Saturday, February 18, 2012

Digital Darkroom: Scanning Black & White Film In Color


This post is about being reminded that while trial and error can be lead one down creative roads that would otherwise remain unexplored, it can also send you down the wrong path without even being aware that there might be a better way.

It's a basic fact of modern photography that almost all film shooters are also digital, even if they don't own a single digital camera. Unless all you want to do is make wet prints in the darkroom to pass around by hand, you are going to need to find a way to digitize your analog images if you want them to get seen. Most people don't even deal with paper prints anymore - they shoot on digital, post images to Facebook, Flickr and etc. or maybe email digital photo files to friends, family and like, all of which get viewed on some type of monitor.

"New Year Light" - January 1, 2012, Chicago, Illinois
(Canonet QL-17, Fomapan 400 film, home developed and scanned)
The final, post-processed "color scan" of the black & white negative. The negative was scanned as a 24 bit color image (RGB) and outputted to a 24 bit RGB TIF file. There is much more detail in the shadow areas and much smoother transitions between tones than any of the Grayscale scans of the same negative, even when saved at the highest Grayscale bit setting.
If you are shooting film, getting that image converted to a digital file means getting a scan somewhere along the line. You can scan prints, of course. And scanning a print is relatively straightforward. Even low-end scanners will generally give you a decent scan of a color or black & white print. But a scan of a print is, at best, a third-hand representation. The negative is used to make the print, which is used to make the scan - each step in the chain represents a potential loss of detail and degradation of the image.

If you are someone who makes actual money of their work  (or a lottery winner), then getting a  drum scan of your negative might be your preferred choice, as a drum scan will give you the best detail and a huge file that can be used to print at very large sizes. But drum scans are incredibly expensive, and most people aren't looking to blow their images up to wall-size prints.

Most places that develop film also offer "standard" resolution scans, such as you get on a Kodak Picture CD. While not generally that expensive - generally an additional $3-$5 - it can be somewhat of a crapshoot when it comes to quality and consistency. The reason for this is that scanning negatives, especially but not only color negatives, is not a "push button" activity. Whoever scans the film, you or the scanning technician, has to make choices about color, contrast, dust, resolution and a host of other variables. Additionally, film negatives are the result of a set of chemical reactions and are, therefore, inherently variable. The exact characteristics of a negative of a particular type of film will vary slightly depending on the chemicals it was developed in, development time, temperature, agitation schemes, the age of the negative, and more.

With all of this to worry about, is it any wonder that film scanning is referred to as a "dark art"? Even though most serious film shooters ultimately end up buying a scanner with the goal of having more control and getting better results, if you ask 10 film shooters about "the best" way to scan any particular negative, you are likely to get 20 answers.

I've used several different film scanners, beginning with a Minolta DiMage Scan Dual III through a Nikon Coolscan 4000 ED. Now that I am shooting both 120 film as well as 35mm, I use an Epson V600. I've generally gotten decent results with my scanners, but I have never been very systematic about really learning the ins and outs of getting good scans. Over the past year, I have been working on my scanning techniques and am beginning to get to the point where I feel like I am maximizing the equipment that I'm using and being able to produce consistent, high-quality scans. No, the scans are not "drum-scan quality," but they are good, consistent scans, and are done well enough that I am not creating unnecessary extra work for myself to do in Photoshop.

So I was a bit surprise to find out that I have been making a big mistake in my technique for scanning black and white negatives - specifically, scanning my black & white negatives in a black & white colorspace ("Grayscale"), rather than a color (RGB) colorspace. 

It really didn't occur to me that I should be scanning my black & white negatives in color. Scanning black & white film in Grayscale makes intuitive sense, and (again intuitively) a 16-bit Grayscale scan would seem to hold the potential for containing at least as much detail information as a 24-bit color scan, since the Grayscale would not need to hold color information. However, as is often the case, intuition doesn't exactly square with reality and it turns out that you can get a lot more detail and better tones in black & white images by scanning them in color.

When I first came across this idea, I really didn't give it much credence. I had been scanning black and white negatives in Grayscale for years with good results, and it seemed like the sort of overly technical advice that would be more hassle than it was worth. However, I kept coming across the same advice, and so decided to put it to the test.

