1 Reply Latest reply on May 5, 2014 6:07 AM by Noel Carboni

    building a composite?

    birdofillomen

      *this sort of pertains to another topic i created about my confusion over resampling

       

      hopefully i convey this simply.

      i like to take found images and build composite images from them - i'll take part of one image and part of another, and then another, and so on, and build and create an entirely new image/object.  i usually do this multiple times over so i end up with a design that is a composite of several smaller composites, and maybe more.

       

      almost always the images are off the internet and are jpegs at varying sizes, resolutions, pixel widths.

      i'm assuming an image built from other images, that all have differing values(sizes, pix, res) won't look as good as one built from the same images with the values (or a value) at a constant,  right?

           ex: having all the source images set at the same pixel width or resolution, or whatever.

      if so, which (pixel width, resolution)?

       

      (from copy) i've always just set the resolution of each image to 300ppi and resampled it to a reasonable size.  if it were super small - 1.5x..., i'd probably increase it to about 4.  if overtly large  - 20x... i'd shrink it to around 12x...

      then i'd cut whatever out that i wanted to use to be placed.

       

      apparently this isn't right at all but i've been lucky up until now with no issues, at least none you could tell.  however, as good as the final products turned out i'm sure they could've been better.

      i've thought about taking each piece that i cut out and saving it as its own separate file, like a png, re opening it and then placing it, but i'm not sure if that's all that necessary or has much of an affect on anything.

       

      any advice is appreciated, thanks.

        • 1. Re: building a composite?
          Noel Carboni Level 8

          Saving as a separate file won't net you much but if the image you're placing is lacking in resolution, opening it first as a separate document (e.g. by dragging it to the border of Photoshop rather than onto the composite image in the workspace), upsampling it with the Detail Preserving setting, then placing the result might yield a slightly crisper look than placing the image at the original pixel size then transforming it to be larger via the Bicubic setting.

           

          Adobe claims the Detail Preserving upsampling method cannot easily be implemented as an available upsampling method within the Transform function for technical reasons.

           

          -Noel