Scanning black and white negatives in 48-bit instead of 16-bit?

[Peter K 75]

Hi,

I recently had a short chat with the Silverfast Support Team. They recommended me scanning my medium format black and white negatives as 48-bit HDR TIFFs instead of 16-bit HDR TIFFs.

According to them, scanning in 48-bit would allow me to capture more detailed information from my negatives, including subtle variations and any details contained in the negative mask. This extra data could be particularly beneficial during post-processing in Photoshop, as it would give me greater flexibility when adjusting tonal ranges and contrasts.

Since I'm pretty obsessed with trying to get as much information from the negatives as my scanner (an Epson V850 Pro) is able to deliver, I was wondering:

Question #1: Does anyone know if this is correct? Should I be able to extract more information from black and white negatives scanning in 48-bit? And why would that be?

Question #2: If yes, how should I post-proces in Photoshop afterwards to utilize this extra information? For example, would it be ok to add a black and white layer to the image early in the workflow? Or would that ruin any potential benefits from scanning in 48-bit?

I might add that I always scan as flat as possible in Silverfast and do all the post-processing in Photoshop afterwards (including any changes to the histogram etc.).

Any help and advice on the subject will be greatly appreciated. Thanks in advance!

- Peter.

 

[ishwanu]

I recently had a short chat with the Silverfast Support Team. They recommended me scanning my medium format black and white negatives as 48-bit HDR TIFFs instead of 16-bit HDR TIFFs.

Apparently you're talking about scanning in 48-bit color vs. 16-bit grayscale.

Unless there's some color information in or on your B&W negatives, I don't see the point.

 

[Overrank]

I’ve heard of people scanning to 48bit colour for image restoration. Some marks can be removed as they only exist in the red, green or blue channels. I imagine that you might be able to use the colour information to a better final image from a good negative.

 

[Svein Eriksen]

I've seen the claim about scanning in color before, don't remember where. I've also heard the claim is BS, don't remember where. If the film isn't completely neutral, then scanning in color may add some info compared to a BW scan. To me it seems unlikely that 48 bit color scans should be visibly better, but if anyone know it should be Silverfast support...

If you get an advantage then I'd keep it as 48 bit in PS as long as possible, but maybe Silverfast support can give you a definitive answer.

 

[Peter K 75]

I’ve heard of people scanning to 48bit colour for image restoration. Some marks can be removed as they only exist in the red, green or blue channels. I imagine that you might be able to use the colour information to a better final image from a good negative.

Thanks for your answer. I'll keep that in mind.

 

[Peter K 75]

I've seen the claim about scanning in color before, don't remember where. I've also heard the claim is BS, don't remember where. If the film isn't completely neutral, then scanning in color may add some info compared to a BW scan. To me it seems unlikely that 48 bit color scans should be visibly better, but if anyone know it should be Silverfast support...

If you get an advantage then I'd keep it as 48 bit in PS as long as possible, but maybe Silverfast support can give you a definitive answer.

Thanks for your answer. Much appreciated.

Yeah, the guys from Silverfast should know. And I'm sure they do. But like you say, it also seems unlikely to me that 48-bit color scans would be visibly better. At least, I can't figure out why that would be from a technical point of view when the negative is in black and white.

I guess I will do some tests to see if I can see any difference. And I will follow your advice to keep the file in 48 bit as long as possible during the workflow in PS.

Thanks again.

 

[D Cox]

I've seen the claim about scanning in color before, don't remember where. I've also heard the claim is BS, don't remember where. If the film isn't completely neutral, then scanning in color may add some info compared to a BW scan. To me it seems unlikely that 48 bit color scans should be visibly better, but if anyone know it should be Silverfast support...

If you get an advantage then I'd keep it as 48 bit in PS as long as possible, but maybe Silverfast support can give you a definitive answer.

Thanks for your answer. Much appreciated.

Yeah, the guys from Silverfast should know. And I'm sure they do. But like you say, it also seems unlikely to me that 48-bit color scans would be visibly better. At least, I can't figure out why that would be from a technical point of view when the negative is in black and white.

I guess I will do some tests to see if I can see any difference. And I will follow your advice to keep the file in 48 bit as long as possible during the workflow in PS.

Thanks again.

Doing your own tests is the only way to find out for certain.

Don

 

[flektogon]

Here is what I found in one article (and it seems to be serious ):

The Dynamic Range and Resolution Of The Human Eye

Although the human eye is capable of a contrast ratio of 1 or 2 million to 1, (about 20 stops), from midnight black to bright sunlit white, yet it takes the eye some time to adjust to these extremes. For the purposes of photography we are interested in its instantaneous dynamic range, about which there is some disagreement, but it is of the order of 1000:1footnote 2 (about 10 stops). And the viewing devices have even more limited dynamic range. LCD monitors around 1000:1 (10 bits), printed paper even less, 500:1. The best are the OLED monitors, still less than 16 bit DR.

