Working with 16-bit Files -- Some Thoughts

[Sepiana]

Question:

Should one work with 16-bit files in Photoshop Elements?

Answer:

I believe the emphasis placed on working with 16-bit files makes a lot of sense if you have Lightroom or Photoshop. However, when you go into Elements "territory", it is a different ball game. This program offers limited support for 16-bit files. In a nutshell, these are its limitations:

- You cannot create a 16-bit file or convert to one.
- You cannot work with layers.
- Quite a few filters won't work (mainly the ones in the Filter Gallery).
- The following tools won't work:

Selection Brush
Magic Wand
Type
Straighten
Recompose
Cookie Cutter
Brush-based tools (Paint, Dodge, Blur, Eraser, etc.)
Shape-based tools (Custom Shape, Rectangle, Ellipse, etc.)

Question:

Why is this option available in Elements?

Answer:

Here is a reason for having the 16-bit option in Elements (although limited) -- third-party plug-ins. Many plug-ins (such as Topaz, Filter Forge, Nik Collection, Auto FX, the Flood filter, Fractalius-version 2, etc.) offer support for 16-bit files.

Question:

What do the experts have to say about this issue?

Answer:

The general consensus seems to be:

Work with a 16-bit file in Elements as long as you can and then change it to an 8-bit file when it becomes necessary. You will know the time has come because some features or tools won't be available.

NOTE

Elements offers two ways to change a 16-bit file into an 8-bit file.

Inside the ACR Converter:

Go to the bottom of the ACR Converter dialog. In the Depth box, select 8 Bits/Channel.

Inside the Editor:

Go to Image > Mode > Select 8-bits/Channel.

[michelb]

Mar 21, 2017 6:41:37 GMT Sepiana said:

Question:

Should one work with 16-bit files in Photoshop Elements?

Answer:

I believe the emphasis placed on working with 16-bit files makes a lot of sense if you have Lightroom or Photoshop. However, when you go into Elements "territory", it is a different ball game. This program offers limited support for 16-bit files. In a nutshell, these are its limitations:

- You cannot create a 16-bit file or convert to one.
- You cannot work with layers.
- Quite a few filters won't work (mainly the ones in the Filter Gallery).
- The following tools won't work:

Selection Brush
Magic Wand
Type
Straighten
Recompose
Cookie Cutter
Brush-based tools (Paint, Dodge, Blur, Eraser, etc.)
Shape-based tools (Custom Shape, Rectangle, Ellipse, etc.)

Question:

Why is this option available in Elements?

Answer:

Here is a reason for having the 16-bit option in Elements (although limited) -- third-party plug-ins. Many plug-ins (such as Topaz, Filter Forge, Nik Collection, Auto FX, the Flood filter, Fractalius-version 2, etc.) offer support for 16-bit files.

Question:

What do the experts have to say about this issue?

Answer:

The general consensus seems to be:

Work with a 16-bit file in Elements as long as you can and then change it to an 8-bit file when it becomes necessary. You will know the time has come because some features or tools won't be available.

I do agree with everything (nice summary)

Since comments are welcome, I'd like to stress a few facts:

1 - If you work in the ACR module, all your calculations are done in 16-bits. The menu choice for 8 or 16 bits is only to choose the format when sending the edited image to the editor. If you start from a raw file you start with 12 or 14 bits precision. The difference between 12 or 14 bits is very hard to reveal. If you start from a jpeg (necessarily 8-bits), the advantage is that you don't get 'rounding errors' in calculations, providing smoother gradients. The other advantage of the ACR module is that you are working in a wider color model as the common sRGB. That is helped with the 16-bits ACR editing. The result is also finer color control. You'll see that editing in ACR first makes 16-bits editing useless (redundant) for ulterior editing in nearly 100% of your shots...
2 - Hard fact: your display and home printer are only 8-bits.
3 - Starting with a 16 bits image, if you convert it to 8-bits (same file), you can't see the difference on display or print, even with a 16-bits printer. That's why the jpeg format with low compression is ideal for printing output.
4 - 16-bits editing is useful, not only for theoretical reasons, but for a very real and annoying limitation in 8-bits: our eyes are apt to grasp the difference between two adjacent areas in very soft monochrome gradients, such as in a sky. The ugly result may be posterization, or banding. With finer tones in 16-bits, the gradient can be smoother. Even with 16-bits, posterization can happen (the cure is to add some noise). This type of banding may be the result of using extreme edits stretching the tones or some layers blending modes like difference. Since 16-bits mode is not supported with layers in Elements, be careful with such edits; however you are safe if you work in 'normal' mode. The lack of 16-bits for 'detailed' adjustments like sharpening or dust removal is never a problem, it's only with wide areas gradients.
5 - So, keep an eye for posterization if you think your image has such smooth gradients. Be aware that it is rather common to see false posterization artifacts only on your display (some are not even 8-bits...), check that on test prints.

