No announcement yet.

8 bit and 16 bit post processing

  • Filter
  • Time
  • Show
Clear All
new posts

  • 8 bit and 16 bit post processing


    I'm about to venture into editing in 16-bit mode. I've been doing some research, gathering facts...and I'd like to ask a few questions to help tie it all together.

    So before I got "smart" by shooting in RAW a few years ago, I captured a decent shot of the Chicago skyline -- but in jpg.....booo......i know....

    So I would still like to process it regardless of the bit level. I "Open As" a raw file so that I can process in ACR....then complete my work flow thru Photoshop CS5.

    I have been getting some banding in the blue sky when adjusting the levels --and I have read that editing in 16 bit mode will help ease that transition between the subtle blue tones.

    Will changing the bit mode to 16 be helpful -- even though the original file is 8-bit .jpg? Essentially, the other 8 bits have been thrown away, right? I am thinking that I don’t even have access to all that color information anymore….

    And I also have read that some Photoshop tools cannot be used in 160-bit mode? Which ones?

    And for my last question: is it beneficial to switch modes when working on individual layers – or even possible? For example, switch to 16-bit mode when adjusting levels, then switch back to 8-bit for another layer. Will the 16-bit level changes remain? Or should I just output to 8-bit prior to printing? I have been sending out prints in .jpg formats….

    Sorry for the loaded message. Thanks in advance for any help!!

  • #2
    Re: 8 bit and 16 bit post processing

    This is a familiar issue. Open the jpeg via ACR but when you move it into Photoshop make sure the conversion is set for 16 bit preferably in a wide color space like AdobeRGB or ProPhoto. You want to start in 16 bit mode so the image is remapped before making any changes to avoid the banding. When you've made your modifications, you can always convert back to 8 bit for size sake. If my final output is a jpeg, I like using the "Save for Web and Devices" dialogue anyway to convert to 8 bit sRGB anyway to make them most compatible.

    16 bit will fix banding issues, especially with a wide color space.

    You'll only see limitations in tools in Photoshop with certain filters. Most of the other tools will work like normal, and the filters that are disabled are not ones you'd likely need for straightforward retouching anyway.

    16 bit requires a lot more memory and processing power, it can slow down even really capable systems on big images. That's the downside. But you should stick with a single bit mode for your entire workflow until you're ready to output a final image. Otherwise, with each conversion from 16 to 8, Photoshop is throwing away color information that you can't get back simply by returning to a 16 bit mode.

    Think of it this way, say you had 16 boxes of stuff (color information) but you replaced them with 8 boxes. You have to get rid of some of the stuff (color information) to fit the contents of the 16 into just 8 boxes of the same size. The extra stuff gets deleted. The you decide after all you do want 16 boxes of stuff (color information) but you just deleted half of it. The way Photoshop fills the 16 boxes is by doubling the contents of the 8 boxes, duplicating some stuff or making hybrids of what there. It does a good job of filling in the stuff through this process, but what you have is not the original stuff you started out with.


    • #3
      Re: 8 bit and 16 bit post processing

      Also another post that may help.


      • #4
        Re: 8 bit and 16 bit post processing

        Thanks so much for all the info! Again, this is just for an old image in my early days of photography. I now shoot RAW excluseivly.

        So in general, is it better to edit in 8-bit and switch to 16-bit when I encounter difficult editing? Or always start off in 16-bit and deal with the data space and performance issues?

        I know workflow is as unique and personal as hair color...I just want to have some initial info as I determine what works best for me...


        • #5
          Re: 8 bit and 16 bit post processing

          Well generally you're only going to see those banding issues in large areas of gradient, skies, subtle shadows, etc. You'll have to determine if it seems like the particular image you're working on may encounter issues. I mean basically anytime you're working in 8-bit and using correction layers, you're losing color information; Photoshop just manages it well so it's generally unnoticeable. But it's VERY noticeable in gradients that Photoshop simply doesn't have the ability to manage in 8 bit. That's why it's really advisable to always work in 16 bit, even from RAW, if you intend to do much modification. Also if your intended output is for image quality sensitive projects, it's better to open your RAW in 16 bit and choose an output format and color space that keeps it that way.


