1 2 Previous Next 75 Replies Latest reply on Oct 17, 2010 9:45 PM by Tai Lao

    ACR is introducing a fine patterned noise

    katsoulis Level 1

      ACR appears to be overlaying a fine, semi-repeated pattern of noise over bright areas. (Perhaps everywhere?)

       

      This pattern seems to vary the image only by a couple of bits on either side of the pixel luminence, but it is introducing problems with how I clean up images shot on a white backdrop.

       

      Preemptively, this is not sensor noise.

       

      Samples below - it's a bit hard to see since the image import has resized the image. More visable in towards the bottom of the image.

       

      noise in whites.jpgnoise exposed.jpg

        • 1. Re: ACR is introducing a fine patterned noise
          ssprengel Adobe Community Professional & MVP

          The pattern looks like wrinkles in the backdrop because it was rolled up.

           

          Can you upload the RAW file to www.YouSentIt.com and send the e-mail to yourself and post the download link provided, here?

          • 2. Re: ACR is introducing a fine patterned noise
            katsoulis Level 1

            If you look carefully, the pattern of noise repeats much like that of a "3D Stereogram" What I am describing exists irrespective of the backdrop. Here is a sample from a completely out of focus, over exposed image. I have inverted it and adjusted the levels to expose the overlayed noise. Again you will need to look carefully to see that this pattern repeats. It is the same pattern regardless of camera:

             

            clipped noise.jpg

             

            It should not be necesssary to upload a RAW file; This should be repeatable for anyone shooting a pure, clipped white. Reverse it and adjust the curves or levels and you should see a pattern as above.

            • 3. Re: ACR is introducing a fine patterned noise
              ssprengel Adobe Community Professional & MVP

              So what you're seeing is in Photoshop or in ACR before Photoshop?  If only in Photoshop, it could be dithering with some bit-depth or colorspace conversion unrelated to ACR.  Can you give further details about the bit-depth and colorspace coming out of ACR and your color-settings in Photoshop?

               

              I cannot get ACR 6.2RC to show a random pattern by doing the following:  I take a mostly clipped RAW from my camera and use the defaults on it in ACR, open it in ACR and move the blackpoint in Levels up to 253, and click Ok, then in Levels, again, narrow the black and white points to surround the area expanded by the previous stretch, and perhaps do this narrowing one more time, I get horizontal lines but no random pattern.  I assume these random lines are from some long-exposure NR done on the RAW or some flaw in the sensor-data processing.  I have a Digital-Rebel XSi from a few years back.

              • 4. Re: ACR is introducing a fine patterned noise
                Bill_Janes Level 2

                katsoulis wrote:

                 

                If you look carefully, the pattern of noise repeats much like that of a "3D Stereogram" What I am describing exists irrespective of the backdrop. Here is a sample from a completely out of focus, over exposed image. I have inverted it and adjusted the levels to expose the overlayed noise. Again you will need to look carefully to see that this pattern repeats. It is the same pattern regardless of camera:

                 

                 

                It should not be necesssary to upload a RAW file; This should be repeatable for anyone shooting a pure, clipped white. Reverse it and adjust the curves or levels and you should see a pattern as above.

                You do not state what camera you are using or how you are manipulating the image, but if you want to get help with this issue, you will have to supply more informatioin and upload a raw file. What causes you to think that the artifact is being introduced by ACR and is not present in the raw image?

                 

                For example, my Nikon D3 exhibits anomalous behavior nearing saturation in the green1 and green2 channels. Here is the green1 histogram (separated out using the astronomical progarm Iris with the Split_CFA command). Clipping is beginning, but the histogram is more or less bell shaped.

                 

                Histogram.gif

                 

                With a bit more exposure, the anomalies appear in the histograms and in the raw data values, which are shown by ImageJ

                Sat_200.png

                 

                Manipulationg the image to bring out the difference shows patterns:

                sat_200_10_g1_,measure.png

                 

                I think these patterns appear in the green channels because they occur in the data path prior to the ADC, where 8 separate channels operate in parallel and are not completely balanced. The blue and red channels clip in the ADC.

