Does PSE6 for Mac support 16-bit files?

[Rich]

I use Canon's Digital Photo Professional (DPP) software to process my RAW files and then convert to 16-bit tiff. Will PSE6 accept these files for editing and leave them as 16-bit files? Thanks.

[ljameso1]

The editing tools are really limited with 16bit files in PSE. Lightroom can handle them.

[Rich]

I'm not familiar with Lightroom, nor PSE6 for that matter. I'm assembling what I need to set up a digital darkroom on a limited budget for "serious" amateur use producing large display prints -- thus my interest in PSE vs. Photoshop.

Is Lightroom better than Photoshop CS3 for post editing of 16-bit image files -- mostly landscapes?

In what way is PSE6 limited in editing 16-bit files? Thanks.

[Michel B]

In what way is PSE6 limited in editing 16-bit files? Thanks.

First it should be noted that 16 bits editing has nothing to do with a better ability to obtain large prints. It is very important if you have to do severe adjustments such as curves which can lead to posterization, that is uneven gradation. Most edits of this can kind can lead to a visible difference between 16 and 8 bits. Otherwise, I would say that since most output device are 8 bits anyway, the last conversion to 8 bits won't show you any difference. If you can use the raw converter to get a proper 8 bits file, psd or tiff for further local edits, my experience is that you almost never get any risk of posterization, since you no longer need to use curves or other tools which produce posterization due to rounding calculations in 8 bits.

Back to PSE6: the 16 bits limitation is mainly the impossibility to use layers. What you can do in 16 bits is important, since this covers enhancements and adjustments which are risky for posterization: levels, shadow/highlight. Most plug-ins can work in 16 bits, noise removal, curves, lens distortion. So is it necessary to export from DPP or ACR in 16 bits? not necessarily, since you can theoretically achieve those results in the raw converter. I find it is best to export in 16 bits to fine-tune the raw conversion, possibly use smartcurves or plug-ins, then I convert to 8 bits to work with layers and tools like the clone tool, healing brush.

Should you use LR as a raw converter, I am positive PSE would be an excellent solution.

[Rich]

Thanks Michel, very helpful, considering that I'm new to most of this.
Most edits of this can kind can lead to a visible difference between 16 and 8 bits. Otherwise, I would say that since most output device are 8 bits anyway, the last conversion to 8 bits won't show you any difference. If you can use the raw converter to get a proper 8 bits file, psd or tiff for further local edits, my experience is that you almost never get any risk of posterization...

So are you saying that certain types of editing in the RAW converter program (Canon's DPP in my case) as a 16 bit image works better than in 8 bit -- and after such editing then converting the image to an 8 bit tiff file for exporting to PSE6 for working on layers and detail editing?

[Michel B]

Yes, that's what I mean.

To answer your first question, yes, you can open the 16 bits tiff produced by DPP, work in 16 bits in PSE and save your work in 16 bits in Elements as tiff or psd.
But...
1- the kind of editing PSE can do in 16 bits is not much more than what you can already achieve in DPP (or ACR), so you won't need it very often.
2- Once you have edited in DPP, most risks of quality degradation are very limited with usual tools
3- For creative use of layers with blend modes etc, there is again a risk of posterization, but that is rare with photo styles like landscape, street scenes, sport...
4- For landscapes, posterization will commonly be found on flat areas like skies. Generally, this can be cared for from the begining in DPP or ACR.
5- working with many layers in 8 bits results in much small working files and increase speed.
6- If you have a situation such as a difficult sky in a landscape, or under or overexposed parts in an image, it is often possible to create two optimum raw conversion in 8 bits, and combine them in different layers in PSE afterwards.

[Rich]

Thanks once again Michel. I think that I'm now getting a general overview of the best work flow for me while utilizing DPP and PSE6.

4- For landscapes, posterization will commonly be found on flat areas like skies. Generally, this can be cared for from the begining in DPP or ACR.
6- If you have a situation such as a difficult sky in a landscape, or under or overexposed parts in an image, it is often possible to create two optimum raw conversion in 8 bits, and combine them in different layers in PSE afterwards.

