Silverfast 9 SE b&w film scanning: 48->24 bit or 16->8 bit or 16 bit HDR Raw for best result?

[Iridium]

Hello,

I'm a bit confused & need help.

I scan my b&w negatives in an EPSON V700 Photo with the Silverfast 9 SE. SE version doesn't have the option of 16 bit. Therefore, my question is: how do I get the best depth output?

1. Choosing 48->24 bit & then converting in Photoshop to Grayscale & 16 bit. Do I gain anything?

2. Choosing the 16 bit HDR Raw & then inverting it in Photoshop.

3. or directly the 16->8 bit.

 

[koraks]

1. Choosing 48->24 bit & then converting in Photoshop to Grayscale & 16 bit. Do I gain anything?

No gain; just a big file at the start.

2. Choosing the 16 bit HDR Raw & then inverting it in Photoshop.

That would be my choice assuming that this will indeed give a 'raw' scan.

3. or directly the 16->8 bit.

You could do that, but personally, my preference is to do any inversion and contrast adjustment manually in post processing, and do as little as possible during actual scanning.

I've always found Epson Scan to be perfectly adequate for my needs.

 

[Alan Edward Klein]

I use 48 bit for color film and 24 bit grayscale for BW film with Epsonscan on my V600 and V850. You can also deselect colors in 8 bit grayscale or BW mode. Never used it. Does Anyone know when they would?

 

[Alan9940]

I'm a little confused... I use Silverfast 8 SE on an old Mac that my Minolta 5400II is attached to, and I scan to 16-bit RAW regularly. Is this a change limitation in Silverfast 9 SE? I, also, scan to 16-bit RAW with Silverfast Ai Studio with an Epson scanner. I convert both 16-bit RAW outputs in PS with the ColorPerfect plugin. Been working up B&W film scans for many years this way and, for me, I found it to be the best solution.

 

[brbo]

I'm a little confused... I use Silverfast 8 SE on an old Mac that my Minolta 5400II is attached to, and I scan to 16-bit RAW regularly. Is this a change limitation in Silverfast 9 SE?

Yes, and? OP never said that he can’t do what you can in SF 8 SE.

He wants to know what people recommend, scanning raw in 16 bit or inverted and processed and saved as 8bit.

I would say it depends on what you value most. Speed, ultimate quality, control over the process…?

 

[Iridium]

Yes, and? OP never said that he can’t do what you can in SF 8 SE.

He wants to know what people recommend, scanning raw in 16 bit or inverted and processed and saved as 8bit.

I would say it depends on what you value most. Speed, ultimate quality, control over the process…?

Ultimate quality for me. As I don't have the option with Silverfast SE to scan directly in 16 bit, I'm trying to find the best alternative & I ask others what they choose when they use Silverfast. Thanx

 

[Alan Edward Klein]

Ultimate quality for me. As I don't have the option with Silverfast SE to scan directly in 16 bit, I'm trying to find the best alternative & I ask others what they choose when they use Silverfast. Thanx

Have you tried Epsonscan to get the 48 bit? It's free.

 

[brbo]

Ultimate quality for me. As I don't have the option with Silverfast SE to scan directly in 16 bit, I'm trying to find the best alternative & I ask others what they choose when they use Silverfast. Thanx

Then stick to 16bit raw and Photoshop.

And as orhers have said, Epson scan will do just as well. Only difference that it might not be able to provide gamma unencoded unprocessed 16bit file.

 

[Iridium]

Have you tried Epsonscan to get the 48 bit? It's free.

Of course I've tried Epson Scan, but the Silverfast is a more dedicated software for film scanning with great results. I have the Silverfast SE version & as I don't intend to upgrade, I just ask from other users their opinion for better results on b&w film scanning.

 

[koraks]

f course I've tried Epson Scan, but the Silverfast is a more dedicated software

Epson Scan is dedicated to your type of scanner and AFAIK offers more than enough control to get an excellent job done. I've used it for years on my 4990 - still do. It's a very effective and no-fuss alternative to SilverFast, which really offers very little additional benefit image-wise.

 

[Iridium]

Epson Scan is dedicated to your type of scanner and AFAIK offers more than enough control to get an excellent job done. I've used it for years on my 4990 - still do. It's a very effective and no-fuss alternative to SilverFast, which really offers very little additional benefit image-wise.

Thanx @koraks . In fact, I didn't make the post to argue for EpsonScan/Silverfast. I use the Silverfast SE & I just asked what other users recommend for output bit depth instead of direct 16bit grayscale which is not available in SE.

 

[koraks]

Well, my suggestion is to use a tool that offers the function you want. No by means of an argument, but as a genuine alternative that seems to fit your needs.

 

[Alan9940]

Yes, and? OP never said that he can’t do what you can in SF 8 SE.

He wants to know what people recommend, scanning raw in 16 bit or inverted and processed and saved as 8bit.

I would say it depends on what you value most. Speed, ultimate quality, control over the process…?