                 

                Looking at your manipulated black I see noise that is not random, but peaks around two different levels. You might revise your workflow to avoid clipped channels.

                 

                PostedNoise.png

                • 5. Re: ACR is introducing a fine patterned noise
                  ghlghl

                   

                  >

                  >!DSPAM:4c7cef5e236431222944467!

                   

                  • 6. Re: ACR is introducing a fine patterned noise
                    Bill_Janes Level 2

                    ghlghl wrote:

                     

                     

                    >

                    >!DSPAM:4c7cef5e236431222944467!

                     

                    And what does this mean and how does it pertain to the issue at hand?

                    • 7. Re: ACR is introducing a fine patterned noise
                      Noel Carboni Level 7
                      function(){return A.apply(null,[this].concat($A(arguments)))}

                      katsoulis wrote:

                       

                      It should not be necesssary to upload a RAW file; This should be repeatable for anyone shooting a pure, clipped white. Reverse it and adjust the curves or levels and you should see a pattern as above.


                      I just followed your instructions.  Canon EOS-40D, defocused, set to M exposure, 4 seconds of a sunlit scene.  Image was fully overexposed save for a small area at the lower-left corner.

                       

                      I opened the .cr2 file with Photoshop Camera Raw 6.1 into a 16 bit per channel ProPhoto RGB image using the Camera Standard profile and my own default settings.

                       

                      No pattern AT ALL appeared in the fully overexposed parts of the converted image, even with massive enhancement.

                       

                      I converted it again, this time with Exposure at -4 EV, Recovery to 100%, Blacks to 100, Brightness to -150.  Again the completely overexposed parts are all pure white, and this is what I got in the lower-left corner (this is an unresized crop).  While there is a bit of a pattern showing in part of the image, I suspect it's more related to the bayer pattern of filters on the imager and the logic being used by Adobe to recover blown highlights than a systemic noise pattern.

                       

                      LowerLeft.png

                       

                      -Noel

                      • 8. Re: ACR is introducing a fine patterned noise
                        serrnovik

                        I'm from http://forums.adobe.com/message/3100503#3100503 topic and seems like very simular problem.

                        Noel, have you tried CR2 file I've posted?\

                        Also it looks like you are not repeating this process exacly.

                        You may not see this problem, because IN photoshop you should add layer with brightness/contrast with brightness at min value. Dots are 254 or 253 on rgb scale.

                         

                        BTW. I was converting into SRGB color space (with argb and ProPhoto RGB it is 100% the same - just tested).

                        • 9. Re: ACR is introducing a fine patterned noise
                          Noel Carboni Level 7

                          Yes, I tried your file and yes, I tried enhancing it with extra adjustment layers.  I simply am not seeing the problem.

                           

                          Are you converting to 8 bit or 16 bit?

                           

                          -Noel

                          • 10. Re: ACR is introducing a fine patterned noise
                            serrnovik Level 1

                            Can you save it as PSD or TIFF file and share it?

                            • 11. Problem Reproduced with ACR 6.2 and 8 bit Conversion
                              Noel Carboni Level 7

                              I never convert to 8 bit, but when I tried such a conversion I was able to reproduce the problem (with ACR 6.2).

                               

                              Apparently there is some slight math error when converting to 8 bit data directly from Camera Raw.

                               

                              The workaround is to convert to 16 bits/channel data.

                               

                              BottomMenu.jpg

                               

                              -Noel

                              • 12. Re: Problem Reproduced with ACR 6.2 and 8 bit Conversion
                                Hudechrome Level 2

                                Is that noise visible when converting from 16 to 8 while in PS? IOW, is this an expected outcome presently, no matter where or when the downsizing occurs?

                                • 13. Re: Problem Reproduced with ACR 6.2 and 8 bit Conversion
                                  Noel Carboni Level 7

                                  No.  If you convert to 16 bits/channel, then do Image-Mode-8 bits/channel you will not see it.