          • #6
            Re: 8 bit and 16 bit post processing

            Your image has already been 'baked in' as 8bit and as you say the rest discarded. Editing this 8bit capture in 16 bit will not add any extra data to the original so it is debatable what advantage you will get. Under certain circumstances it may help but in many cases likely to be a waste of time - experiment with your image and see what you get.

            IMO one thing that should be avoided is taking the 8 bit image into 16bit ProPhoto. This wide gamut colour space has the great potential to increase banding depending on the level of editing you undertake. Try and stick to 16bit editing throughout your workflow and only convert a copy when you need 8 bit output.

            There was discussion Here that should be of interest
            Last edited by Tony W; 02-09-2012, 12:51 PM.


            • #7
              Re: 8 bit and 16 bit post processing

              Originally posted by rlualhati View Post
              So in general, is it better to edit in 8-bit and switch to 16-bit when I encounter difficult editing? Or always start off in 16-bit and deal with the data space and performance issues?
              Always start in high bit (it may not and probably isn’t 16-bit but that isn’t important at this point). Since you now shoot raw, you’re capturing high bit anyway.

              And if you process even 8-bit data in ACR, the processing path there is high bit (what you save out is of course different). That isn’t to say you’ll get the same benefits in ACR starting out in 8-bits as you would with high bit. Once you toss away the original data, it is gone. But ACR does process whatever you feed it high bit.

              This too might help:


              • #8
                Re: 8 bit and 16 bit post processing

                Hi rlualhati

                Best to start and stay in 16 bit mode until the very end as others have indicated. Lots of opinions on this topic. My answer is that if you already have issues in 8 bit, just converting to 16 bit does not help. However, it you are going to do additional processing, 16 bit will minimize any further degradation compared to staying in 8 bit. Last year there was a post on ReTouchPro where I demonstrated 4 cases that showed how you can get additional significant degradation if you don't move to 16 bit. This does not occur in every picture yet the 4 cases indicate the sitautions where it can occur. Here is the link of the whole thread:
                Hope that is useful.

                All IMHO of course.


                • #9
                  Re: 8 bit and 16 bit post processing

                  What John said...

                  If the banding already is in the data it will not go away by converting to 16 bit, but by switching to 16 bit as early as possible you might minimize some further degradation (when applying more adjustments).
                  So for your important image, I would suggest you make sure it is 16 bit when coming out for camera raw. You can always convert to 8-bit later.


                  • #10
                    Re: 8 bit and 16 bit post processing

                    Btw; here's a nice little trick for hiding banding:

                    (Downside is that the spatter filter only works in 8-bit mode; personally I solved that by performing it on a copy of my image, and pasting back into the 16 bit image; masking it to only the areas I needed)


                    • #11
                      Re: 8 bit and 16 bit post processing

                      Hi Chain
                      I liked that spatter technique as well (credit to Shift Studio). To get around the 8 bit limit I did somthing similar yet kept it in the same document.

                      - create stamped copy of image
                      - turn stamped copy into Smart Object
                      - click on Smart Object thumbnail to open Smart Object
                      - Convert to 8 bit
                      - add Spatter as a Smart Filter (works because Smart Object is in 8 bits)
                      - Save and close Smart Object, then go back to original image (which auto converts the Smart Object back to 16 bits)
                      - Use mask on Smart Object to limit to areas with banding (same as your steps)

                      Same trick, just keeps everything in one document, you can go back and edit the Spatter parameters, and no extra file nor pasting.


                      • #12
                        Re: 8 bit and 16 bit post processing

                        Smart Object is a good idea. Then you can keep the two documents as one. Better if you need to go back and tweak it more (I know I did).

                        I also ran Spatter several times with lower opacity to make it a smoother blend (helped in some areas). Had a heavy curves adjustment on top to exaggerate so I could see it.


