Screen 1030AI and color negatives

[joseppe]

Hello,

after searching the last two weeks about the Screen 1030AI`s ability scanning color negatives (in my case MF and 4x5), i`m now just more confused.
The general tenor is, that drum scanning color negatives is basically more difficult, then transparencies.
Particular for the screen 1030AI, the opinions are divided, in first place because of the 8bit, the scanner provides. Some say, that the point light, that is send through the medium, is full of intensity, that 8bit from a drum scanner can`t be compared with 8bit from a flatbed. Latter can`t recognize shades in that range, because it`s light source isn`t just as bright.

But apart of the technical aspects, how does it look on software site? Is it really so difficult to get satisfying results? I have often read, that the black and white point is very necessary. But what should i do, if I have photographed a completely bright scene, without any dark areas, respectively vice versa.

So dear screen 1030ai users, whats your opinion with scanning color negatives? Has anyone some scans to show up?

regards, Joseph

[vinny]

I don't have the negacolor software on my 1030ai system so i can't do color negs. There are folks who scan negs as positives then Invert them using software.

[Rich14]

Joseph,

Obtaining good results from scanning color negatives has nothing to do with drum scanner vs flat bed scanner technology or the brightness of the lamp.

Theoretically, it is possible to invert the color image from the 1030AI in software after the scan to achieve a reasonable representation of the original scene. The inversion process has to compensate for the orange mask that is typically inherent to color negatives and other characteristics of the color curves of the particular brand of film.

However, there are (at least) two huge obstacles that will defeat you. The first, which is minor compared to the second, is the problem of non linearities (cross overs) of the characteristic color curves of the emulsion. These can make it impossible to fully correct the image without elaborate (hand tweaked) masking. Even then, the results can look garish. A scanner that has only positive scanning capabilities cannot set its sensitivities properly (white and black points, contrast/gamma, etc), resulting in a capture that is so "stretched" or "limited" in its histograms as to make the post processing software correction futile. A scanner that can invert and apply its curves and set points to the needs of the negative at the time of the scan can deliver a file that needs only small corrections in post.

The second problem is the issue of the 8 bit file from the 1030. And this applies to any scanner producing an 8 bit file. At 8 bits, the scanner must produce an almost perfectly color corrected scan (positive or negative) right out of the scanner. There is almost no "room" for color manipulation after the scan without degrading the quality of the image. At 8 bits, the image contains 256 "gray" levels in each color channel (2^8). It is necessary to have at least 256 "bit depth" or gray levels for the image to appear high quality to our eyes (on screen or print). Any manipulation of gray levels (color correcting a color image or changing the contrast of a B&W image) results in tone expansion in some parts of the curve and tone compression in others. And that results in loss of gray levels because "synthetic" levels have to be generated from adjacent pixels to "fill in" the compressed or expanded areas. And that results in posterization and pixelation.

If the image contains more than 8 bits of information, "real data" is available in the compression/expansion process so no synthetic data has to be generated. A file holding 12 bits of image information has 4096 gray levels per color channel (2^12), 16 times the information needed to produce a final 8 bit file for final output. A file holding 14 bits of image information has 16,384 gray levels. When color moves are made in such high bit files (Curves or Levels adjustments in Photoshop or contrast changes), significant changes can be made without fear of image quality loss. Of course, it is always best if the image is as close to its final tone values out of the scanner before any corrections are done.

Scanners and digital cameras producing high bit files, of course, output 16 bit files but there is no scanner that outputs more than 12 image bits and no camera that outputs more than 14 image bits in those 16 bit formats. And to be pedantic, there aren't actually 12 real, useable bits in scanner files or 14 real useable bits from digital cameras as there are losses in the analog to digital converters resulting in a little less real image info. But there is plenty of "headroom" in such files to allow post processing tweaking.

Rich

[coisasdavida]

A while back there was a thread about this subject (color negs on the 1030) inside the 1045 Yahoo Group.

[Bruce Watson]

The second problem is the issue of the 8 bit file from the 1030. And this applies to any scanner producing an 8 bit file. At 8 bits, the scanner must produce an almost perfectly color corrected scan (positive or negative) right out of the scanner. There is almost no "room" for color manipulation after the scan without degrading the quality of the image.

This.

[David Rowland]

I have a 1030 and have scanned quite a lot of colour negs on it in various formats. I scan as a positive and convert the image in photoshop using a drop on layer set a retoucher and former scanner operator gave me so long as you crop off the border of the neg it is quite easy with practice.