fractals or any other program is not a realistic means to counter the disadvantage you're working under.

Quote
Apology accepted.

Your naivete does not constitute grounds for me to issue an apology. Request denied.
« Last Edit: January 22, 2008, 04:25:54 PM by Jonathan Wienke » Logged

digitaldog
Sr. Member
****
Offline Offline

Posts: 8639



WWW
« Reply #4 on: January 22, 2008, 04:41:44 PM »
ReplyReply

Quote
You're expecting software to fix a problem that can only be solved by capturing more real image data instead of "faking" image information by making educated guesses with software.

Exactly.

That an agency would allow this but disallow a smaller capture device shows how clueless they've become with respect to photography and imaging.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
walter.sk
Sr. Member
****
Offline Offline

Posts: 1328


« Reply #5 on: January 22, 2008, 04:49:32 PM »
ReplyReply

If you are using a PC, or an Intel Mac, you might try a deconvolution program called Focus Magic.  I don't have the link handy but a Google search will turn up their website.  They have a downloadable trial version.

The program is very helpful in countering the blur caused by A-A filters, and by measuring the width of the blur, it can actually reclaim the data within the blur, giving the impression of refocusing the image.  The effect is subtle, and very data-intensive but I include it in my workflow for almost every image.

You might have good results if it were used before uprezzing, or possibly after doing so, or, even before and after.  It might give your files the crisper look.  You have nothing to lose in trying it.
Logged
TravellingLight
Newbie
*
Offline Offline

Posts: 4



« Reply #6 on: January 22, 2008, 07:07:16 PM »
ReplyReply

Quote
One of the Stock agenices I use requires a minimum of 48 megs for an 8 bit image. I shoot with Pentax K10D's and a 16 bit TIFF file is roughly 50-60 megs, half that when I drop it to 8 bits. At their suggestion, I'm using Genuine Fractals to bump up the image size and I'm getting a lot of QC rejects for "soft definition". I know the originals are fine because I have no QC issues with the same shots from other agencies who do not require interpolation.

I'm using the Express option, because I thought it would be simpler and less of a learning curve, but I guess that's been a mistake. Under the option of "Pixel dimension" which works in terms of %, I increase it by whatever amount it takes to get to the file size I need.

Anybody have any experience with this plug in, in terms of settings etc., if I use the full, rather than Express option? Thanks in advance.

That sounds like Alamy, and their recommendation of Genuine Fractals is years out of date. Most people using less than 16MP cameras simply upsize in Photoshop/Lightroom using Bicubic Smoother. If you're using ACR you can do it during conversion by choosing 5120 x 3413 as the output size.

We're still selling pics on Alamy upsized from a 6MP *istD, although they might not get through QC any more. We've only had one K10D pic through QC so far, but there are some more in the queue.

Colin
Logged

Kirk Gittings
Sr. Member
****
Offline Offline

Posts: 1543


WWW
« Reply #7 on: January 22, 2008, 09:01:33 PM »
ReplyReply

Quote
their recommendation of Genuine Fractals is years out of date. Most people using less than 16MP cameras simply upsize in Photoshop/Lightroom using Bicubic Smoother. If you're using ACR you can do it during conversion by choosing 5120 x 3413 as the output size.

The latest version of GF is much improved over previous versions and by my tests slightly outperforms ACR, CS3 or Lightroom for uprezing.
Logged

Thanks,
Kirk

Kirk Gittings
Architecture and Landscape Photography
WWW.GITTINGSPHOTO.COM

LIGHT+SPACE+STRUCTURE (blog)
TravellingLight
Newbie
*
Offline Offline

Posts: 4



« Reply #8 on: January 22, 2008, 10:54:49 PM »
ReplyReply

Quote
The latest version of GF is much improved over previous versions and by my tests slightly outperforms ACR, CS3 or Lightroom for uprezing.
[a href=\"index.php?act=findpost&pid=168909\"][{POST_SNAPBACK}][/a]

That's interesting. Is that at the 1x to 2x order of linear increase? GF used to be better for very large blow-ups, but showed no consistent advantage at these lower orders.

Colin
Logged

andyptak
Full Member
***
Offline Offline

Posts: 240


« Reply #9 on: January 23, 2008, 11:10:15 AM »
ReplyReply

@Travelling Light - Thanks. Nice to hear from someone who's not in a bad mood.

