kennylovrin
Well-known
Hey
I don't know if this is of value for anyone, but it was for me so I might as well share it..
I've been trying to simplify my scanning workflow for some time (for me film is purely for the fun of it, so the more I can automate the better). However, I can barely stand the scanning software, and want to go as quick as possible into Aperture with as much of the basic work already done.
So I decided on linear scanning as it is the fastest for me because I just have to hit scan basically, and do everything else in another software.
It seems many people use negfix8 to convert linear scans to positive gamma adjusted ones, but personally I always get horrible results from it, I just can't get it to give me anything that makes sense.
So I played around with the same workflow, but going straight at Image Magick.
So here's a few quick comparisons (they all are from the same 16bit linear scan, scanned with Silverfast 8):
1. Converted with negfix with contrast stretching enabled (-cs flag)
2. Converted using my own Image Magick command, with normalization/contrast stretching (convert -gamma 2.15 -negate -channel ALL -normalize IN_FILENAME OUT_FILENAME)
3. Converted using my own Image Magick command, without normalization (convert -gamma 2.15 -negate IN_FILENAME OUT_FILENAME)
4. Same as no 3, but with auto curves and auto levels applied in Aperture (so basically Aperture is doing the contrast stretching)
In my opinion, the first one is horrendous and unusable for any post processing without severe quality loss. The second one is much better, but by allowing Image Magick to do the contrast stretching, you'll basically always end up with a frame that spans black to white, and for some shots there is no black or white in the frame, and it becomes trickier to pull it back in post.
So I think because Aperture auto adjustments gets so close to the final result, the combination of 3 and 4 works best for me, as it will retain all the full tone "as shot" but won't pull grays to either black or whites unless you want it on purpose.
I haven't tried this on color shots yet, theoretically the same commands probably won't work because of the orange mask. On the other hand Aperture might be able to offset the mask with it's auto levels, so I'll look into that when I have the need.
Anyway, not very scientific, but perhaps of value to someone!
Kenny
I don't know if this is of value for anyone, but it was for me so I might as well share it..
I've been trying to simplify my scanning workflow for some time (for me film is purely for the fun of it, so the more I can automate the better). However, I can barely stand the scanning software, and want to go as quick as possible into Aperture with as much of the basic work already done.
So I decided on linear scanning as it is the fastest for me because I just have to hit scan basically, and do everything else in another software.
It seems many people use negfix8 to convert linear scans to positive gamma adjusted ones, but personally I always get horrible results from it, I just can't get it to give me anything that makes sense.
So I played around with the same workflow, but going straight at Image Magick.
So here's a few quick comparisons (they all are from the same 16bit linear scan, scanned with Silverfast 8):
1. Converted with negfix with contrast stretching enabled (-cs flag)

2. Converted using my own Image Magick command, with normalization/contrast stretching (convert -gamma 2.15 -negate -channel ALL -normalize IN_FILENAME OUT_FILENAME)

3. Converted using my own Image Magick command, without normalization (convert -gamma 2.15 -negate IN_FILENAME OUT_FILENAME)

4. Same as no 3, but with auto curves and auto levels applied in Aperture (so basically Aperture is doing the contrast stretching)

In my opinion, the first one is horrendous and unusable for any post processing without severe quality loss. The second one is much better, but by allowing Image Magick to do the contrast stretching, you'll basically always end up with a frame that spans black to white, and for some shots there is no black or white in the frame, and it becomes trickier to pull it back in post.
So I think because Aperture auto adjustments gets so close to the final result, the combination of 3 and 4 works best for me, as it will retain all the full tone "as shot" but won't pull grays to either black or whites unless you want it on purpose.
I haven't tried this on color shots yet, theoretically the same commands probably won't work because of the orange mask. On the other hand Aperture might be able to offset the mask with it's auto levels, so I'll look into that when I have the need.
Anyway, not very scientific, but perhaps of value to someone!
Kenny
DrTebi
Slide Lover
It is definitely worth trying to experiment as you did.
One thing that I find a must is to start with a linear scan. After that... there are many ways to get to a satisfactory result.
I have used ColorPerfect quite a bit, although lately I am finding that it's really best for color negatives, for black and whites I have had great results by "manually" converting the linear scan in Photivo. There are so many settings to play with, and with time it's in my opinion really worth it to use Photivo for just about any post-editing of scans. Give it a try (it's free... so there is nothing to loose.
One thing that I find a must is to start with a linear scan. After that... there are many ways to get to a satisfactory result.
I have used ColorPerfect quite a bit, although lately I am finding that it's really best for color negatives, for black and whites I have had great results by "manually" converting the linear scan in Photivo. There are so many settings to play with, and with time it's in my opinion really worth it to use Photivo for just about any post-editing of scans. Give it a try (it's free... so there is nothing to loose.