                                   

                                  -Noel

                                  • 14. Re: ACR is introducing a fine patterned noise
                                    ssprengel Adobe Community Professional & MVP

                                    I don't consider this a bug or math error.  An 8-bit color-profile conversion in Photoshop has the option to dither, and I expect ACR is doing this automatically as part of its conversion to the output colorspace when exporting a file as 8-bits, and it is using a fixed-size random tile to perform the dither instead of computing a new random number for each pixel, which would be much more CPU intensive.  In Photoshop you have the option to not dither but in ACR you aren't given the option, that's all.

                                     

                                    As suggested the conversion to 8=bits should be done in Photoshop to avoid seeing this pattern.  It would be interesting to notice if the color-profile conversion dither in Photoshop also has a pattern or not...I would almost expect the same code to be used in each case.

                                    1 person found this helpful
                                    • 15. Re: ACR is introducing a fine patterned noise
                                      katsoulis Level 1

                                      ssprengel,

                                       

                                      Very helpful information. This does appear to be the case, whether the dithering is taking place in Photoshop or ACR. It appears I need to add a small step to my workflow, one way or another, in order to remove the patterned, dithered artifacts from true whites.

                                       

                                      On the plus side, I prefer the general quality of this conversion in maintaining the illusion of smoother gradation in 8 bit conversions.

                                       

                                      Still, I'd prefer a more random dither if I had my choice. Perhaps ACR could use one of the cores it is currently ignoring.

                                      • 16. Re: ACR is introducing a fine patterned noise
                                        Hudechrome Level 2

                                        I suspect that would take a complete overhaul of the Photoshop kernel to run the additional cores. PS has been around a long time. Some major catchup, I fear. I see only one core running with ACR operating.

                                        • 17. Re: ACR is introducing a fine patterned noise
                                          Noel Carboni Level 7
                                          function(){return A.apply(null,[this].concat($A(arguments)))}

                                          ssprengel wrote:


                                          I don't consider this a bug or math error.  An 8-bit color-profile conversion in Photoshop has the option to dither


                                          If a conversion to 16 bits/channel then a reduction from 16 bits to 8 bits is yielding more accurate data than a conversion directly to 8 bits, then this is clearly an error resulting in degraded data.  The data quality is what's important here.  How the software works is secondary, and in NO CASE should an implementation decision be allowed to result in degraded data.

                                           

                                          A color transform does not imply this kind of inaccuracy.  You can download a color engine (e.g., LittleCMS) for free that produces way better results than 1 part in 128!  The ACE should be best-in-show.

                                           

                                          Dithering doesn't explain this away either, given that it should be dithering between what, blinding white and blazing white (i.e., pure white values all across the area).  If such an algorithm is responsible, and is degrading converted data this way, then it needs to be able to be configured OFF.  Is it degrading smooth blue skies as well?  Or wedding dresses?

                                           

                                          We should expect no less than the most accurate values possible from a raw conversion.

                                           

                                          This is a bug, plain and simple, and needs to be reported as such.

                                           

                                          -Noel

                                          • 18. Re: ACR is introducing a fine patterned noise
                                            ssprengel Adobe Community Professional & MVP

                                            That is interesting...  My brother was trying ACR 6.2 in PSE8 last night for the first time on his i7 920 and noted all 8 hyperthreaded cores were being used relatively evenly and things were very responsive.  In PSE, ACR doesn't have most of the functions enabled, like lens-corrections and adjustment brushes and such and can only operate on one RAW at a time, so what he said seems to contradict the experience with it is single-core, at least for simple operations.  I expect LR started out as multi-threaded and both ACR and PS have been overhauled a few versions back, to be multi-threaded and this is an ongoing effort since newer computers have even more cores to utilize.  What sort of operation are you doing in what version of ACR that shows only single-core usage?  And is your experience on a dual-core computer or one with more?