Could you please expound a little further on #4, caring for posterization from the begining in DPP or ACR; and #6, the creation of two optimum 8 bit raw conversion files for layers in PSE? I guess what I'm asking is how do you optimize a RAW image, or portions thereof, in your conversion program in order to avoid posterization later in PSE?

[Michel B]

#4 - Avoiding posterization in DPP or ACR. That means spotting the areas which have smooth gradation and trying to get them right without any need to use curves, levels, shadow/highlight or such afterwards in PSE. Working in 8 bits increases the risk of rounding errors creating visible differences. You can check this by looking at the produced 8 bits file at 100% magnification (that's a case where pixel peeping is good...) At any other magnification you cannot trust your display. Theoretically, posterization is prone to happen in very dark parts, but it can be seen on light parts, such as very light studio backgrounds.

#6 - Let's suppose you are able to get the risky part right, but that is bad for the rest of the picture. Then you could produce 2 different 8 bits files optimized for each part. Combining them in PSE is a matter of layers and masks, but you'll keep the best of both. Don't worry, this case is not very common unless you are looking for very dramatic look or you are dealing with extreme contrasts.

Also, posterization may happen even in 16 bits, and there are ways to deal with this... by adding some noise to mask the sharp differences showing in a gradient.

[Rich]

Appreciate the details, Michel -- well explained. With time and practice I should be able to spot those areas in my images that may be prone to posterization if edited later in PSE.

I'm hoping that once I adjust an area in RAW editing, let's say the sky, that it will appear the same once it is converted to 8 bit and opened in PSE -- requiring no further editing that may cause posterization.

However, if the image changes in some way, then that could be rather frustrating trying to "guess" how my edited RAW image will appear in PSE. Have you heard of any reports of such problems? Thanks again.

[Michel B]

.
However, if the image changes in some way, then that could be rather frustrating trying to "guess" how my edited RAW image will appear in PSE. Have you heard of any reports of such problems? Thanks again.

I don't recall if there is such a way to pre-visualize the 8 bit result while still in DPP; there is no way in ACR, you have to open it in PSE. My advice would be to check for this as soon as you open your file in PSE before you plan your future edits. You won't lose much time. Similarly, it is a good habit either to keep your edited file in PSE without flattening the layers and save them at least temporarily as tiff or psd until they are printed. However you should be safe by just examining your image at 100% magnification when finished with editing.

[Kvon]

Wow - good info, I just stumbled across this thread. I am also debating whether to splurge on the full-blown PS in order to get the 16 bit functionality. (The tips in this thread just might save me some ca$h!). As a follow-up question, can anyone provide a list (or a link to a list) of all of the capabilities in PSE 6 that are limited to 8 bit? Haven't been able to find this anywhere.

Thanks!

[TonyW]

I don't know of a list but as long as it doesn't require a layer quite a lot of functions are supported. You might be in interested in this post from the Lightform forum here and the reference in it to Matt's comments on 8 bit vs 16 bit editing.

http://www.elementsvillage.com/forums/showpost.php?p=425589&postcount=2

Admittedly you can do a lot more 16 bit RAW file editing in Lightroom than you can in PSE's Camera Raw editor but it does question the need to post process in 16 bit.

Tony

[Kvon]

Very helpful, thank you!

So if I understand correctly -- if I have a situation that demands a 16 bit file (heavy color or tone corrections required), I can either
a) do those manipulations in Lightroom and export an 8 bit file to PSE6, OR
b) export a 16 bit file to PSE6, do the color/tone corrections, then convert to 8 bit for the layers work...?

Thanks again!

[spigmu]

Also, it's important to note that spot application of filters via tools is disabled in 16 bit. This kicks me since I tend to do a lot of it in PS on my work Mac. On my personal MBP I get around it as best I can if I need it by applying the filter to a selection to varying degrees, altering the selection and filter parameters. This is coming out of Aperture, but I find the spot tools there unsatisfying.