I read this statement "SE version doesn't have the option of 16 bit" by the OP as that he or she would like to scan to 16-bit grayscale, but that SE version 9 doesn't allow that option. I don't scan color, but IIRC the SE version of Silverfast software doesn't allow scanning directly to 16-bits per channel. However, it will (or, at least, v8 will) allow scanning to a 16-bit HDR file; I do this all the time with 35mm B&W film.

For the record, I always recommend scanning to the highest bit depth possible. But, what does that matter if your scanning software doesn't allow for doing that? And, also for the record, scanning 8-bit then converting to 16-bit ain't gonna buy ya nuttin; it's still only 256 levels of tone now stretched out over ~65K.

 

[Iridium]

I read this statement "SE version doesn't have the option of 16 bit" by the OP as that he or she would like to scan to 16-bit grayscale, but that SE version 9 doesn't allow that option. I don't scan color, but IIRC the SE version of Silverfast software doesn't allow scanning directly to 16-bits per channel. However, it will (or, at least, v8 will) allow scanning to a 16-bit HDR file; I do this all the time with 35mm B&W film.

For the record, I always recommend scanning to the highest bit depth possible. But, what does that matter if your scanning software doesn't allow for doing that? And, also for the record, scanning 8-bit then converting to 16-bit ain't gonna buy ya nuttin; it's still only 256 levels of tone now stretched out over ~65K.

Thanx @Alan9940 I do exactly stick to the "16-bit HDR Raw" option. It seems that I get better grayscale rendering for the 35mm b&w film.

 

[brbo]

However, it will (or, at least, v8 will) allow scanning to a 16-bit HDR file; I do this all the time with 35mm B&W film.

And OP clearly stated that he has this option.

And, also for the record, scanning 8-bit then converting to 16-bit ain't gonna buy ya nuttin; it's still only 256 levels of tone now stretched out over ~65K.

If you do heavy editing, converting 8bit files to 16bit before editing WILL help.

 

[Sirius Glass]

And OP clearly stated that he has this option.



If you do heavy editing, converting 8bit files to 16bit before editing WILL help.

How? There is only a fixed amount of data and more bits of the same image will not increase in quality.

 

[Alan9940]

How? There is only a fixed amount of data and more bits of the same image will not increase in quality.

Exactly! Thank you. I've had this very argument so many times that I'm not going into the technical details and reasons why here.

 

[brbo]

Exactly! Thank you. I've had this very argument so many times that I'm not going into the technical details and reasons why here.

Sometimes it helps to verify "theory" in practice.

Starting from the same 8bit gradient I left one file in 8bit and converted the other to 16bit. Then applied exactly the same (simple) curves to both. Here are both saved to jpg 8bit (click for full resolution), the difference is more than noticeable (if you don't notice the difference you need a better monitor).

8bit editing


16bit editing

 

[Sirius Glass]

It only appears to have higher resolution, but the amount of data is the same so if there is no more data, it is only an illusion.

 

[MattKing]

The reason to consider upscaling to 16 bit before editing isn't to increase the data.
It is to improve the appearance of the result after the edit.
Yes, the apparent increase in resolution comes purely from the interpolation added by the up-scaling process.
But if you are going to have to make the edit anyways to obtain the result you require, and you just have an 8 bit file to start with, at least the final result's appearance doesn't draw as much attention to the relative lack of information in the source file.
If you try to do this in the reverse order - edit first, then attempt to upscale the result to deal with the unwanted artifacts, the result will be less likely to appear visually satisfying.

 

[brbo]

It only appears to have higher resolution, but the amount of data is the same so if there is no more data, it is only an illusion.

No. Both files ahave exactly the same resolution.

 

[Sirius Glass]

No. Both files have exactly the same resolution.

So as I said there is no gain.

 

[Alan Edward Klein]

You would end up with 8 bits with sRGB anyway if that matters. The web basically uses 8 bit sRGB if I understand correctly.

 

[koraks]

So as I said there is no gain.

Yet, there is since the posterization in the 2nd example is considerably less. What this proves is that even if you demonstrate something, people will still want to believe whatever the heck they want.

The web basically uses 8 bit sRGB if I understand correctly.

To a large extent, but png is now widely supported which exceeds this 16 bit limit. Moreover, what @brbo demonstrates above is that there's merit for temporarily moving to 16 bit space for editing before (optionally) moving back to 8 bit for final display/export.

 

[Alan Edward Klein]

Yet, there is since the posterization in the 2nd example is considerably less. What this proves is that even if you demonstrate something, people will still want to believe whatever the heck they want.

To a large extent, but png is now widely supported which exceeds this 16 bit limit. Moreover, what @brbo demonstrates above is that there's merit for temporarily moving to 16 bit space for editing before (optionally) moving back to 8 bit for final display/export.

I scan at 48-bit 2400bpi and adjust in Lightroom as 48-bit sRGB then create a sRGB final for the web at a smaller resolution, let's say 1600 wide. If I'm using a digital photo for part of a video slideshow, I'll reduce the final sRGB to 3840x2160 (16:9) (that's 4K to fill a TV screen or monitor) to match any 16:9 video clips that might be part of the show. No black bars.