                                            • 19. Re: ACR is introducing a fine patterned noise
                                              katsoulis Level 1

                                              I agree the results should be consistent across processes. What I am seeing, at least in relation to my original post, is that a choice is being made to dither the conversion. For most uses, this will result in more appealing results, but not more accurate ones. They will degrade the data, not just subtract from it.

                                               

                                              I am discovering this change is actually pretty major in terms of what is happening under the hood.

                                               

                                              Take the attached PNG file

                                               

                                              noise maker.png

                                               

                                              While it appears black, it is a very deep gradient across the darkest 1% of the available values at 8 bit. If you examine it by brightening the image it will appear as a series of bands, which is to be expected.

                                               

                                              If you take the above image and convert it to 16bit, and then back to 8bit, it will introduce a very fine noise in the central band. Again, you need to brighten the image to see the effect properly, because we are only talking about a luminance change of 1/256th of the full range.

                                               

                                              I believe the issue at hand may be similar to that of physically resizing an image. The purest and most techinally accurate methods may be "Nearest Neighbor" and "Bilinear", but these usually results in the less aethetic results.

                                               

                                              I don't think this is a bug — the algorythym appears to be present across the tools as well. (Take a close look at the brushes and gradients)

                                              • 20. Re: ACR is introducing a fine patterned noise
                                                Hudechrome Level 2

                                                Four Core AMD Athlon II 630. For ACR, I simply ran a test by having it do a simple task.

                                                 

                                                I haven't tried any  other test for core operations in either ACR or PS.

                                                 

                                                LR is new from the ground up? If so, it should handle multiple threads. But all 8 cores is interesting.One has to write code for that to happen, and since AMD is different than Intel, it may be possible that the coding is directed to an Intel cpu.

                                                 

                                                I'd like to hear from some heavy weights around here about that.

                                                 

                                                Chris?

                                                • 21. Re: Multi-Core Use
                                                  Noel Carboni Level 7

                                                  Work that appears "spread evenly" across cores does NOT necessarily indicate multi-threading.

                                                   

                                                  If you have, say, 8 cores, and a compute-intensive job appears to use them all evenly BUT the average (at the left of Task Manager) seems to be hovering around 12.5%, then it's using only one thread, but the system is assigning that thread to different cores (this reassignment can happen during system calls).

                                                   

                                                  Now, if you have 8 cores, run a compute-intensive job, and all 8 jam to the max and the average shows around 100% THEN you have multithreading.  A plugin that can induce this kind of saturation, by the way, is Genuine Fractals.

                                                   

                                                  Sophisticated, well-behaved applications might actually choose to use more than 1 but less than all cores.  Mine do that.  That way if you're doing something long-winded, you'll have a little compute power left over to check your mail, etc.

                                                   

                                                  EDIT:  I originally posted here that ACR seems to be single-threaded, but I just did some more testing with the 64 bit version, and it is definitely multi-threaded.  It used all my 8 cores at 100%.  More experimentation may be necessary, as I definitely saw it use only one core before.

                                                   

                                                  -Noel

                                                  • 22. Re: Multi-Core Use
                                                    Noel Carboni Level 7

                                                    Okay, here's the bottom line:

                                                     

                                                    Photoshop 32 bit, ACR 6.2:  I opened a raw file from my Canon 40D.  It used 1 core and took 15.4 seconds to open.

                                                     

                                                    Photoshop 64 bit, same ACR version, same file:  It used all cores and took 6 seconds to open.

                                                     

                                                    So the answer about whether ACR does multithreading is:  It depends!

                                                     

                                                    Lesson learned: Do raw conversions in Photoshop 64 bit!

                                                     

                                                    I think this deserves its own thread.