To summarize, don't make your worklow harder than needed and don't think you have to upgrade to Photoshop for that reason.
Work in ACR first, know when you risk posterization and which kinds of hard edits may produce it (especiall in black and white).

Just for play:
Choose a file with a soft sky gradient.
duplicate the layer and use the filter adjustment >> posterize

Set layers with the values : 128 (7  bits), 64 (6-bit), 32 (5 bits)

That gives you an idea of the loss of gradient smoothness with lower bit depth.

[Sydney]

This is some very interesting reading. Thank you both for taking time to shed some light on working with 16 bit files. I am going to print this one out for my notebook of good things to know.

[PeteB]

I’ve been playing and experimenting with 16 bit images since Mark Galer first posted a PSE 10 tutorial about a 16 bit workflow.
www.youtube.com/watch?v=sA2_pEg-8pQ

When working in PSE there are work arounds that make it possible to create layers and add masks. Also, Nik Color Efex and Topaz filters can be applied. When I read this post, I decided to open a recent photo and play with it as a 16 bit image. Below i posted a screen shot of my workflow. The top red arrow shows that it is a 16 bit image. The arrow pointing to layer four reads … stamp visible+mask+topaz b&w. I also sharpened with high pass filter and added a vignette.

Conclusion, you can edit a 16 bit image in PSE … BTW, my Epson printer allows for 16 bit printing and I’ve tried it but I did not see any visible difference (comparing to a printed 8 bit image).  

I should also note that once a new layer is created, Stamp Visible becomes available. A mask can be created by first creating a selection and then selecting the mask icon.

2812blog.files.wordpress.com/2017/03/harbor-bikes.jpg

[Sepiana]

Mar 21, 2017 9:14:35 GMT michelb said:

I do agree with everything (nice summary)

Michel, thanks! Glad to hear we are in agreement.

Since comments are welcome, I'd like to stress a few facts:

1 - If you work in the ACR module, all your calculations are done in 16-bits. The menu choice for 8 or 16 bits is only to choose the format when sending the edited image to the editor. If you start from a raw file you start with 12 or 14 bits precision. The difference between 12 or 14 bits is very hard to reveal. If you start from a jpeg (necessarily 8-bits), the advantage is that you don't get 'rounding errors' in calculations, providing smoother gradients. The other advantage of the ACR module is that you are working in a wider color model as the common sRGB. That is helped with the 16-bits ACR editing. The result is also finer color control. You'll see that editing in ACR first makes 16-bits editing useless (redundant) for ulterior editing in nearly 100% of your shots...
2 - Hard fact: your display and home printer are only 8-bits.
3 - Starting with a 16 bits image, if you convert it to 8-bits (same file), you can't see the difference on display or print, even with a 16-bits printer. That's why the jpeg format with low compression is ideal for printing output.
4 - 16-bits editing is useful, not only for theoretical reasons, but for a very real and annoying limitation in 8-bits: our eyes are apt to grasp the difference between two adjacent areas in very soft monochrome gradients, such as in a sky. The ugly result may be posterization, or banding. With finer tones in 16-bits, the gradient can be smoother. Even with 16-bits, posterization can happen (the cure is to add some noise). This type of banding may be the result of using extreme edits stretching the tones or some layers blending modes like difference. Since 16-bits mode is not supported with layers in Elements, be careful with such edits; however you are safe if you work in 'normal' mode. The lack of 16-bits for 'detailed' adjustments like sharpening or dust removal is never a problem, it's only with wide areas gradients.
5 - So, keep an eye for posterization if you think your image has such smooth gradients. Be aware that it is rather common to see false posterization artifacts only on your display (some are not even 8-bits...), check that on test prints.

