Here is a question I received today:

We’re planning to buy the 35mm film negatives from our wedding photographer, and we would like to have them scanned into digital format (as a back-up, and so friends/family can order prints). A friend of a friend offered to help us out—and “scan them at 16-bit resolution” for free.
Does that refer to bit depth (i.e. 65,536 colors) – or something else?
Would 11″x14″ prints from a “16-bit resolution” scan look ok?

My response

He is referring to the bit depth per channel, so that’s 65,536 levels/channel (red green blue) which works out to many times 16 million colors (2^(16*3)). (16 million colors = 24 bit = 8 bit per channel). Technically he means “dynamic range” and not “resolution,” but many people make that mistake.

(BTW, that is well past color range of the eye which can only distinguish around 10 million colors or so. It is useful to scan at this level because you can then adjust the color or exposure without “posterization”. In other words, an 8-bit scan is good enough for most applications, but you want 16-bit for editing/retouching or when scanning a film negative.)

To determine if an 11×14″ print is okay, you need the know the right “resolution” of the scan, not just the dynamic range. To get that, realize multiply the dpi of the scan by 35mm / 14 inches (0.0984251969) and make sure that number is greater than or equal to 300: it should be greater than or equal to 3048dpi. This will ensure that you can view this without any pixelization at a viewing distance of 1 foot or greater. It also happens to be the maximum effective resolution of the best commercial photo printers.

(Note that scanners have two dpi numbers because resolution in one direction is mechanically limited: You just care about the lower number only.)

Personally, for a print that large, you can be a little generous and say the magic number can be as low as 200 since viewing distance for a print that large is well over a foot and that resolution can resolve negative film grain easily. Thus I feel that anything above 2000dpi is going to be acceptable, while 4800dpi is going to be stellar and going to resolve even the grains of the best slide film.

If the resolution meets your minimum bar, what is going to matter most is the type/quality of the scanner itself and the software. Slide scanners are the best, but recently some flatbed scanners with removable backs and slide mounts have gotten very good—Canon, for instance, has stopped manufacturing slide scanners because they feel the 9950F¹ is just as good.

After that, the software is going to make a huge difference, especially since you are scanning the negatives. Negatives don’t have the actual color of the shot, so software has to process each image—like a photo lab machine does—to restore the colors that you want. Also the software can eliminate dust and scratches that inevitably get onto the negative, especially if they aren’t well cared for.

A followup

Caitlin points out that my post was all very confusing, so I wrote a followup:

My advice is to scan the 35mm negatives in at 16-bit and at least 2000dpi, preferably over 4000dpi. The rest of it depends on the how good the particular scanner model is at slide scanning, and how good the software is at processing a negative. After that, if you wish, you can save space by storing them as 8-bit, 100% for the 17″ prints and 8-bit 50% for stuff that doesn’t need to be above a 10″ print.

The rest of it is because I don’t like to give an opinion unless I show the reasoning as to how I formed it. 😉

What do you think?

What do you think. I’ve been digital so long, I wasn’t too sure what to say. Was the advice good? Any corrections or comments?

Thanks!