So, theoretically 16, but practically only 10 bits (per channel) of resolution is needed. The best digital cameras use just 14 bits (RAW). Thus 14 bits for B&W images is more than enough. The films themselves have pretty limited dynamic range, not exceeding 12 bits (Dmax ~ 3.6).

Now, scanning B&W through RGB filters won't bring any additional information.

--
Regards,
Peter

 

[Bob]

Color has three channels: red, green, and blue. If each channel is 16-bits, then you have 3x16=48 bits.

B&W (greyscale) only has 1 channel. Say that’s 16 bits too.

So if there is no color information at all on the negative, then yeah, you’re better off scanning 16 bit greyscale. It will be faster and a smaller file.

 

[flektogon]

Hi,

I recently had a short chat with the Silverfast Support Team. They recommended me scanning my medium format black and white negatives as 48-bit HDR TIFFs instead of 16-bit HDR TIFFs.

According to them, scanning in 48-bit would allow me to capture more detailed information from my negatives, including subtle variations and any details contained in the negative mask. This extra data could be particularly beneficial during post-processing in Photoshop, as it would give me greater flexibility when adjusting tonal ranges and contrasts.

Since I'm pretty obsessed with trying to get as much information from the negatives as my scanner (an Epson V850 Pro) is able to deliver, I was wondering:

Question #1: Does anyone know if this is correct? Should I be able to extract more information from black and white negatives scanning in 48-bit? And why would that be?

Question #2: If yes, how should I post-proces in Photoshop afterwards to utilize this extra information? For example, would it be ok to add a black and white layer to the image early in the workflow? Or would that ruin any potential benefits from scanning in 48-bit?

I might add that I always scan as flat as possible in Silverfast and do all the post-processing in Photoshop afterwards (including any changes to the histogram etc.).

Any help and advice on the subject will be greatly appreciated. Thanks in advance!

- Peter.

Well, I already replied that:

Now, scanning B&W through RGB filters won't bring any additional information.

But on the second thought, while there will be no more details (or so) retrieved, this "triple" scanning may yield lower noise caused by the scanning process. The grain contained in the film might be affected as well, though this is questionable.

Regarding the "right" post processing, I think (just my guts are telling me) that you don't need to do anything, treat the files as if they were "coloured". I do the same (i.e. nothing) when I digitally re-shoot B&W prints. Definitely don't activate any (automatic) orange layer removal, which the colour negatives contain.

--
Regards,
Peter

 

[Peter K 75]

Color has three channels: red, green, and blue. If each channel is 16-bits, then you have 3x16=48 bits.

B&W (greyscale) only has 1 channel. Say that’s 16 bits too.

So if there is no color information at all on the negative, then yeah, you’re better off scanning 16 bit greyscale. It will be faster and a smaller file.

Thanks for your answer. That makes sense.

But just to make sure: When I shoot black and white film (FP4+, HP5+ etc.), I guess there can't be any color information to scan? I mean, the negatives will always be monochromatic, right?

Or am I missing something? Can black and white film contain color information that makes scanning in 48 bit relevant?

I know this is a noob question. Sorry for that. I am relatively new to film photography. So I am just trying to understand the technical side of it all

Thanks again.

- Peter.

 

[Peter K 75]

Well, I already replied that:

Now, scanning B&W through RGB filters won't bring any additional information.

But on the second thought, while there will be no more details (or so) retrieved, this "triple" scanning may yield lower noise caused by the scanning process. The grain contained in the film might be affected as well, though this is questionable.

Regarding the "right" post processing, I think (just my guts are telling me) that you don't need to do anything, treat the files as if they were "coloured". I do the same (i.e. nothing) when I digitally re-shoot B&W prints. Definitely don't activate any (automatic) orange layer removal, which the colour negatives contain.

Thanks (for both of your answers). Much appreciated!

So while I won't get any additional information by scanning B&W film in 48 color, this proces might yield less noise. That's very interesting and definitely something I will A/B test.

I will also try and post process the files as regular color files as you mention (no black and white layers and no conversion to grayscale).

I don't think my software (Silverfast and Photoshop) is set to any kind of automatic orange layer removal. But I will double-check on that.

Thanks again.

- Peter.

 

[Overrank]

Color has three channels: red, green, and blue. If each channel is 16-bits, then you have 3x16=48 bits.

B&W (greyscale) only has 1 channel. Say that’s 16 bits too.

So if there is no color information at all on the negative, then yeah, you’re better off scanning 16 bit greyscale. It will be faster and a smaller file.

Thanks for your answer. That makes sense.

But just to make sure: When I shoot black and white film (FP4+, HP5+ etc.), I guess there can't be any color information to scan? I mean, the negatives will always be monochromatic, right?

Or am I missing something? Can black and white film contain color information that makes scanning in 48 bit relevant?

There is discussion here https://www.photrio.com/forum/threads/the-beauty-of-scanning-black-and-white-in-color.162149/ that scanning in colour and then dumping two of the channels results in a sharper image (I’ve heard this elsewhere too). But then it’s spoiled because some say just keep the green channel, some just the blue. The blue sounds *slightly* more plausible because of the shorter wavelength, but tbh the wavelength of blue light is so much smaller than the scanning resolution that I can’t really see it.

