Just wanted to share some results from a new program I am trying. I know I have seen this program discussed a little before but never saw any kinds of examples posted.
I am impressed if I do say so. The examples below are 640x480 100% crops from the two files. The first is the original out of the camera (Canon Digital Rebel) and the second is the resized image resized to 20in. x 30in. @ 300ppi. I'm getting much better results with this program than what I was getting using the 5% incremental steps method.
640x480 Crop of Original File (2048 x 3072 @180ppi)
640x480 Crop of Resampled File (6000x9000 @300ppi)
Please let me know what you think of the results. I trial version of this software can be downloaded from //www.extensis.com.
I may post some more results as I continue to try the software if anyone is interested. This software is going to make it a breeze to resample photos for use with DPCPrints. I really am impressed with it so far.
I hope I don't sound rude, but from your samples the quality still looks somewhat poor.
I'd be interested to know, though, if you've tried to print anything that's been "blown up" with this?
I don't know...I think there's just some limit to what software interpretation can do.
There are companies out there that will make a slide out of a digital shot. It's expensive, but can be done. Would it be better quality to have a shot captured to slide, and then scan that slide at a super-high resolution?
I know that there will be a degradation of quality that way, but I can't imagine it would be worse than what a software program can "guesstimate." I'm not familiar enough with large-format printing to know.
I hope I don't sound rude, but from your samples the quality still looks somewhat poor.
I'd be interested to know, though, if you've tried to print anything that's been "blown up" with this?
I don't know...I think there's just some limit to what software interpretation can do.
There are companies out there that will make a slide out of a digital shot. It's expensive, but can be done. Would it be better quality to have a shot captured to slide, and then scan that slide at a super-high resolution?
I know that there will be a degradation of quality that way, but I can't imagine it would be worse than what a software program can "guesstimate." I'm not familiar enough with large-format printing to know.
The best possible algorithm can't invent information, the goal is to produce something that looks smooth, retains all detail present, and avoids artifacts. the results that Bill presented certainly meet that goal.
Transferring to film & scanning would yield significantly poorer results, since noise would be introduced twice, and of course the resolution would be no better.
I think a really great test of the software interpolation methods would be to use a high-contrast, variable freequency line pattern with lines at angles slightly off vertical or horizontal. the best algorithms should retain detail near the nyquist limit of the original file, and avoid aliasing, moire, and "jaggies" on diagonals.
I've downloaded the free trial, and in its first test, it seems to deliver. I probably wouldn't spring for the full version yet, needs more testing on my part.
