What can be done about sloppy cut negatives?

Question

I recently picked up my ten rolls of film from the local lab, which previously has done a fine enough job developing my film, but this this time whoever cut the negatives really did a poor job -- many images are spliced by a diagonal hair with a small portion of the image falling on the wrong side of the cut line. Some are worse than others, but there really shouldn't be any. This was all ten rolls of film had this problem.

Now my "high resolution" scans appear to be made prior to the sloppy cut job (or the hardware/software cropped the image to the smallest fitting rectangle?) but that is a poor compromise. The only reason I ordered the scans was so I could identify the ones I wanted printed -- printed from negative, not from "high resolution" scan from my local lab.

So my question is, simply, how do I recover from this? Is there anything I can do to print the damaged images from negative without empty space occurring where the sloppy cuts were made? For damaged images I believe my only recourse is to take both strips of negative to a professional and have them scan both ends of the image and reconstruct it in software at a sufficient resolution to print. But maybe one of you have seen this before and know an old trick to get past this?

Questions 2, who do you go to in NYC who does NOT do this, does not damage the film during development and does not damage the film after development?

Answer 1

How high is the resolution?

The nature of film is that it will record about 2500 x 4000 pixels.

Try this experiment: Take a single good film frame with somewhat extreme exposure range, but one that you wouldn't have to dodge or burn to get an acceptable print.

1. Produce a print from the negative.

2. Scan the negative at successive resolutions, starting at a level well below the known resolution of film, and continuing up to, say 4000 x 6000 pixels. It is critical that the neg not move between scans.

3. From each scan make a print, blowing up a crop of the image.

4. Examine them: Can you see a difference.

Take the scans. Expand and interpolate each to the size of the largest scan. Now use photoshop, or netpbm and subtract one image from the other, and then renormalize the contrast. You may have to do this both ways A-B and B-A.

This will give you a measure of the detail difference between images.

My expectation is that if the high res scan is over 2500 x 4000 x 12 bit depth that the differences between images are going to look like sandpaper. If there were real differences there would be some degree of ghost image.

Now: from your scans, can you print an image that is comparable to the one printed from the negative. This will require calibrating a printer, and really needs a photograde printer to do so, OR access to an online system for printing.

If you can pass these tests, then the high resolution scan should be sufficient. Proceed from there.

Answer 2

Brew lotsa coffee. You're going to need it.

The short answer is that the results will be a compromise between what is possible and what you expect. Here're the details…

NO, not if you ever expect anything that will become un-noticeable at ANY level of scrutiny. When I tried this after ONE similar tragedy, I gave up after many hours of multi-level stitching, and sampling in PhotoShop and accepted what I had then. I may be picky; but, every time I see it, it screams at me.

YES, with today's "healing" effects, you can achieve an acceptable nearly automated remedy for your mishap in software. Oh, …and one more thing, Good Luck.