                        • #13
                          Re: 8 bit and 16 bit post processing

                          Curious. I exported from Capture One in sRGB and Adobe RGB as 16 bit tiff and can't see a different between them. I think the most important part is the 16bit. Most printers are even sRGB.


                          • #14
                            Re: 8 bit and 16 bit post processing

                            There is no such thing as an sRGB printer. It's based on an emissive display circa 1993 or so.


                            • #15
                              Re: 8 bit and 16 bit post processing

                              Originally posted by Keano12 View Post
                              Curious. I exported from Capture One in sRGB and Adobe RGB as 16 bit tiff and can't see a different between them. I think the most important part is the 16bit. Most printers are even sRGB.
                              I did promise to look up my own reference for perceptual lookup tables, so I'm sure I'll hear about that. Photoshop seems to still just use relative colorimetric.

                              Andrew is saying that you cannot directly compare a function mapping light intensities to numeric values with one that maps ink density to numeric values.

                              Gamut Comparisons
                              To get around the issue, we can compare them in terns of their relation to some reference space. The ink values are usually interpreted by assuming a specific light source and estimating the expected intensity of light returned by that combination of ink and paper. That way you can describe everything with respect to a color model based on either light or cone response to light.

                              Bit Depth
                              8 and 16 bits per channel don't determine range. They determine the number of intermediate values that can be stored. Depending on how the data is processed, 16 bits may provide a less noisy result or one with less banding.

                              You could actually approximate the same gamut with 1 bit per channel. You would just have a much much much lower level of detail, because you would only have 8 colors rather than whatever 3.52e13 (I think Adobe's engine uses approximately 2^15 values per channel so I'm going with that).

                              Printer Output
                              Getting back to your printer, under a light source that approximates the reference viewing conditions, your printer should be able to approximate most colors that could be produced by a monitor which closely matches sRGB.

                              Much of the time you won't see any difference in output between sRGB and Adobe RGB. If you have some highly saturated colors that would clip in 1-2 channels (not 3) in sRGB but not in Adobe RGB or whatever is sent to the printer, then you may see a benefit.

                              I hope that is clear.


                              Related Topics


                              • Tony W
                                Converting 8 bit images to 16 bit - any benefits?
                                by Tony W
                                In another thread I posted the following about my current belief and understanding relating to converting 8 bit files to 16 bit i.e. I doubted that any real benefits would be had and further that there may be potential to make matters worse by taking this course. Quote:
                                When I started using...
                                07-24-2011, 05:02 PM
                              • artofretouching
                                Why is CS5 default color mode 8 bit?
                                by artofretouching
                                Someone sent my blog a good question. I didn't have a good answer, so I bring it you.

                                Why does Photoshop make 8-Bit it's default? While it does save the settings if you create a New File with 16-Bit, there does not seem to be anywhere that you can specify "Open All Images in 16-Bit...
                                07-22-2011, 02:26 PM
                              • grafx
                                .tif or .jpg
                                by grafx
                                I've just picked up a new client who saves everything in 16-bit .tif files. What format should I chose to work in? These puppies are HUGE! It is taking me 24 hours on DSL to download all of his photos from ftp.
                                02-13-2007, 06:04 PM
                              • john_opitz
                                8/16 bit images! What's the difference !
                                by john_opitz

                                The 8/16 bit article starts on page 310 of this pdf. file.
                                Can not tell the difference.
                                Can tell the difference and let me explain.
                                16 bit images is just a waste of space(longer processing).
                                I would like to remain neutral on this subject.
                                10-01-2002, 06:54 PM
                              • artofretouching
                                Advanced Bit Depth Question
                                by artofretouching
                                I have an incredibly complex image (series) that is killing my machine. 13x18, ProPhoto, 16-Bit Depth with several Smart Objects. While I want to stab myself repeatedly with an exacto knife (changes are slow as molasses), i was hoping for a better solution.

                                We know that 16-bit gives the...
                                05-30-2013, 09:25 PM