This came up on another posting...if the microsites can 'tell' if you uprezed/interpolated an image.
My thoughts are that when you interpolate/uprez a pic, it doesn't 'add' any detail, it only gets rid of the jaggies. If you started with a low rez image and scaled it way up, what you would see is not so much added 'artifact' but rather that an image at the larger scale should have a certain amount of detail...if it's not there, it wouldn't make much sense that the image was taked at that (higher) resolution. IE the output from a c-sized dslr will  always have more detail than un uprezzed 3 megapixel point and shoot. I suppose the only way they can 'tell' is thus just viewing the image closely, for even with a lousy lens,an image at ,say, 30 meg file should have a certain amount of detail.
 Having said that,there is no discernable difference if used selectively.
 For example, I interpolate my Canon 30d images from it's native 240 dpi up to 300 dpi with Photozoom Pro 2. This program offers a very high quality interpolation engine.I usually do the interpolation last and THEN sharpen the image accordingly (with Focalblade).
 Would be great to hear your thoughts ,knowledge and experience on this as I am only really taking a guess here...

Well in Photozoom it does...you are changing the dpi but NOT the image dimensions. Would that not therefore increase its size?I think you are talking about in PS where when you change the dpi it chnges the image size accordingly.

The neat thing I like about Photozoom is it's elegantly simple interface. You drop an image on it and it shows you the interpolation process as you move about the image....(hard to explain) you should check out the demo. Yes of course it adds artificial detail...but when it comes to digital information there are quite sophisticated algorithms that are quite good at figuring out what those missing pixels are most likely to be (in color and level). A lot of publishers have used it when printing from lower res images...for ex. a lot of photo journalists had to (until recently with higher meg) shoot on quite low rez point and shoots when a SLR or larger camera were not appropriate to use.

Well in Photozoom it does...you are changing the dpi but NOT the image dimensions. Would that not therefore increase its size?I think you are talking about in PS where when you change the dpi it chnges the image size accordingly.

oh on reading this again, i think when your program enlarges an image, it figures out that your image is 1000 X 1000 piels at 1000 dpi then it will print at 1", so if you want to double the resolution, it will just make it 2000X2000 pixels and then change the dpi to 2000 DPI so that it will still print at 1"...

same difference, and yes photoshop can do this as well

and the interpolation engine in Photozoom (s-spline) is much better than PS's bicubic.

Photozoom's s-spline algorithm is just a slightly weighted version of the bicubic interpolation that CS2 uses. Every unbiased comparison I've seen says they are almost identical.

Genuine Fractal's algorithm has always been considered the best interpolation algorithm available.

Check out this comparison: http://www.ronbigelow.com/articles/interpolation3/interpolation3.htm

If you actually read the whole article you'll see that the page I linked to was the only one that you could see any difference at all. It was also at 400% which is way beyond what you could get away with if you wanted to upload it to a stock agency. At the lower levels Photozoom and PS are identical.

I hate to tell you this but s-spline "xl" is more hype then improvement. The mathmatics behind it can't be improved much more then they already had. Photozoom uses a nearest neighbor, weighted, low-degree polynomial algorithm. The only thing they can do is tweak the weighting (which is what they did).

You're welcome to believe whatever you want, but you're going to have a hard time proving to me that splining produces better results than fractalization, when they are done correctly. I wrote my master's thesis on use of fractals in image rasterization.

You do know that you can see a difference though, regardless of which algorithm you use. That was the whole point of the post right? You want to upsize photos for microstock. If you are upsizing more than 20% then I'm willing to bet they get rejected for one of the following: "over filtering", "pixelation", or for the photo being too soft.

yep i have had a number of images rejected because they saw the image come from a canon 10D and it was too large of a file...... HOWEVER... they didn't take the time to realize that it was a collage of a number of images which is why the size was so large

Well...Photozoom Pro 1 had the problem in that it stripped all exif data...something they fixed in version 2. I believe I can edit whatever exif data I want in Iview media pro...so I could always just take out that info if I wanted too. When you mean a 'collaged' image from 3 to 6...do you mean interpolating?

I made collages to make one new image out of a few older images..... not just pasting two images on a white background.

so you could edit the EXIF data to make it look like the image was shot on a Canon 5D

would this work?

Why stop there - now what was the name of that 100MP camera?

I have upsized a couple of images and they have been approved even at IS.  I also often do small upsizing to fit minimum size requirements.  Not all images upsize well, though.  I had a very good experience with ClearerZoom, but I didn't compare it to other software.

Regards,
Adelaide