4 Replies Latest reply on Apr 27, 2010 2:04 AM by AntaraNSi

    Simple alpha matte filter

    AntaraNSi

      Hi all,

       

      I'm pretty new to pb so please don't shoot! I have a jpg that has two images. The actual image and the mask. I need to cut out the image using the mask in pb and return the result. I've managed using two images, but been unable to find out how to pick a region...

       

      any help appreciated.

        • 1. Re: Simple alpha matte filter
          Kevin Goldsmith Level 3

          not 100% sure I understand the question, the part about "but been unable to find out how to pick a region...". Do you mean region, like Pixel Bender regions, or do you mean something else. Do you mean that you want to use the mask image to extract a different sized image from the original image? Is this for Flash or for PS/AE?

          1 person found this helpful
          • 2. Re: Simple alpha matte filter
            AntaraNSi Level 1

            I have an original image that is 300x300, and a mask for that image that is 300x300. I combine them adjucent and create a jpeg that is 600x300. Now I want to give that image to a PB Kernel and use the part of the image x:300, y:0, width:300, height:300 to mask the image at x:0, y:0, w:300, h:300.

             

            I just need to get from the combined image the two original parts and mask with one the other.

             

            Thnx for helping

            • 3. Re: Simple alpha matte filter
              Kevin Goldsmith Level 3

              you could mask out the left side of the image using the region regioning functions in Pixel Bender for PS/AE, but for Flash, you will always process the entire input image, so you'd need to incorporate that loging into the kernel itself.

               

              Is there a reason that you are combining the mask with the input image into a single image instead of passing them into the kernel separately?

              • 4. Re: Simple alpha matte filter
                AntaraNSi Level 1

                The description of the problem was just for presentation purposes. The actual image has 8 frames and 8 masks all small in size. Given they have the same color table their total individual sizes is at least 4-5 times bigger (in KB) than one combined, and also there is the http requests issue. Combining 8 in one image the whole loading operation is 10fold faster (and less expensive - CDN on the other end). Now I am doing this in AS3 but i thought it would be faster in PB. I can separate them in flash i guess and pass them in kernel as separate images.

                 

                Thnx for helping.