[matthew]

I think I may have promised that I wouldn't do the film-versus-digital thing - so that isn't quite what this is. Yes, one image is scanned from film, and the other two are from a digital camera, I'm not trying to compare pixels. Instead, I'd really like to know which of these images feels better. (Please ignore the slight differences in the cropping, and I realize that this isn't a scene that many people are going to love.)


Image #1:



Image #2:



Image #3:



These photos span about forty years of image-capture technology, but are all processed digitally. The film image has been put through a 4000dpi scan and adjusted in Lightroom. The other two are digital captures, one if which was from the best in-camera exposure (and then tweaked in Lightroom) and is typical of how digital capture stands in for film. The other digital shot was heavily processed with additional software, which also needed some fundamental differences in how the photo was shot. I'll say a bit more about that later, but for now I'm just wanting to know which image feels the best.

[EnglishBob]

For me #2 has the truest feel for texture and color.

[enviroguy]

I believe that #2 displays the texture just a hair better than #3, but they are really very close. The colour is not important since that can be manipulated in any direction.

Regards.....Dennis

[Don Schaeffer]

either 2 or 3 is significantly better than 1

[Wedding Shooter]

2 and 3 are the best - without seeing the original scene it is hard to judge which has the truest colour. The colour/brightness is the only real difference at web size between 2 and 3.

[shuttertalk]

Interesting thread! You can see the grain/lines on the white much better in #2 and #3, but because you asked which ones "feels" better, I would say #1 because it looks cleaner...

[matthew]

Thanks for the responses, I have to admit that I'm surprised by how well #2 did. I realize that it's nearly identical to #3, but 2 is the straight-from-camera file - or at least as close as I can some when I'm shooting raw. My personal preference is for #1 or #3, with the first being an emotional favourite as it's the film image. It just feels a little more natural. But ultimately my true inclination is for the extra detail in the third image.

(I'm someone who always looks for complicated answers to simple problems, to the point of making the problems themselves more complicated than they need to be.)

The third image has been put together using some software that I'm just starting to experiment with. It's called photoacute, and it's designed to work on multiple images of the same scene to increase the amount of detail and reduce noise, and can also be used for HDR images and distortion correction. So while the second image in my initial post is a single frame, #3 is the result of 10 frames shot sequentially. I set my camera to take a five-frame exposure bracket at 0.3EV change, which is pretty mild for HDR, but does give me extra headroom in the final result. (For ten frames, I shoot it twice.) The noise reduction can be quite impressive, and it only removes the random variation so fine detail isn't reduced. The 'super resolution' option is also very interesting, and it does work - but I'm going to have to compare some prints to see how much real difference it makes for me. I haven't been as impressed by the distortion correction, but that's actually the only thing that was really challenging about this scene. You can see some change in the shapes of the vertical strokes in the "n" and "u" in these samples, and in the original files I can still spot some of the complex distortion that this lens has at wide angles.

I'm going to work with the Photoacute software some more, especially with prints, to see how it really stacks up. The original files from my camera are 10 megapixels, and the resulting images are about 35MP after the software has cropped the merged images. It certainly isn't giving a 400% increase in detail, but there is maybe 50-80% improvement depending on the original files. I like that it uses DNG files for both the source images and the final output, making it easy to integrate into my existing workflow.

For comparison, the scanned film has dimensions in the 20 megapixel range, but doesn't seem to hold any more detail than a native file from my 10mp camera. Better film might make more of a difference, so that's something else that I'll keep experimenting with.

[Don Schaeffer]

Why should the digital image be more detailed than thwe film one? IO thought film would have the higer resolution.