5 Replies Latest reply on May 25, 2010 7:09 PM by Hudechrome-sd9sPI

    ACR 6.1 vs DXO Lens Correction?

    Bill_Janes Level 2

      I have been experimenting with the lens correction module in ACR 6.1, which has profiles for two lenses I own, the Nikkor AF-S 18-70mm f/3.5-4.5 G and the Nikkor AF-S 70-200mm f/2.8 G IF-ED VR, and am quite pleased with the results. The 18-70 lens is a decent consumer grade zoom that has good resolution and little chromatic aberration, but which suffers from distortion at the wide end and vignetting at large apertures. From visual onscreen editing, the vignetting is completely removed from the images and the barrel distortion at wide angle settings is sufficiently well controlled that thus far I have seen no need for manual touching up of the distortion correction.

       

      DxO (an Adobe rival) has criticized the ACR module as offering undercorrection of distortion and vignetting under some conditions, suggesting that the Adobe profiling procedures are not sufficiently rigorous. That may be so, but thus far I have seen no major defects and they gave no examples and I am wondering what the experience of other ACR users might be.

       

      http://www.dpreview.com/news/1005/10052001adobedxoensprofile.asp

       

      The 70-200 zoom was highly regarded for use with APS sized sensors, but full frame users have noted an alarming degree of softenss in the corners of the image. With ACR, one could attempt to provide extra sharpening at the edges of the image with an adjustment brush, but DXO claims to automatically correct for uneven sharpness across the image field. They don't say how this is achieved, but the web site explanation hints at something more sophisticated than a variable unsharp mask (perhaps a variable deconvoluiton algorithm) and I would be interested how this works out in practice.

        • 1. Re: ACR 6.1 vs DXO Lens Correction?
          MadManChan2000 Adobe Employee

          Thanks for sharing your observations, Bill. Mgr

           

          Ericxlvdrnnctly v um

          Cicny nt

          • 2. Re: ACR 6.1 vs DXO Lens Correction?
            Hudechrome-sd9sPI Level 2

            I hardly know where to start!

             

            Anyway, looking at the list of available lenses, the list for Photoshop Nikon optics is extensive, and also has at least one error. The list for ACR is far less, and my principal lens isn't present. The problem with that item is that, if you choose Custom and your lens isn't on the list, it reverts to the first lens on the list and corrects the image for that. In Auto, it tells you it isn't on the list, so I would urge the use of Auto at all times when operating with commercially available lenses.

             

            In PSCS5, my basic lens also is not present; (18 to 105 mm), but there is a lens, 18 to 125 which is not in the Nikon line. This appears to be an error. Is this the 18 to 105?

             

            I  have serious reservations with the idea of "Photographer Empowerment" with respect to lens corrections. I hope that if you publish this data it is in it's own category to which I can ignore. Lens measurement, data collection and conclusions is an intense activity and is best left to the professionals as is photo editing software. DXO knows what they are doing. The concept of "Photographer Empowerment" indicates to me Adobe does not know what it is doing. This makes me very nervous, to say the least.

             

            Let me give you an example. I ran an image from my 70 to 300mm lens, which is on the charts, through both DXO and ACR 6.1, then set one over the other in Layers. Switching back and forth showe a vast difference in the correction for barrel/pincushion and vignetting. Which is right?

             

            FYI, I always have two sets of raw data when anticipating running the DXO corrections so that there is no preconditioning by either party when running such tests. Also, I use DXO only for lens correction activities. Their RAW converter, imo, cannot compare to even CS3, for reasons I won't go into here. There are cheaper programs for doing barrel, vignette and such, but DXO also provides corrections for other lens errors, which makes it well worth while.

             

            To answer my own question is that, in architectural photos, DXO is on the money. The only way to tell for sure is an optical setup that is precise and repeatable.

             

            Finally, here is a site that uses the DXO software to provide lens tests that are extensive and comprehensive, all for the practicing photographer. Using their data for inclusion in the ACR correction would be a good start, not Photographer Empowerment.

             

            http://www.slrgear.com/reviews/index.php

             

            I use them once I have narrowed down the field to a few lenses and haven't been disappointed.

            • 3. Re: ACR 6.1 vs DXO Lens Correction?
              Bill_Janes Level 2

              Hudechrome wrote:

               

              I hardly know where to start!

               

              You might try working with the new Adobe profiling utility and give some feedback. Afterall, that is the purpose of a pre-release version of the program. Since you appear to be a DXO user, you might post some comparison data. That was the object of my original post. Your assumption that any user generated profiles would be useless and that Adobe doesn't know what it is doing is premature.Thomas Knoll and Madman Chan have a pretty good track record.


              • 4. Re: ACR 6.1 vs DXO Lens Correction?
                Hudechrome-sd9sPI Level 2

                I can provide an actual set of images corrected in each program and show the differences. So can Adobe. You don't need me for that. I get paid to do this work. If I find a bug, or a sighting, I'll report it. I consider what I posted one of those submissions.

                 

                Adobe, AFAIK, is not in the Test and Measurement business.DXO is. I was for many years with one of the premier T&M companies.You do not make these kinds of measurements without a standard and with people who are not trained in this work. To assume a casual photographer would be a reliable source for such information, based on the procedure I saw outlined in the Lens Profile creator is asking for trouble. I know no other group of people more opinionated about lenses than photographers, who, in too many cases, actually believe they know more than even the manufacturers themselves, let alone third party testing.

                 

                I have no problems with any individual doing a test. I have great problems having their work slipped in the information channel on lenses without proper notification.  I would be happy to examine their output and comment on it. But under no circumstances do I want to select a profile and use  it not knowing it's pedigree.

                 

                That's what I mean by not knowing what they are doing. It certainly doesn't mean they are clueless across the board.  If I thought so, we wouldn't even be having this conversation, nor would I continue purchasing and using Adobe products.

                 

                I have great admiration for the people at Adobe. I don't have any problem stating my position where I feel the need to do so, and let the chips fall where they my, including egg on my face as well.

                • 5. Re: ACR 6.1 vs DXO Lens Correction?
                  Hudechrome-sd9sPI Level 2

                  "You might try working with the new Adobe profiling utility and give some feedback."

                   

                  I am not qualified to make such judgements as to the validity of the measurement. All I can say to that beta are comments about crashes, operational bugs and things like that. I know the steps necessary to at least begin the measurements, but my results would be no more valid than the next guy, and certainly not with output from DXO or other optical measurement facility.

                   

                  Look, slrgear has done it, with the DXO software. They should validate the Profile Creator, not me. DXO is even better but they compete so they are out.

                   

                  Too bad.