Digital BW, The Print

On 03/06/2010 Michael wrote:
> I recently read a scanner review that said to scan 35mm b&w negs at 
> no more than 3200 dpi since anything more is just enhancing the 
> existing grain in the negative. 

This is fairly nonsensical if you consider that the grain is the image : 
the information contained by grain will be degraded some. I've done 
comparisons between scans at 2,700, 4000, 8000 and 12000ppi. It is true 
that scanning beyond 4,000ppi is very much a matter of diminishing 
returns, and that 4,000ppi gets almost all the useful image info assuming 
scanner optics that are up to it. But more ppi does help avoid grain 
aliasing artifacts, which are a recurrent (and insoluble) problem at 
2,700ppi and an occasional one at 4,000ppi with some materials.

> I'm not an expert on this and was 
> looking for advice for my own scanning efforts; I believed his 
> analysis. Also, you should make a few test prints from scans of 
> various sizes to see what the lowest dpi scan setting is that gives 
> you the quality you're looking for at the finished size; in this case 
> you said 8x10". You scan for the largest size you anticipate 
> printing. 

Personally I scan at max optical res possible, just because I don't want 
to have to do it again. Too much data is never a problem, it's just disk 
storage space. Too little is a problem.

> A flatbed scanner with film adapter that can scan 24 frames 
> at a time will save an enormous amount of time.

Not as much as you'd hope. For one thing you can only multiple frames like 
that at some median exposure, which is only satisfactory if all your film 
has similar density range. But the main issue is that scanning is only the 
start. You will normally have to spend considerable time in 
post-production on each image if you want any sort of decent quality. 
Defects have to be repaired, black and white points (levels) have to be 
set, and B&W neg film in particular needs lots of curve work because its 
density curve is S-shaped, to match the inverse characteristics of bromide 
paper. Exposure and development determine which parts of the curve are 
used, and the OD of the image. Scanners transfer with a simple gamma 
function. Usually B&W scan tonality is horrible unless curves are worked upon.

-- 
Regards

Tony Sleep
http://tonysleep.co.uk