It would be pretty simple to test - do a B+W scan and then a colour scan, and then subtract each layer separately from the B+W scan and see what was the difference.

I know this is a noob question. Sorry for that. I am relatively new to film photography. So I am just trying to understand the technical side of it all

Thanks again.

- Peter.

 

[Karl Heinz W]

Hi,

I recently had a short chat with the Silverfast Support Team. They recommended me scanning my medium format black and white negatives as 48-bit HDR TIFFs instead of 16-bit HDR TIFFs.

According to them, scanning in 48-bit would allow me to capture more detailed information from my negatives, including subtle variations and any details contained in the negative mask. This extra data could be particularly beneficial during post-processing in Photoshop, as it would give me greater flexibility when adjusting tonal ranges and contrasts.

Since I'm pretty obsessed with trying to get as much information from the negatives as my scanner (an Epson V850 Pro) is able to deliver, I was wondering:

Question #1: Does anyone know if this is correct? Should I be able to extract more information from black and white negatives scanning in 48-bit? And why would that be?

Question #2: If yes, how should I post-proces in Photoshop afterwards to utilize this extra information? For example, would it be ok to add a black and white layer to the image early in the workflow? Or would that ruin any potential benefits from scanning in 48-bit?

I might add that I always scan as flat as possible in Silverfast and do all the post-processing in Photoshop afterwards (including any changes to the histogram etc.).

Any help and advice on the subject will be greatly appreciated. Thanks in advance!

- Peter.

Where is the problem?

Try it out and see what is possible with your setup.

I tried it with my Nikon Coolscan and I can't see any advantage in color scanning black and white negatives. What I do see is that I was able to improve the quality a bit by scanning at the maximum resolution Silverfast offers (even if it is higher than the scanner's maximum resolution) and scaling down in Photoshop.

I had the CMS 20 in my scanner. I assume you won't see any difference with “standard” films like a HP5.

 

[flektogon]

Well, I already replied that:

Now, scanning B&W through RGB filters won't bring any additional information.

But on the second thought, while there will be no more details (or so) retrieved, this "triple" scanning may yield lower noise caused by the scanning process. The grain contained in the film might be affected as well, though this is questionable.

Regarding the "right" post processing, I think (just my guts are telling me) that you don't need to do anything, treat the files as if they were "coloured". I do the same (i.e. nothing) when I digitally re-shoot B&W prints. Definitely don't activate any (automatic) orange layer removal, which the colour negatives contain.

Thanks (for both of your answers). Much appreciated!

So while I won't get any additional information by scanning B&W film in 48 color, this proces might yield less noise. That's very interesting and definitely something I will A/B test.

I will also try and post process the files as regular color files as you mention (no black and white layers and no conversion to grayscale).

I don't think my software (Silverfast and Photoshop) is set to any kind of automatic orange layer removal. But I will double-check on that.

Thanks again.

- Peter.

Peter,

One advantage of scanning B&W film through RGB filters can be, that in PP you can modify magnitude of individual "colours", thus simulate effect of any filter, like yellow or green, or whatever can be used in the B&W photography.

--
Regards,
Peter

 

[Peter K 75]

Peter,

One advantage of scanning B&W film through RGB filters can be, that in PP you can modify magnitude of individual "colours", thus simulate effect of any filter, like yellow or green, or whatever can be used in the B&W photography.

Thanks for this input also.

These are great ideas. Now I understand, that even though the extra information obtained by scanning B&W film as 48-bit color may be limited, it still gives me more flexibility in other areas of the post-processing phase.

I am definitely moving towards scanning as 48-bit color in the future to keep all doors open. Even though the file size is obviously bigger.

Thanks again,
Peter.

 

[Peter K 75]

But just to make sure: When I shoot black and white film (FP4+, HP5+ etc.), I guess there can't be any color information to scan? I mean, the negatives will always be monochromatic, right?

Or am I missing something? Can black and white film contain color information that makes scanning in 48 bit relevant?

There is discussion here https://www.photrio.com/forum/threads/the-beauty-of-scanning-black-and-white-in-color.162149/ that scanning in colour and then dumping two of the channels results in a sharper image (I’ve heard this elsewhere too). But then it’s spoiled because some say just keep the green channel, some just the blue. The blue sounds *slightly* more plausible because of the shorter wavelength, but tbh the wavelength of blue light is so much smaller than the scanning resolution that I can’t really see it.

It would be pretty simple to test - do a B+W scan and then a colour scan, and then subtract each layer separately from the B+W scan and see what was the difference.

Thanks a lot for the link! That thread contains a lot of the answers that I was looking for when it comes to the reasons for scanning B&W film in 48-bit color.

I haven't had a chance to do the test that you mention yet. But I will certainly do so as soon as I get the chance.

When I started this thread, I really didn't understand the reasoning for scanning B&W film in color. Now I do. And I'm definitely leaning towards doing that in the future to keep all doors open in the post processing phase.

So thanks again. Much appreciated!

- Peter.