To summarize, don't make your worklow harder than needed and don't think you have to upgrade to Photoshop for that reason.
Work in ACR first, know when you risk posterization and which kinds of hard edits may produce it (especiall in black and white).

Just for play:
Choose a file with a soft sky gradient.
duplicate the layer and use the filter adjustment >> posterize

Set layers with the values : 128 (7 bits), 64 (6-bit), 32 (5 bits)

That gives you an idea of the loss of gradient smoothness with lower bit depth.

Thanks once again! Your knowledgeable contribution adds valuable information to the 16-bit file issue. It emphasizes the benefits from starting your workflow in ACR as well as the risks of posterization and hard-editing.  Not too many Elements users are aware of these risks.

[Sepiana]

Mar 21, 2017 21:43:48 GMT PeteB said:

I’ve been playing and experimenting with 16 bit images since Mark Galer first posted a PSE 10 tutorial about a 16 bit workflow.
www.youtube.com/watch?v=sA2_pEg-8pQ

When working in PSE there are work arounds that make it possible to create layers and add masks. Also, Nik Color Efex and Topaz filters can be applied. When I read this post, I decided to open a recent photo and play with it as a 16 bit image. Below i posted a screen shot of my workflow. The top red arrow shows that it is a 16 bit image. The arrow pointing to layer four reads … stamp visible+mask+topaz b&w. I also sharpened with high pass filter and added a vignette.

Conclusion, you can edit a 16 bit image in PSE … BTW, my Epson printer allows for 16 bit printing and I’ve tried it but I did not see any visible difference (comparing to a printed 8 bit image).  

I should also note that once a new layer is created, Stamp Visible becomes available. A mask can be created by first creating a selection and then selecting the mask icon.

2812blog.files.wordpress.com/2017/03/harbor-bikes.jpg

Pete, thanks for sharing this tutorial! Great workaround! It enables Elements to duplicate layers and add masks when working with 16-bit files. Unfortunately, this program's limitations still remain -- a few filters (mainly the ones in the Filter Gallery) won't work; several of the tools cannot be applied; and you cannot work with layers. If you save your 16-bit file as a multi-layered one, when you open it back in Elements, you will get this message. 

[Chris]

Thank you all for sharing this very interesting information. Some time ago I did an experiment correcting an overexposed image from a raw file in ACR compared to correcting an 8 bit jepg in ACR. The results speak for themselves.




My personal workflow preference is to correct what I can as a raw file and then continue to edit as an 8 bit file in PS Elements in the Adobe RGB colour space.

It's also worth mentioning that 16 bit PSD files can be more than double the size of an 8 bit file.

Kind regards
Chris

[michelb]

Jun 8, 2024 19:55:04 GMT Chris said:

My personal workflow preference is to correct what I can as a raw file and then continue to edit as an 8 bit file in PS Elements in the Adobe RGB colour space.

It's also worth mentioning that 16 bit PSD files can be more than double the size of an 8 bit file.

Kind regards
Chris

Hi Chris,

I agree with your workflow except for working in the Adobe RGB color space.

That's already four or five versions which no longer allow choosing that workspace from a raw original, contrary to older versions and to the help doc which has not been corrected.

This huge bug has been confirmed by Camera raw experts, but there is no sign it will be cared for. In a nutshell, the only way according to the help doc is to set the menu for color settings in the PSE editor is to always optimize for print. Whatever choice you are entering, you only get sRGB. Just check it yourself. Converting sRGB to ARGB afterward does not recover the loss of color range. Probably not a showstopper, but it's good to know.

(Thanks for showing the difference with a jpg.)