I'm glad to have found this forum and thread, because I have an important question that I can't get an answer to, even on the Adobe site. Which is: when Elements (I'm on the current 6 for OSX) converts an image from 16 to 8 bit, how is it accomplishing this? Is it simply truncating the bottom eight bits? Or is a dithering algorithm applied? I'm an audio guy by trade, and going to a lower bit resolution in audio by truncating elicits screams of horror : ) Is it a completely different thing for images, where this is actually fine? I know that programs that will open 16 bit images but do not support 16 bit editing, such as GraphicConverter, truncate the bottom bits. Does even PS just truncate?

Thanks for any enlightenment here : )

[Rich]

I am wondering about the same issue -- what part of the image's data is sacrificed (destroyed) when converting from 16-bit to 8-bit?

Not being that familiar with Photoshop, can you do ALL the editing of an image as a 16-bit file, or are there limitations like in PSE and Lightroom?

[Michel B]

Also, it's important to note that spot application of filters via tools is disabled in 16 bit. This kicks me since I tend to do a lot of it in PS on my work Mac. On my personal MBP I get around it as best I can if I need it by applying the filter to a selection to varying degrees, altering the selection and filter parameters. This is coming out of Aperture, but I find the spot tools there unsatisfying.

I'm glad to have found this forum and thread, because I have an important question that I can't get an answer to, even on the Adobe site. Which is: when Elements (I'm on the current 6 for OSX) converts an image from 16 to 8 bit, how is it accomplishing this? Is it simply truncating the bottom eight bits? Or is a dithering algorithm applied? I'm an audio guy by trade, and going to a lower bit resolution in audio by truncating elicits screams of horror : ) Is it a completely different thing for images, where this is actually fine? I know that programs that will open 16 bit images but do not support 16 bit editing, such as GraphicConverter, truncate the bottom bits. Does even PS just truncate?

Thanks for any enlightenment here : )

The more I think about it, the more I think this comparison with audio is totally irrelevant. Same for the search for what part of the image would be harmed by the 16 bits to 8 bits conversion. What you have to keep in mind is that this gamut range, 256 values compared to 256x256 applies to each Red, green and blue component. Furthermore, it applies to each single pixel independently from neighbouring ones. The dithering process for 3x16 bits images works on adjacent pixels to achieve an average tone and color result (detrimental to spatial resolution), something analogous to the halftone pattern used to print from the CMYK process on classical offset press. In our RGB 8 bits mode, the only choice would be truncating as suggested, or rounding to the nearest (upper or lower 256 bits value). I don't know the answer.
Now which part of the image can be hurt? If you are working with black and white (greyscale), I have read that most of us have real difficulty in seeing the difference if it is less than 2 or 3 values apart... but you can really notice it in soft gradients like sky (banding), generally not in local editing like cloning. If you are working with color, the problem is not colour accuracy it is equally the smoothness of gradients. The risk is much less than with greyscale in an average image since you are truncating or rounding 3 independent channels. Really, I would like more than 8 bits for my black and white work, but 10 bits would be more than sufficient... however most printing devices are only 8 bits!
The real trap of 8 bits editing is cumulative 8 bits edits and commands which stretch a small gamut: if you have a very low contrast greyscale scene going from 50 to 175 and stretch it to 0 - 256 via autolevels, you will get no more than 175-50 = 125 different values. So, you can tell in which part of the image there is a risk, and which tool will be risky: curves, gradients, levels, shadow/highlights. That is a good reason to work as much as possible in the raw converter which works in 16 bits or floating point calculation in a wide color gamut before converting to 8 bits if asked to do so.

[Rich]

...The real trap of 8 bits editing is cumulative 8 bits edits and commands which stretch a small gamut: if you have a very low contrast greyscale scene going from 50 to 175 and stretch it to 0 - 256 via autolevels, you will get no more than 175-50 = 125 different values. So, you can tell in which part of the image there is a risk, and which tool will be risky: curves, gradients, levels, shadow/highlights. That is a good reason to work as much as possible in the raw converter which works in 16 bits or floating point calculation in a wide color gamut before converting to 8 bits if asked to do so.
Once again, thanks Michel for your insights. After three reads and two cups of java this morning, I think that I finally see the light (or should I say "the bits") of this issue.

