How should the beginner scan? (Nikon Coolscan)

[Take2]

Vuescan has a multiple exposure option. It will scan the film more than once (two or three times, not configurable as far as I can tell) with different exposure levels. It drills down into the shadows to capture extra detail, and also tries to fine the extra gradations in the highlights.

An excerpt from the Vuescan manual:

"The basic idea is scanning each pixel more than once and averaging the pixels. Each doubling of the number of pixels increases the effective number of useful bits of data by one. For instance, if you have a 10-bit scanner like the Nikon LS-30 and you read the CCD 4 times at each pixel position, you get effectively 12 bits of useful image data.

There are several ways of achieving multiple image samples. The first is single-pass multi-scanning. Some scanners are capable of reading each pixel position multiple times before advancing the scan head to a new position. The second technique is multi-pass multi-scanning, which most scanners are capable of (however, some can't accurately reposition each scan pass, so this may not always work well)."

URL: Maximizing Image Quality

 

[Take2]

artobest -

Fascinating. I'll definitely have to try that out...

When you say scan as "RBG positive" I take it you mean the slide film setting? (eg. some transparency)

Expect feedback and probably more harassing questions!

 

[Tim Gray]

Quoted directly from Hamrick's user manual:

"If you're scanning several images in one go, scan each image once and save the raw CCD data file. "

URL: Using Raw Scan Files

But then afterwards he explains that "raw" is in fact a TIFF

URL: Output tab

Seems the DNG file wrapper is altogether optional....

RAW files from most digital cameras are just grayscale images that haven't been debayered yet. That's all the image data is - a bunch of tonal values. Debayering takes that data and knowledge of the color filters over the pixels to reconstruct the color image.

In most scanners, like your Coolscan, the CCD is an RGB device. It outputs an RGB values for each pixel. It doesn't have a Bayer filter so it's not a RAW file like what comes out of a camera. It *does* have linear gamma, so needs some processing before it looks good. So in this case it's not quite like a RAW image. It's pretty much exactly like a TIFF The closest thing to it out of a digital camera is probably a Linear DNG, which is a RAW image that has been debayered (demosaiced) and saved as a DNG file.

So, there's no real point in scanning to DNG files other than the fact that Adobe Camera Raw and Lightroom will store your adjustments inside the file and not in an XMP sidecar. Though they'll do the same for a TIFF file too I think. So there's no real point in my mind.

Lastly, I don't really see the advantage of taking a non-inverted linear scan (a raw scan if you will) into one of these programs. In my experience, they really don't provide the necessary tools to color correct it properly and many times you'll end up with some color crossover in the shadows. There's no real substitute for setting the black point properly, which calls for a full fledged curves tool, and Lightroom doesn't offer that.

That being said, you can go a long way to recreating Color Perfect's inversions with Photoshop or even Image Magick if you know what you are doing. Change the gamma of the file from 1.0 to 2.2, invert, and properly set the black and white points and you are 90% of the way there. I have a set of actions that does this in Photoshop. If you want to do it for free, check out the negfix script (relies on Image Magick).

negfix8

 

[Take2]

I don't use Vuescan but I'd be surprised if there weren't a per-channel levels function.

I checked and AFAIK, There isn't. But you can adjust levels using a slider similar to the one in photoshop.

The black and white adjustment layer is for transforming colour images to b&w. It effectively replicates the use of colour filters in traditional b&w photography, with greater subtlety and variation.

I tried scanning a b/w negative as per your suggested workflow:

- mode: color transparency
- bits/c: 16 (48-bit RGB file)

I then compared it to the same frame scanned as 16-bit grayscale negative, with all other settings unchanged (black/white point, resolution, etc). Both images were brought into photoshop, levels adjusted and the RGB version was inverted.

Verdict:

The RGB variant was more contrasty with substantially less highlight/shadow separation. The grayscale variant showed slightly longer tonal range.

I tried playing around a bit with the channel mixer, and while I see your point as to how it differs from the B/W adjustment, I've no advantage to using it with grayscale images. (or at least the one I scanned/tested)

Thus, in my configuration (LS4000 for 35mm b/w negs + Vuescan Pro) I can find no factual basis for scanning a b/w negative in RGB or for that matter, in anything other then a 16-bit grayscale TIFF file. I'm actually quite enthused by the results obtained in this way, where tonal range and image quality is concerned. This, especially since the negative I used to test this workflow is rather crappy!

Images attached -

Reduced scan RGB: http://dl.dropbox.com/u/1578355/scan-compared/full-rgb.jpg

Reduced Scan GRAY: http://dl.dropbox.com/u/1578355/scan-compared/full-gray.jpg


--

1:1 CROP RGB: http://dl.dropbox.com/u/1578355/scan-compared/zoom-rgb.jpg

1:1 CROP GRAY: http://dl.dropbox.com/u/1578355/scan-compared/zoom-gray.jpg

 

[chuck94022]

An excerpt from the Vuescan manual:

"The basic idea is scanning each pixel more than once and averaging the pixels. Each doubling of the number of pixels increases the effective number of useful bits of data by one. For instance, if you have a 10-bit scanner like the Nikon LS-30 and you read the CCD 4 times at each pixel position, you get effectively 12 bits of useful image data.

There are several ways of achieving multiple image samples. The first is single-pass multi-scanning. Some scanners are capable of reading each pixel position multiple times before advancing the scan head to a new position. The second technique is multi-pass multi-scanning, which most scanners are capable of (however, some can't accurately reposition each scan pass, so this may not always work well)."

URL: Maximizing Image Quality

Stormbytes, you are confusing multisampling with multiple exposure. I have read the manual. I know how it all works I just haven't found any configuration options to control it (where "it" is multiple exposure, not multiscanning). From reading your many responses here, perhaps your fingers are getting ahead of your synapses a bit.

Have any *experienced* viewscan users fully understood multiple exposure and figured out how to control it and best exploit it post-scan?


---
I am here: http://maps.google.com/maps?ll=25.058191,121.529632

 

[Take2]

From reading your many responses here, perhaps your fingers are getting ahead of your synapses a bit.

Funny... As far as I can tell I only responded to *your* post just the once. This might be a good time to double check that prescription.

 

[Tim Gray]

I've only used multi exposure when scanning Kodachrome slides. Pretty much all the other color I scan is color negative, which a Coolscan has no problem with digging into the shadows. As far as I know, it's either on or off. You don't have any control over it. However, in my tests, it did the right thing - just dug into the shadows more.