Scanning bw film - greyscale or RGB?

[thomas_park]

I usually scan my B+W film in RGB mode, but suddenly started to wonder if this

is the best course of action.

 

Is there an advantage to scanning in greyscale over RGB? The scanner is using

the same sensor, so I'd assumed that the greyscale scan would simply use the

tonal values for the RGB channels and represent them as separate grey values. I

don't see any reason why scanning in greyscale would cause the sensor to be more

sensitive to grey tones.

 

In other words, my assumption is that scanning in greyscale is the same as

scanning in RGB but with perhaps without the ability to apply different output

levels to the separate color channels. Is this a reasonable or correct

assumption, or is there actually an advantage to scanning in greyscale for BW

images?

 

Thanks,

Thomas park

[bernardwest]

<i>without the ability to apply different output levels to the separate color channels</i><br><br>

 

If there's no colour to begin with, then there is no use in seperating the colour channels. Also, greyscale should give you a smaller file size (therefore, faster!).

[thomas_park]

I'd argue that the film base does contribute some color, at the very least, and that some scanners may have slightly more noise from one color channel than the others, so it may be desirable to clip the output from the channels unevenly. Probably not an issue with the newer and higher end machines, but this is an ingrained habit by now =)

[obakesan]

Hi

<P>

I believe that it depends on the scanner. My Epson flatbed for example provides better sharpness in the green channel than it does the other channels. We are talking major differences here. However when compared with scanning in colour and greyscale I think that the scanner does ignore the red and blue and defaults to the green (as it looks just like scans from the green).

<P>

I've also found that my epson scan software operates better when scanning as a positive and inverting (unlike my Nikon software). If you were using Vuescan I suspect it would make no difference choosing to scan as positve (then invert) or scan as negative. So far I have found scanning with Epsons scan and then doing everything in Photoshop works best for me. I invert and apply layers for curves and levels in there plus some small sharpening (say 100 ~ 200% sharpening, 0.7 ~ 1.4 pixel radius, 0 threshold) as my first step post scan.

<P>

Try reading <A HREF="http://www.largeformatphotography.info/1800F-bw.html" target="_blank"><b>this fellows findings</b></A> on his scanner

or pick around on <A HREF="http://cjeastwd.blogspot.com/" target="_blank"><b>my blog</b></A> for bits n pieces as I post them

[carter_hopkins]

I use a Nikon Super CoolScan 9000 & 4000 and get better results scanning in RGB. For most negatives Grayscale works just fine, but when I have a negative of lesser quality or if I am going to be making a large print, I find that I do get a better result with RGB; there is a slight color cast which needs to be dealt with in Photoshop. Can't say I have seen an increase in quality scanning a negative as a positive though.

[eric_brody]

Why don't you try it yourself both ways and carefully look at the results instead of relying on the highly variable quality of others' opinions. This is an issue easily resolved by an individual using the tools you must already have, a computer, a scanner, software, and a monitor. If you try it both ways and cannot see a difference on the screen or much more importantly in a test print of two, do it the easier way.

 

Good luck.

 

Eric

[john_wilson10]

Using my 10 year old Arcus II, I get a huge increase in quality scanning as a positive and inverting in photoshop (I blame the crummy Agfa scanning software). Even so, I spent two days trying to figure out the optimum settings for that scanner, and I couldn't get any scan to come out better than "embarassingly bad".

 

I tried the trial of vuescan, but I couldn't get a scan to output at a large DPI, and since the main problem I had with the scans was a smearing and jagging of fine detail, I couldnt evaluate whether it was the software or the scanner that was causing the problem at such small resolutions.

 

YMMV with older scanners; you may get better results scanning as a positive than as a negative, but as for me, I'm saving up for a new scanner.

[JeffOwen]

On my Canon 4200F I find using greyscale gives a better result.

[godfrey]

I use both RGB and Grayscale settings depending upon what film I'm scanning. I find that for

traditional B&W film emulsions there's no practical difference, but for C41 process films I

seem to get just a hair more tonal separation with RGB scans.

 

Whether there is any value in RGB vs Grayscale capture probably also depends on what

scanner you're using and the specific scan application/methodology you use.

 

Godfrey

[mendel_leisk]

You can always scan in color and desaturate later. In fact, you always *are* scanning in color: if you tell the scanner to scan in black and white it will still scan in color, desaaturate on the fly, using it's method, over which you likely have no control.

 

Also, it depends on your workflow. I know using a Vuescan Raw File workflow, if you output the initital raw file in 16 bit greyscale, there is a marked degradation of the quality of subsequent scan-from-disk output.

[bobmichaels]

Mendel makes the key point when he reminds us that the scanner always scans in color and simply desaturates. This is true even when the film being scanned is monochrome.

 

I use Vuescan also and can never tell any difference between the Vuescan desaturation and what I can do with channels. So I let Vuescan desaturate and save as a greyscale file simply because of the size.

 

Some point out that they do better saving as RGB and making the monochrome conversion themselves using an image editor. I think that is a function of the scanning software one is using.

[charles_gravely]

I use viewscan to scan from my 5400.

 

In documentation for Viewscan, when scanning B&W, viewscan will turn down the gain (or the light source?) to accommodate for B&W negatives. I can't seem to find the reference now though. Maybe there is more to the algorithm than just de-saturating the RGB image.

 

Anybody else recall seeing a reference to appropriate gain for scanning B&W negs versus color?

 

CG

[thomas_park]

My original bias was towards scanning in RGB, and I'm still standing behind that feeling. The 16-bit greyscale scans from my scanner (Epson Perfection, admittedly not the most powerful engine) resemble the green channel from an RGB scan of the same image.

 

In the RGB scan, the blue and green channels have the highest acutance,with the blue channel being slight lighter. I think this is due to the color of the film base, however.

 

The green channel looks very similar to the blue, only darker. The red channel is softer than the other two.

 

So, with my film on my scanner, it seems to make sense to scan RGB and mix the channels by hand. YMMV depending on your scanner, of course.