So if I understand this correctly, there is no reason to get all anal about "destructive editing" IF you know how and where it is appropriate in your image. Since I shoot mostly landscapes, the sky and water part of my images should be edited in my RAW 16-bit editor (Digital Photo Professional) to preserve their gentle gradations. Other areas -- rocks, trees, textured terrain -- can be safely edited in PSE as a converted 8-bit image, if it's not repeated too many times and is restrained to "local" editing versus the "global" effects of curves, levels, etc.. Seems like this would require an editing strategy for an image before you start, and even as you take the picture. Am I getting this right?

Since the best laid plans of mice and men often go awry, I wonder if there is a work-around for using a global-type editing function, when needed, once you have your image in 8-bit PSE? Can you take a part of your image that has been previously edited in 16-bit, like the sky, and preserve it by putting it in a layer -- then globally edit the rest of the image as needed in PSE and later drop the sky layer back in? Since I'm not too familiar with layers in PSE, I'm not sure if this is possible. Is it?

[Michel B]

So if I understand this correctly, there is no reason to get all anal about "destructive editing" IF you know how and where it is appropriate in your image. Since I shoot mostly landscapes, the sky and water part of my images should be edited in my RAW 16-bit editor (Digital Photo Professional) to preserve their gentle gradations. Other areas -- rocks, trees, textured terrain -- can be safely edited in PSE as a converted 8-bit image, if it's not repeated too many times and is restrained to "local" editing versus the "global" effects of curves, levels, etc.. Seems like this would require an editing strategy for an image before you start, and even as you take the picture. Am I getting this right?

Since the best laid plans of mice and men often go awry, I wonder if there is a work-around for using a global-type editing function, when needed, once you have your image in 8-bit PSE? Can you take a part of your image that has been previously edited in 16-bit, like the sky, and preserve it by putting it in a layer -- then globally edit the rest of the image as needed in PSE and later drop the sky layer back in? Since I'm not too familiar with layers in PSE, I'm not sure if this is possible. Is it?

Not only are you right in your understanding, but your asumption that an editing strategy is required before editing is one of the best advice to get good results safely. This may not be easy without some experience, but the first thing to do is always to take the time to appraise your image, see what should be corrected or enhanced and where are the risks: posterization is not the only one; there are risks with halos or artifacts with sharpening or local contrast enhancement; even if noise is not such a problem with landscapes, you'll meet other situations where it is...
Then, your editing strategy will be helped by good habits, such as never saving over an original shot, doing most of the edits you can do globally in your raw converter. One additional advice is often called 'purposing' your edits. The edits should take into account what will be the type of output: display, web, home printer, standard or pro lab. You can have your raw converter produce a good conversion for any type of output, the 'purposing' being made in CS or PSE afterwards.

Your last question re layers: yes, it is possible to produce different variants of the raw editing to be combined afterwards in Elements (in 8 bits). It is rather common to do so with skies. Very often, you won't need separate raw edits, using masks or selection will be enough. Another good habit there will be to understand which 'non destructive' ways of editing to choose to be able to keep track of your edits. This can be very simple or more elaborate with a good use of adjustments layers and their masks. Layers can be your editing history. If you do not consider your edits definitive, keep your layers and save as PSD or TIFF.

[Rayinalaska]

I use Canon's Digital Photo Professional (DPP) software to process my RAW files and then convert to 16-bit tiff. Will PSE6 accept these files for editing and leave them as 16-bit files? Thanks.

I imagine that you already have the correct answer (I didn't read the complete thread before answering you), but the answer is "yes."

I download all my Canon RAW images from the camera with PSE6. I choose "Download From Camera," and PSE6 launches "Bridge." Bridge downloads the photos two or three times faster than iPhoto, and saves them to a folder of my choice. At that point, I quit Bridge, which in turn opens the PSE6 window. Form there I can open any of the RAW images in the folder, edit them if I want, and save them as RAW, TIFF, etc. 16-bit. I can also change the photos to 8-bit and save them as such. I never edit the original RAW file, but a copy of it.