Each of the images below is a straight, uncorrected scan of a single 35mm frame of Fomapan 400. I chose an image that not only had a lot of detail, but was also high in contrast. This one, with the left side being bright from the window and the far right in deep shadow, seemed perfect.

Other than re-sizing the images for the web, no post processing was applied. I did not even spot the scans below for residual dust - as the incredibly annoying bit of fluff in the center of the upper frame will attest. The scans were done with Vuescan 9.0.75 on my Epson V600, using the betterscanning.com ANR inserts for the Epson's 35mm film trays, at 2400 dpi. The scans also utilized Vuescan's "Multi-Exposure" pass feature, which is great for pulling out additional detail from a negative. Finally, "Infrared Clean" (aka digital ICE) was set to Low, which is consistent with my normal workflow.

After scanning the images were re-sized with the Lightroom 3 export feature to JPEG files, using Lightroom's "standard" sharpening for web (also consistent with the standard workflow I use for getting images ready for posting to the blog. The differences between the scans are not quite as apparent as in the full-size TIF files, but the main difference between the 24 bit color scan and any of the Grayscale scans remains quite stark.

16 bit Grayscale scan, output to 8 bit TIF file; total file size 7.6 MB
16 bit Grayscale scan, output to 16 bit TIF file; total file size 15.3 MB
24 bit RGB scan, output to 16 bit Grayscale file; total file size 15.3 MB
24 bit RGB scan, output to 24 bit RGB file; total file size 22.9 MB
Note that, even though this is scanned in a color format, the image is still black and white. (Although I do seem to detect a slight toning of the image, a very slight yellow/green tint.)
24 bit RGB scan, output to 24 bit RGB file, converted to 16 bit Grayscale in Photoshop;
total file size 15.4 MB
This is from the same scan as the image above. After scanning, the image was brought into Photoshop and the color information was simply removed. The file size is now basically the same as 16-bit Grayscale images above, but the additional detail that had been captured in the RGB scan remained.
It should be clear, even in these reduced-size jpegs, that there is a lot more detail in the last two images, than in the first three. This is particularly apparent when you compare the detail and highlights present in the lower right of the image. The Christmas ball in the Grayscale images is almost totally in shadow, with only brightest of the highlight detail coming clearly through. In the RGB images, however, the shape of the ball is much more apparent and the highlight detail is more prevalent and shows a broader tonal range.

In all of the first three images, the shadows are bunched up and muddy compared to the bottom two images. Now, if I hadn't seen the 24 bit RGB scans, I would have been pretty happy with any of the Grayscale images. My eye can detect very little difference between the first three images, even looking at the original TIFs at 100 percent on my desktop monitor. Indeed, one of the reasons for my skepticism of the idea of scanning black & white negatives in color, was that several years ago I had tested scanning at higher and lower Grayscale bit rates, had been able to see almost no difference and so had been saving all my black & white negative scans as 8 bit Grayscale files since the larger file size didn't seem justified by any appreciable increase in detail. 

However, the RGB scans are clearly superior, and when you look at the Grayscale scans side-by-side with the RGB scans, the Grayscale begins to look like junk. It may be that you want a more contrasted image or less detail in the shadows because of a creative choice you make in post, but I would rather start with as much detail and as good of a tonal range as possible and then make my creative decisions from there. 

While I won't be going back and re-scanning all my black & white negatives, there a few that I would like to see how much of a detail difference this technique actually brings out. And from here on out, I will definitely be scanning all my black & white negatives in an RGB color space. The slightly larger file sizes is more than made up for by the enhanced detail.

2 comments:

Ruby said...
This comment has been removed by the author.
Ruby said...

Digital is the paradigm now: from technology to medium or platform. Which is why it is a prerequisite to learn and look into things from that vantage, not only in terms of aesthetics but data sizes. They may not consume as much physical area, but they have to be made to fit in storage drives, or be accommodated by them. Which is fine, because there's lots of mechanisms on that end.

Ruby Badcoe