Author Topic: This might be a dumb (scanning) question! (Read 1940 times)

Nigel · « **on:** February 11, 2009, 08:08:02 PM »

I was scanning some b&w negs and I seemed to remember reading somewhere that (maybe with some scanning software) that you get better results on b&w using the colour settings. I think it said because colour scans in 24 bit instead of 16 bit.

Has anyone else read that, or have I imagined it?

If this is complete nonsense forget I asked.

Nigel

Francois · « **Reply #1 on:** February 11, 2009, 11:12:48 PM »

Not nonsense...
Regular B&W is 8 bit.
True Color is 24 bit.

If your software handles it, scanning 16 bit B&W is probably better.
I scan my color at full 48 bit color (3x 16 bit for RGB).

Rule of thumb: use the biggest numbers your scanner can dish out if you want maximum quality

Scanning resolution can be adapted to the enlargement you want.

LT · « **Reply #2 on:** February 12, 2009, 08:48:23 AM »

I a bitof a dumbkopf when it comes to matters digital - so please correct me if I'm on the wrong track here ... But, if the choice is between a 24 bitcolour scan or a 16 bit mono, isn't the mono the more detailed of the two because the colour scan is only 3 x 8 bit scans two of which will be discarded anyway if you're converting to greyscale?
I've tried scanning in colour with my tan & stain negs as that is supposed to realize the full benefit if the stain's effect, but to be honest, I didn't notice any perceptible difference, it was more of a pain actually cos I had to mess about to get a reasonable black and White conversion, a stage I don't have to bother with in 16 bit mono scans.

The again, I only scan to work out darkroom printing plans and assess the potential of a negative so don't worry about detail for electronic output.

Skorj · « **Reply #3 on:** February 12, 2009, 09:31:45 AM »

Not sure myself, Ed I think knows the business though. I do know I scan in color and at the highest number of bits as possible. I forget who told me this, but it was either Leon (who humbly knows more than he admits), or previously said Ed. Skj.

Pete_R · « **Reply #4 on:** February 12, 2009, 11:45:53 AM »

To add my two cents worth...

Leon's correct, 16 bit mono is better than 24 bit colour because 24 bit colour is just three channels of 8 bit.

It probably depends on what your scanner/software is capable of. Choose whatever gives you the highest number but always divide the colour bit depth by 3 to get the real figure.

Other things to consider are:

File size - RGB files will be much larger.

If you want to digitally tone the result, you need a RGB file. [Edit] But as Leon says, you can convert to RGB later.

If scanning as a colour neg, the scanner is probably expecting to find an orange film base colour and may try and correct for it giving you a wierd coloured scan. This can also cause saturation of one of the colour channels (brain hurting here but I think the blue channel is given more gain to compensate for the orange base colour, but if the orange base colour isn't there, the blue channel saturates). I know I used to get this with a previous scanner I had.

LT · « **Reply #5 on:** February 12, 2009, 12:41:04 PM »

my process is to scan at 16 bit black and white. do all my messing about with local and global contrast, then flatten and convert to 24-bit rgb and add toning at that point using colour balance in PS. Dont know if that is the best way to do it - as with all things in PS ... there are a million ways to achieve the same effect so it's best to find a method that gives you reliable and satisfactory results, then stick with it.

Francois · « **Reply #6 on:** February 12, 2009, 03:17:28 PM »

Actually, from what I can figure out, the scanner scans at higher bit depth and "squeezes" the colors into 24 bit format (for 24 bit color).

Scanning at higher bit depth is dependent on scanner and software. 16 bit B&W and 48 bit color are not formats compatible with every program.

The question is usually "how much do you need". 8 bit grayscale is not recommended because it has only 256 tones. So if your scanner or software doesn't support 16 bit grayscale (which gives 65,536 tones of gray), you're better off scanning in 24 bit color and then desaturate the image to get grayscale (leaving the image in RGB mode). This will put more data in the image (including the off color casts) and that's what we want.

If I put my choices, I prefer to scan B&W in 16 bit gray. If I can't, then I choose 24 bit color and desaturate.

As for toning, it's quite easy using Duotones. There's probably also an adjustment layer trick that's even easier.

Skorj · « **Reply #7 on:** February 12, 2009, 10:48:08 PM »

Too many options! But, clearly I can be trying something else. Arigato. Skj.

Nigel · « **Reply #8 on:** February 13, 2009, 08:32:43 AM »

Mmm. As Skorj says a lot of opinions.

As there isn't a definitive "it's definitely better using colour" I think I'll stick with my 16 bit B&w scans. An interesting discussion though. Thanks for all the advice guys.

Nigel