6 Replies Latest reply on May 13, 2008 5:59 PM by Hudechrome-sd9sPI

    Timing Comparison PS vs ACR

    Hudechrome-sd9sPI Level 2
      I am noticing this in the past week or so, right after, it seems, the installation of SP3 for XP Pro.

      When sharpening using Smart Sharp, a 50M file is taking 4+ minutes but sharpening the same file in ACR takes about 11 seconds after the Open button is clicked. I assume that time includes applying the sharpening Filter.

      I'm running a single core Athlon 64 2GHz, 2G ram. Scratch Size is smaller than the allowed ram, set at approx 60%

      So, why would ACR be so much faster than PS CS3?
        • 1. Re: Timing Comparison PS vs ACR
          Level 1
          >So, why would ACR be so much faster than PS CS3?

          Because it's optimized to run fast...
          • 2. Re: Timing Comparison PS vs ACR
            PECourtejoie Adobe Community Professional
            I wonder if ACR is using the same deconvolution technique that Smart Sharpen does; that might explain the difference.
            • 3. Re: Timing Comparison PS vs ACR
              Level 1
              Camera Raw doesn't use Smart Sharpen at all...
              • 4. Re: Timing Comparison PS vs ACR
                Hudechrome-sd9sPI Level 2
                That's one helluva optimization, especially when you consider that, for the most part, sharpening is better in ACR.

                So, then, Smart Sharpening ought to go away in favor of a similar method as available in ACR. It's ok to go to ACR, but I have to flatten first, and that's a drag.

                The other option is to drop to 8 bit first in Smart, but that also is counterproductive in spades!
                • 5. Re: Timing Comparison PS vs ACR
                  Dennis 1111 Level 2
                  Lawrence,

                  If you like the results of a deconvolution sharpener like Smart Sharpen, then I think you are going to have to pay the piper and maybe get a more powerful machine. You could try Focus Magic or Focus Fixer, which are also deconvolution sharpeners, but I suspect they will also be slow (although I find Focus Magic reasonably snappy on my AMD X2 4800). The deconvolution process is compute intensive.

                  On the other hand, if you like the results from ACR but you want easier access to it from PS, it sounds like PK Sharpener would be a good fit for you. ACR sharpening was influenced by pass 1 of the 3-pass sharpening technique developed by Pixel Genius, makers of PK Sharpener. I have not experimented much with PK Sharpener but I suspect it would run significantly faster than Smart Sharpen.
                  • 6. Re: Timing Comparison PS vs ACR
                    Hudechrome-sd9sPI Level 2
                    I know that Smart Sharpen is slow, but the speeds it's running now with a 50M 16 bit file is roughly equivalent to a scanned 2400dpi 120 image back when I first installed CS3. When I got the Nikon D-80, I was delighted with the fact that Smart Sharpen now moved smartly along. Now this!

                    I alternate between Smart and ACR, letting the image tell me what looks best.

                    I did a trial with PK when it first came out and I was and am impressed. That is, I was until I started running it on b&w and I found I had to convert hundreds of well scanned, well adjusted files to RGB, just to use PK. It was a no-brainer for me. When I told the folks about my reservation, they claimed to be looking into a gray scale version, and would I be interested in participating in a Beta test? I agreed, then never heard another thing.

                    I think that sharpening, given the nature of digital, needs to be a number 1 priority in a high end graphic editing program. Paying another 100 bucks just for that seemed outrageous especially since "creative" sharpening was of no interest.

                    A little side note: When I ran the Surface Blur filter, I also saw a slowdown. But when I sharpened in ACR then ran the Surface Blur, it was back to it's usual speed.

                    Go figure!