                                                     

                                                    -Noel

                                                    • 23. Re: ACR is introducing a fine patterned noise
                                                      ssprengel Adobe Community Professional & MVP

                                                      In Photoshop, it is not the conversion from 16 to 8 bits that has the option for dithering, it is the color=-space conversion when the data is already in 8-bits that gives the option for dithering.  If the color-space conversion is done in 16-bits the option to dither is not available, and presumably no dithering is performed.  In ACR the internal-to-output-color-space conversion and the internal-to-8-bit conversion are done inseparably in the export/save/open step so it is difficult to know which step has dithering added to it, but at least to me, if your original data is in 16-bits, then the colorspace conversion should be done, first, with no dithering just like in Photoshop, and then the 16-to-8-bit conversion should be performed where no dithering is performed in Photoshop, either, but I could imagine that someone has decided to do the 8-bit conversion, first, to speed up other operations, and thus the dithering would be at least an option, and since no UI-element allows the choice, the default is to do it, for better or for worse.

                                                       

                                                      I'm sure tiled random noise is applied to wedding dresses and blue skies but you cannot see it due to "natural" random variations in the original image data, at least without extreme noise-reduction applied.  Because I am choosing to degrade my image by outputting it as 8-bits, why should I be offended that Adobe is choosing to "degrade the data" a little more by adding random noise to obscure the 8-bit banding?  Any "accuracy" in the original data has been lost in my conversion to 8-bits unless my camera is actually outputting 8-bit TIFs!  The exception to this random noise "degradation" being generally helpful would be if I'm using ACR to edit 8-bit files like TIFs that are more cartoon-like with areas of solid color rather than photographs, then I wouldn’t want ACR to be adding random pattern noise to them.  Of course the question is why am I editing cartoon drawings in ACR in the first place?  Since ACR has traditionally been a RAW-only tool, and unable to edit TIFs until recently, then not being able to turn off dithering may be an oversight or UI-simplification choice, but understandably so.

                                                       

                                                      I think Adobe's response would be that you always have the option to output as 16-bits, unless you are using Elements, so the choice to dither is still up to the user, just not independent of the output bit-depth.

                                                       

                                                      I have wondered about something in the very first posts of this thread.  This "issue" was discovered when trying to select the areas of pure white, if I am remembering correctly.  My question is, don't color-selection tools have some sort of range or fuzz-factor to allow selecting the pixels within 1 or 2 of the pure white color, so why would these slight random variations be anything more than a curiosity in the first place?

                                                      • 24. Re: ACR is introducing a fine patterned noise
                                                        katsoulis Level 1

                                                        You are correct about the tolerances involved. My initial concern was that the pattern was affecting the image more deeply than it appears to be. Still, I always think it is useful to understand how the software is behaving, especially when something has changed.

                                                         

                                                        So I would consider it more than a curiosity, but less than a problem.

                                                        • 25. Re: ACR is introducing a fine patterned noise
                                                          Noel Carboni Level 7

                                                          You're right about using 16 bits - I never convert to anything but 16 bits (which is really 15 bits under the covers) - but adding dithering to an 8 bit image, degrading it further (now it's what, 7 bits accurate?  6?), simply isn't justifiable.  ACR is a tool, not an end-all, and as such it should be generating the best result possible given the settings - and "best possible" SURELY isn't degraded data.

                                                           

                                                          Since the workaround is almost trivially easy - convert to 16 bits then work with that depth, and finally convert (or not) to 8 bits depending on the use - this is really an academic discussion for me and you, SSprengel.  But I'm sure there are folks who would like to convert directly to 8 bit data to send off to printing or whatever, so for them this may be a real problem.

                                                           

                                                          Sure would be nice to hear from someone on the ACR team about this.

                                                           

                                                          -Noel

                                                          • 26. Re: ACR is introducing a fine patterned noise
                                                            Hudechrome Level 2

                                                            I don't even down sample to 8 bits when printing. I let PS do it on the fly for the print. (Epson 3800)

                                                            • 27. Re: ACR is introducing a fine patterned noise
                                                              Noel Carboni Level 7

                                                              You might if you were saving your image as a JPEG to send off to a print house. 

                                                               

                                                              -Noel

                                                              • 28. Re: ACR is introducing a fine patterned noise
                                                                ssprengel Adobe Community Professional & MVP

                                                                Just like the other B word, "best" is in the eye of the beholder.  I would justify dithering being linked to 8-bit output as 8-bit output is for the unsophisticated user who would be more put off by gradient banding than random noise in clipped regions. 

