Color or grayscale for scanning B&W film?

[Particular]

But the Minolta 5400, for example, has the 3 channels pretty equal, resolution-wise. Fernando

Perhaps the Minolta engineers succeeded in putting an apo-chromatic lens in their scanner, while the nikon guys didn't quite nail it. It surely was an issue they grappled with.

 

[tsiklonaut]

Color and also scan as a linear negative in 16 bit for the best advantages in PP.

Scanners scan it as positive anyway, natively, making it "positive" during the scan from a negative always involves losses since the scanning software has to do it on-the-fly and not to give you a faulty or defected scans on slower-CPUd computers due to data inflow the on-the-fly conversion can use a compromised "cut-corners" algorithms.

Unless you use some high-end scanning software then scanning as B&W especially as positive can lose you a lot of juice from the original media since the scanning software tends not to be as "lossless" as pro editing programs such as Photoshop etc in conversion.

For me it's about building the strongest "chain" and removing any possible weak links - scanning with the best native setting on the scanner to get the best out of hardware - full color RGB with the highest bit-depth the hardware allows and perform the inversion it in the best possible software environment: any pro photo editing app working natively in 16 bit or 32 bit environment. There's another advantage doing so: by keeping a copy of the analog "digital negative" (linear 16bit RGB TIF scan) will keep you in time as the image manipulation softwares evolve - you can use the same files 10 years later in some future PP app much more capable and take comfort knowing you've gotten the best out of your scanner hardware with no software compromises or cut corners.

Conversion procedure: with smart usage of curves for overall tonality I don't take any certain channel preference, but use a b&w conversion through a channel mixer where I observe the overall sharpness, grain and tonality depending on the particular film - usually it tends to be the spectral peak of the film depending it's native color (backing) after inversion, it's also depending on the developer you used the Ilford films tend to be warm (yellow-orange so red-channel biased mixture gives the best conversion results), some Efke films are almost blue in some developers (then blue-channel dominated mixture give the best B&W conversion for sharpness) etc. So I try to find optimum from particular film's media. This method I think is the most "scientific" one yet gives you massive creative control over the B&W workflow, i.e. with all the original info available by a selective channel mixer you can tune it to the maximum visual sharpness and detail or to reduce grain, or find other unique "awkward" tonality or increase softness you can bias on the opposite channels or anything between etc, something you have no control over with a simplified "direct" or single channel B&W conversion most people use, especially doing it through a scanning software. Also you have loads of editing headroom thanks to the linear "flat" scan.

After B&W conversion I also keep it in RGB space in PP till the end, I almost never work in 100% grayscale workflow. This is a personal preference, definitely not mandatory.

 

[thereabouts]

I scan in colour for the following reasons:

1. It gives more options for post-processing in Photoshop – some features are unavailable in grayscale.
2. I like to make the black and white conversion myself and this gives me more control.
3. I don't care about file size.
4. When I send images to be printed to a high-end injket, they want files in RGB, as they use multiple shades of black ink for best results.
5. I'd rather my cheap desktop scanner and scanning software captured as much data as possible and allowed me to throw away any 'uneccesary' digital information in Photoshop.

16 bit grey scale at the highest optical resolution on my Coolscan V. When I tested that against the colour and select a channel approach I couldn't see an advantage in either and so chose the one that gave me the smallest files.

I once asked Ed Hamrick of Vuescan fame if he could have a mode in the software that scanned negatives as positives and then inverted the scan. His response was that was what the scanner did anyway so there was no point.

Interesting. I've never understood the whole 'scan as positive and invert' argument. We are dealing with digital here – ones and zeros. All we would be doing is inverting data, surely (more or less)?

Even assuming it makes any difference (which I don't believe), I'm guessing that most of us are using scanners costing no more than a few hundred (perhaps a couple of thousand) dollars? That being the case, I wonder how much of this makes any difference? For absolute quality, a drum scan by a skilled operator is required anyway. I used to work in design and pre-press and, compared to drum scans, the home scanned images I work with are not even worth trying to compare.

