My question is three fold: What is the best way to scan 4x5 film (B&W) on a flat bed scanner---- do it in grayscale or in RGB mode and convert it to grayscale in photoshop. Secondly, is it better to use chromogenic film instead of regular black and white film for scaning. I use a very good pro lab to develop my B&W films. I now want to switch to T-Max 100. Could I do the film rating test with the lab processing the negatives for me as I do not have the time nor the space to do my own processing and then obtain densitometry readings of the negatives to establish the EI for T-Max 100.

Responses will be appreciated.

Mo Kenny

-- mo kenny (mokenny74@hotmail.com), May 09, 2001

Answers

Hi Mo, I know nothing about digital stuff, but I know T-max since it came around. It's a great emulsion and quite sensitive and responsive to development variables. If you really trust this lab, maybe this practice of sensitometry may work. Short toe helps detecting exposure limits and testing your metering techniques. But contrast control can be maddening if things in the lab aren't kept on clever hands. I hope you're lucky and don't become one more to hate this film. Regards,

Cesar B.

-- Cesar Barreto (cesarb@infolink.com.br), May 09, 2001.

At least with my flatbed scanner, a Microtek ScanMaker 4, I get much better results by scanning in color, than adjusting the levels in Photoshop for each color (set white and black points), then converting to greyscale. If I select greyscale directly from the scanner, the results just don't seem very good, perhaps because the histograms of the three colors don't line up.

If you are going to have your actual T-Max 100 negatives developed by a lab, then it makes sense to have your test negs processed the same way.

-- Michael S. Briggs (michaelbriggs@earthlink.net), May 09, 2001.

Most scanner software seems to take only the green channel when scanning in B&W, and this is the channel that also seems to have the most noise. Stupid or what?
It's my experience that scanning in RGB and desaturating in PS gives a slight edge in quality.
As for the difference between chromogenic and 'normal' films, I don't think the difference is worth going out of your way for. XP2super scans very well, but then so does FP4+. I suppose if you need the 400 ISO rating then XP2 is better than Tri-x or HP5+.
I don't recommend Tmax if scanning is your main aim. I find it nearly always needs more fiddling with the curves than FP4+, in order to get the tonal range right.

If you don't already have a scanner, take a look at Canon's new 2400 dpi job.

-- Pete Andrews (p.l.andrews@bham.ac.uk), May 09, 2001.

I do not believe you can extract more or different information by scanning a B&W negative in Lab or RGB mode. Scanning a color neg or slide and then examing the channels (RGB or CMYK) is a different matter. The major thing to watch for: flatbed scanners do an exceptional job with thin negatives (even ones so thin we would likely not try conventional printing), while they have a much more difficult time with dense negs. The well-exposed but not over-developed neg does best here also, as will a finer-grained film.

-- David Stein (DFStein@aol.com), May 09, 2001.

Here is a tip I learned from Tom Lopez who is the Chair of the Digital Imaging Fine Art Department at the University of Miami.

Scan using the highest resolution you can. Preferably the native optical resolution of your scanner. Try not to let the scanner interpolate. Scan in RGB and push your white and black points all the way ou. Essentially you are scanning very flat. This will record maximum detail in the highlights and shadows.

Next import the image into Photoshop and convert to LAB using Image/Mode/LAB. One you have converted the image go to the channels box on the right and drag the "b" channel into the trash can. Next drag the Alpha 2 channel to the trash can. Bow change the image to Grayscale using Image/Mode/Grayscale.

This technique produces a very nice looking grayscale image that can be further modified in Photoshop as required.

Oh, BTW, it is the Blue channel that contains most of the noise witin an image, I believe someone said it was the green channel.

I hope this technique helps, I saw it demonstrated at FotoFusion in January. Tom had even the Photoshop expert Julianne Kost watching on in awe.

Mike

-- Mike Kravit (kravit@bellsouth.net), May 09, 2001.

David. Scanning B&W negatives in colour really has nothing to do with RGB or L*a*b* colour spaces. It has to do with fineness of greyscale resolution and noise reduction.
If you take only the signal from (say) the green channel, then the noise is fixed by the S/N ratio of the CCD and its A/D converter, but by taking all 3 channels and combining them, then the noise is reduced.
Since noise is random, it tends to cancel when added, but image data is real and constant over the 3 channels, and does not cancel. The result is that the S/N ratio is increased on average by root 3, and you get ~0.57 times the noise that the data from one channel alone gives.
There may also be some increase in tonal resolution, since the gain of the 3 channels won't be absolutely matched. These gains will sum to an average of the 3 channels, and the A/D dithering will likewise be averaged. If scanning is done in 48 bit mode, there should be the possibility of gaining some additional 'interpolated' bits of greyscale resolution from the sum of the 3 RGB channels. This might give a little extra 'headroom' when manipulating curves or levels, before converting to 12 or 8 bit greyscale.

Mike. I've also read that the blue channel is supposed to contain most of the noise data, but this is contrary to experience. What I see in actual scans are green speckles in the shadow areas, plainly indicating excess noise in the green channel. There's absolutely no good reason for electrical noise to favour one channel over another, since the only difference between channels is in the optical filtering of the CCD elements, so I suspect that the noisiest channel may vary with CCD type, or with the layout of the scanner circuit.

-- Pete Andrews (p.l.andrews@bham.ac.uk), May 09, 2001.

Pete,

The reason I see noise in the blue channel and not the green channel may be because I am using a drum scanner (PMT) not a CCD scanner. am not an electical engineer, so my observations are highly subjective.

I find absolutely no green "speckles" in the shadow areas.

-- Mike Kravit (kravit@bellsouth.net), May 09, 2001.

If color separation is performed with optical filters over the same CCD, then differing amounts of gain will probably have to be applied to the three channels to take into account a non-constant CCD response with wavelength. Perhaps the green (or blue) channel needs to be boosted more, and thus noise is amplified.

Good grief, if we're going to talk SNR here, then I'm going to quit spending my work breaks reading this site! ;)

-- John H. Henderson (jhende03@harris.com), May 10, 2001.

The best tactic will vary with the scanner hardware, firmware and software, so general prescriptions aren't much help except to highlight things to look out for.

When I scan B+W on my flatbed (Epson 1200) it is definitely better to use the colour negative mode and not the B+W, but I suspect this is entirely a software issue. The red channel is always low contrast and noise-free, while the green and blue are progressively more contrasty and noisy. Both of these I put down to greater gain (digital, not analogue) for those two channels (which makes sense when you think about orange mask removal). Shadow noise can also be amplified via 'noise pumping' when the image is inverted, especially if it's done with 8-bit maths. The excess noise negates any benefit from combining channels once the image is in photoshop.

With the Epson software the 'best' way is to scan as a slide, and do the inversion manually. That way all the channels are treated equally. It can be difficult to set black and white points though, so better driver software (Vuescan for me) makes a lot of sense if you are going to be scanning B+W regularly.

-- Struan Gray (struan.gray@sljus.lu.se), May 11, 2001.