You're right, it is Alamy. Why they do this I don't know, but they have their reasons, I guess. As you've probably guessed, I'm fairly new to digital - been an old film hanger-on for years. My mind is still stuck on, have a decent camera, good lens and a low ISO film, and the rest will take care of itself. Pixels and image dimensions and all of that stuff didn't enter into it - enlargements were all up to the lab and you could blow something up until the film grain ruined it.

Anyway, it's a bit of a learning curve, but I'm enjoying it.  Thanks again.
Logged
Kirk Gittings
Sr. Member
****
Offline Offline

Posts: 1543


WWW
« Reply #10 on: January 23, 2008, 12:18:04 PM »
ReplyReply

Quote
That's interesting. Is that at the 1x to 2x order of linear increase? GF used to be better for very large blow-ups, but showed no consistent advantage at these lower orders.

Colin
[a href=\"index.php?act=findpost&pid=168920\"][{POST_SNAPBACK}][/a]

I only tested it for 2x. As I said the difference is slight. It slows down the workflow over doing it in ACR, but for what I do digitally (architecture, primarily for magazines with a 5D), all aspects of the workflow are critical for best results.
Logged

Thanks,
Kirk

Kirk Gittings
Architecture and Landscape Photography
WWW.GITTINGSPHOTO.COM

LIGHT+SPACE+STRUCTURE (blog)
plugsnpixels
Sr. Member
****
Offline Offline

Posts: 295



WWW
« Reply #11 on: January 24, 2008, 05:03:23 PM »
ReplyReply

I've got some screenshots showing the extreme of onOne's GF effects on a 4mp image here. I haven't done print tests or intentional head-to heads yet, but the same sort of application can be seen for Alien Skin's BlowUp here.

Other related resampling options, some with similar screenshots, are listed here under "Resampling/interpolation".
Logged

Free digital imaging ezine
http://www.plugsandpixels.com
digitaldog
Sr. Member
****
Offline Offline

Posts: 8639



WWW
« Reply #12 on: January 24, 2008, 05:07:34 PM »
ReplyReply

Quote
I've got some screenshots showing the extreme of onOne's GF effects on a 4mp image here. I haven't done print tests or intentional head-to heads yet, but the same sort of application can be seen for Alien Skin's BlowUp here.

Other related resampling options, some with similar screenshots, are listed here under "Resampling/interpolation".
[a href=\"index.php?act=findpost&pid=169343\"][{POST_SNAPBACK}][/a]

Screen shots are basically useless. Print it (and test it next to a Photoshop BiCubic smoother upsize).
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
dandeliondigital
Full Member
***
Offline Offline

Posts: 220



WWW
« Reply #13 on: January 24, 2008, 08:05:26 PM »
ReplyReply

Quote
The latest version of GF is much improved over previous versions and by my tests slightly outperforms ACR, CS3 or Lightroom for uprezing.
[a href=\"index.php?act=findpost&pid=168909\"][{POST_SNAPBACK}][/a]

Hi,
I concur with your findings when comparing GF to PS CS3 uprez.

It's subtle but there is a nice quality about the GF that side by side makes it the winner.

So long for now, TOM
Logged

www.dandeliondigital.com posters and prints
dandeliondigital
Full Member
***
Offline Offline

Posts: 220



WWW
« Reply #14 on: January 24, 2008, 08:08:26 PM »
ReplyReply

Quote
Screen shots are basically useless. Print it (and test it next to a Photoshop BiCubic smoother upsize).
[a href=\"index.php?act=findpost&pid=169346\"][{POST_SNAPBACK}][/a]

Andrew,
Could you explain why?

If I am viewing each uprez at 100% and letting PS CS3 Match Zoom and location?

Thanks, and so long for now, TOM
Logged

www.dandeliondigital.com posters and prints
digitaldog
Sr. Member
****
Offline Offline

Posts: 8639



WWW
« Reply #15 on: January 24, 2008, 08:12:50 PM »
ReplyReply

Quote
Andrew,
Could you explain why?

If I am viewing each uprez at 100% and letting PS CS3 Match Zoom and location?