And then we are dealing with old cameras and lenses, curved film, dust scratches, the various foibles of analogue development and so on.

I'm not saying that trying to get high quality isn't worth the effort, but is this kind of trickery really going to make much difference? I doubt it. But digital 'perfection' is not why I use film. If I want that, I use digital. 😉

 

[xavoy]

I haven't read this whole thread, but have skimmed over it, and as a result made some adjustments today whilst scanning some difficult (high contrast) images. Thought I'd chime in with my experience.

I have always scanned and saved in grayscale, using a a plustek 8100 and Vuescan. After reading this thread, I played around and found that I could indeed get quite a bit of extra detail out of both highlights and shadows when I set the INPUT as 24 bit RGB (rather than 16 bit gray). I tried using 48 bit RGB as the input but it offered no noticeable benefit (to my eyes). And while there was very slightly more information in the 48 bit RGB TIFF files, it wasn't all that much, and so I will continue to output as a 16 bit gray TIFF.

This is scanning Tri-X (developed at home in Rodinal, if that makes a difference)

 

[Chriscrawfordphoto]

I've never understood the whole 'scan as positive and invert' argument. We are dealing with digital here – ones and zeros. All we would be doing is inverting data, surely (more or less)?

Nope, you're completely wrong here. Film scanners, and I mean real film scanners, not flatbeds that can also scan film, are made to be able to scan the very wide density range of a color transparency.

A black and white neg should never have anything as dense as the blacks in a transparency. So, when the scanner scans the neg, which has a smaller density range, it gives a very flat scan that looks muddy because there are no whites or blacks in the image.

Most scan software, when set to scan BW negs, applies a levels correction behind the scenes that is separate from any levels or curve corrections the user sets in the software. This makes the scan look more like a normal BW photo, but with very high contrast negs it will clip some tones.

Scanning as a transparency gives a scan without any background corrections, thus capturing all of the data in the film in cases where the film has a wider density range than a normal BW neg should have.

 

[thereabouts]

Nope, you're completely wrong here.

I'm willing to accept that I might be wrong in various ways. But I'm not sure that what you have said particularly addreses where I might be incorrect.

You appear to be saying that we ought to scan without using the scanner's software 'corrections', such as level adjustments and so on. And that by tricking the scanner into thinking it is scanning a transparency, it will cover a wider density range. This is, of course, perfectly possible (and advisable) without setting the scanner to scan as a positive transparency. Simply disable all the software filters and most scanners will scan the image equally, regardless as to the media being used. And, if scanning black and white negatives as 48bit colour, I would say that all bases are being covered.

Moreover, as has been stated by a few people upthread and elsewhere, scanners ('real' or otherwise) seem to work the same way, in that they scan as positives and then 'invert' as neccesary, automatically. That being the case, I'm still yet to see any evidence that scanning negatives as postiives gives any advantage. On the contrary, it seems to be double handling.

 

[Chriscrawfordphoto]

I'm willing to accept that I might be wrong in various ways. But I'm not sure that what you have said particularly addreses where I might be incorrect.

You appear to be saying that we ought to scan without using the scanner's software 'corrections', such as level adjustments and so on. And that by tricking the scanner into thinking it is scanning a transparency, it will cover a wider density range. This is, of course, perfectly possible (and advisable) without setting the scanner to scan as a positive transparency. Simply disable all the software filters and most scanners will scan the image equally, regardless as to the media being used. And, if scanning black and white negatives as 48bit colour, I would say that all bases are being covered.

Moreover, as has been stated by a few people upthread and elsewhere, scanners ('real' or otherwise) seem to work the same way, in that they scan as positives and then 'invert' as neccesary, automatically. That being the case, I'm still yet to see any evidence that scanning negatives as postiives gives any advantage. On the contrary, it seems to be double handling.

Not all scanners can be set to disable all software corrections. In that case, scanning as a transparency gives nearly the same result they would get if the software would allow scanning as a neg without further internal corrections.

 

[Damaso]

I've always scanned in grayscale at 14 bit on my Coolscan V and made exhibition size and quality prints...

 

[borge]

I scan at 16-bit grayscale directly.