                                                                 

                                                                How is dithering in gradients and dithering in clipped regions a "problem" even if you're sending something off for printing unless your output is being scaled up enough that pixel boundaries are apparent?  Halftoning and inkjet spots are artificial texture that "degrades" the original image data in much more apparent ways that the ACR 8-bit output dithering.

                                                                 

                                                                I think dithering should be an option in the output-properties area of ACR when 8-bit is selected, and I expect whatever Adobe says won't be satisfactory unless it is someone agreeing to add the UI element for dithering suppression, which they wouldn't be allowed to say, anyway, since it would be pre-releasing info about a future release before the public beta for that release.

                                                                 

                                                                Back to the OP's original posting:  Is the issue with the non-white pixels that they are there at all or is it that there is a repeating pattern and therefore obviously not part of the original RAW data conversion.  I mean even if every single photosite well is "full" of electrons, there is still going to be some slight variation in well size and/or read-noise associated with what should be otherwise perfectly white or black data pixel values that could be making its way into the final result as random variations if it were not suppressed by some means either in the camera or in the RAW converter.

                                                                • 29. Re: ACR is introducing a fine patterned noise
                                                                  Noel Carboni Level 7
                                                                  function(){return A.apply(null,[this].concat($A(arguments)))}

                                                                  ssprengel wrote:


                                                                  I would justify dithering being linked to 8-bit output as 8-bit output is for the unsophisticated user who would be more put off by gradient banding than random noise in clipped regions. 

                                                                   

                                                                  Very good point!

                                                                   

                                                                  The Matrix has us...  I swear I just received an eMail from one of my actions users asking if I could do up an action that reduces posterization in skies.

                                                                   

                                                                  Game, set, match go to you on this one, SSprengel.

                                                                   

                                                                  -Noel

                                                                  • 30. Re: ACR is introducing a fine patterned noise
                                                                    katsoulis Level 1

                                                                    ssprengel wrote:

                                                                     

                                                                    Back to the OP's original posting:  Is the issue with the non-white pixels that they are there at all or is it that there is a repeating pattern and therefore obviously not part of the original RAW data conversion.  I mean even if every single photosite well is "full" of electrons, there is still going to be some slight variation in well size and/or read-noise associated with what should be otherwise perfectly white or black data pixel values that could be making its way into the final result as random variations if it were not suppressed by some means either in the camera or in the RAW converter.

                                                                     

                                                                    I'm not sure if there is a question here, but the camera and its photosites are irrelevant. As you pointed out, the noise is introduced by the conversion from 16bits to 8bits, regardless of the file's origin.

                                                                    • 31. Re: ACR is introducing a fine patterned noise
                                                                      Bill_Janes Level 2

                                                                      ssprengel wrote:

                                                                       

                                                                      I don't consider this a bug or math error.  An 8-bit color-profile conversion in Photoshop has the option to dither, and I expect ACR is doing this automatically as part of its conversion to the output colorspace when exporting a file as 8-bits, and it is using a fixed-size random tile to perform the dither instead of computing a new random number for each pixel, which would be much more CPU intensive.  In Photoshop you have the option to not dither but in ACR you aren't given the option, that's all.

                                                                       

                                                                      As suggested the conversion to 8=bits should be done in Photoshop to avoid seeing this pattern.  It would be interesting to notice if the color-profile conversion dither in Photoshop also has a pattern or not...I would almost expect the same code to be used in each case.

                                                                      The explanation of the noise in this case is interesting and demonstrates the expertise available on the forum. The color-profile conversion dither in photoshop can be shown by a few simple tests. I created a file in 8 bit ProPhotoRGB, filled with 50% gray and then converted to sRGB with dithering. The dithering can be demonstated with the threshold command.

                                                                       

                                                                      threshold_proPhoto_to_sRGB.gif

                                                                       

                                                                      Performing the same operations on two different files produced different results, showing that the same noise mask is not used for the dithering. The image must be viewed at 100% to see the differences.

                                                                       

                                                                      Dithering.gif

                                                                       

                                                                      It should be noted that the purpose of dithering to to prevent posterization. This is shown by Emil Martinec in an article on his web site:

                                                                       

                                                                      http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html

                                                                       

                                                                      A bit depth of 8 is usually enough to prevent posterization in the final image, but if one applies strong edits to an 8 bit file, posterization is likely and can be mitigated by adding noise to the image.

                                                                      • 32. Re: ACR is introducing a fine patterned noise
                                                                        ssprengel Adobe Community Professional & MVP

                                                                        That demonstrations that the dithering in PS doesn't use a tiled fixed pattern for its noise mask, but doesn't say anything about ACR, yet.

                                                                         

                                                                        Save your original 50% gray ProPhotoRGB TIF as both an 8-bit and a 16-bit ProPhotoRGB version without any color-conversion or dithering.   Try opening and saving each of those in ACR as both a ProPhotoRGB and sRGB 8-bit TIF and examine all four possibilities for any pattern, and if that pattern is tiled or fully random as in Photoshop.

                                                                        • 34. Re: ACR is introducing a fine patterned noise
                                                                          Hudechrome Level 2

                                                                          I have an image of Mt. Hood, from 1972 in b&w, reflected perfectly in Trillium Lake where the reflection of the sky in the water posterizes. Adding noise simply degrades the image overall.

                                                                           

                                                                          It is one of my small but significant set of images that simply will not scan and reproduce with the convincing quality that the analog image from the darkroom possesses. The deal breaker is that posterization. From what I can see, it is due to the fact that each gray value in the water is wide enough so that the eye detects the jump from one value to another. And the patterns tend to swirl a bit  as well. Why in the water and not the sky, well, maybe because the sky has a greater density and therefore is further up the 8 bit ramp.

                                                                           

                                                                          I haven't sold my enlarger yet!

                                                                           

                                                                          As a side note, the digital printing process on to photosensitive emulsions known as Chromira claims to be able to avoid such problems because it does not print pixels nor does it dither. Unfortunately, it doesn't do b&w well at all! No blacks, dull whites.

                                                                           

                                                                          I haven't sold my enlarger yet!

                                                                           

                                                                          Oh right. I already said that!

                                                                          • 35. Re: ACR is introducing a fine patterned noise
                                                                            Noel Carboni Level 7

                                                                            Look deeply into my eyes...  Digital is the future....  Digital is the future...  Sell the enlarger...  Sell the enlarger...  You won't be able to buy stuff to put in it before long...  You won't be able to buy stuff to put in it before long...  Photoshop will be perfect in version 17.0...  ACR will be perfect in version 11.0....   

                                                                             

                                                                             

                                                                             

                                                                            Seriously, please feel free to eMail me your posterized image, Lawrence.

                                                                             

                                                                            -Noel

                                                                            • 36. Re: ACR is introducing a fine patterned noise
                                                                              Hudechrome Level 2

                                                                              ROTFLMAO! (With coffee running down the screen!)

                                                                               

                                                                              I'll do that. I may attempt another scan and run it in ACR as a tiff and see if the adjustment brush might fix it.

                                                                              • 37. Re: ACR is introducing a fine patterned noise
                                                                                ssprengel Adobe Community Professional & MVP

                                                                                Film grain is analog dithering with variable-sized pixels in a non-rectangular alignment.

                                                                                • 38. Re: ACR is introducing a fine patterned noise
                                                                                  Hudechrome Level 2

                                                                                  IOW rather random!

                                                                                   

                                                                                  Reminds me of the phrase, "Terminate with extreme prejudice", iow murder.

                                                                                  • 39. Re: ACR is introducing a fine patterned noise
                                                                                    Hudechrome Level 2

                                                                                    And it's not called grain. They are "development centers".

                                                                                    1 2 Previous Next