Thanks, and so long for now, TOM
[a href=\"index.php?act=findpost&pid=169380\"][{POST_SNAPBACK}][/a]

Cause like rubber hitting the road, you have to print out the differences (there's no reason to upsize files to this size just to view them).

In the examples I see of the different interpolation methods above, I see one side that shows very well defined but "pixelated" rendering and one that looks smooth and smeared. The later I suspect is using the GF upsizing methods. Neither looks good. They both look different. How do they look when we do what we'd eventually do with the upsized files, print them out?
« Last Edit: January 24, 2008, 08:13:20 PM by digitaldog » Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
Jonathan Wienke
Sr. Member
****
Offline Offline

Posts: 5759



WWW
« Reply #16 on: January 25, 2008, 09:49:16 AM »
ReplyReply

What you see at 100% on a monitor is not exactly how the print will look. For a good print, one usually needs to sharpen to the point that the file looks somewhat "crunchy" or oversharpened onscreen. The same applies to upsizing software. Artifacts that look unpleasant on the monitor at 100% may look fine when printed, or vice versa. The only way to tell what really works best is side-by-side print comparisons.
Logged

dandeliondigital
Full Member
***
Offline Offline

Posts: 220



WWW
« Reply #17 on: January 25, 2008, 11:44:33 AM »
ReplyReply

Quote
Cause like rubber hitting the road, you have to print out the differences (there's no reason to upsize files to this size just to view them).

Hi Andew,
OK, but can you consider the following? (Please -- I know this may be less scientific and more subjective and narrative, but bear with me). Thanks.

My workflow is to take my file and uprez with each method (with PS CS3 (bicubic or bicubic sharper), Blow up, and GF) and then use PK Sharpener as I usually would on the uprezzed files. I then examined them alongside each other (using PS CS3 Match & Zoom) at 100%, and especially examine any area of importance vital to the success image.

I did this work in the same "spirit" as soft proofing, knowing that the soft proof is not perfect, but looking for the "pop." With soft proofing it's usually subtle color effects. With this uprez/sharpening comparison, I am looking for the image that has more, for lack of a better word "life."

Because we are dealing with rather large images 36x30 inches or larger, and to print out all these versions to arrive at a conclusion would be very costly in time consuming and materials, we thought we had a good solution to this problem.

Are you saying, I really have a faulty workflow, and have made poor choices that really aren't valid because until the "rubber hits the road" you cannot really know what's happening by looking at the screen at 100%?

(ASIDE: Over the past 9 years, I have witnessed that the "uprez" solutions evolve and change. There have been times at which different solutions work better than others, but I have been comparing these on the screen at 100%, and thought that was a safe way to go. I fully understand that at different magnifications Photoshop is aliasing the files in ways that would negate such an approach. But I thought 100% was the real deal. Am I wrong about that?

Thanks, and so long for now, TOM
Logged

www.dandeliondigital.com posters and prints
digitaldog
Sr. Member
****
Offline Offline

Posts: 8639



WWW
« Reply #18 on: January 25, 2008, 11:52:45 AM »
ReplyReply

Quote
Are you saying, I really have a faulty workflow, and have made poor choices that really aren't valid because until the "rubber hits the road" you cannot really know what's happening by looking at the screen at 100%?

Well yes, unless your final output is a HUGE upsized image you want to view on a display (doubtful).

You can print out an 8x10 section of each, no need to output both to an enormous print.

And in the end, there's no substitute for real data. These products are crutches or to be used when there's a gun to your head.

For fun, take the original capture resolution and output an 8x10 section. Then downsize that a great deal, and upsize back using the products. I think you'll see the don't come close to actual captured data.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
plugsnpixels
Sr. Member
****
Offline Offline

Posts: 295



WWW
« Reply #19 on: January 25, 2008, 01:03:59 PM »
ReplyReply

Andrew's got good points, of course. I think of interpolation software as "hamburger helper".

As for showing their capability on a website (such as I do), I chose to use extreme examples for visibility's sake. I wonder if the more realistic printed out side-by-sides would be as visible once scanned and posted online? I don't tend to make prints (I mainly do web and PDF publishing).
Logged

Free digital imaging ezine
http://www.plugsandpixels.com
Pages: [1] 2 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad