3 Replies Latest reply on May 20, 2017 12:25 PM by D Fosse

    Upper limit for photomerge?


      Hi, I'm using the latest photoshop and have an issue creating very large photomerges (panoramics).


      If I load say twenty raw images from my camera it doesn't blend the images and I end up (after quite some time) with a very disjointed looking image. If I cut it down to say the middle eight images it blends them perfectly.


      (I'm not at home at the moment to post comparison screenshots but hopefully this makes sense)


      Is there an upper limit to the amount of images I can load (or at least the resulting resolution - as it would be rather large granted).


      I don't think it's a constraint of my machine, I have 16GB RAM and the process seemed to peak around 1.8GB.

        • 1. Re: Upper limit for photomerge?
          John Waller Adobe Community Professional & MVP

          Moved from Cloud forum.

          • 2. Re: Upper limit for photomerge?

            Hi I am using photomerge to stich larges images  (10472Ko 4008x2672 pixels) taken with a electron microscope camera. There is specific softwares to do such job mostly in ImageJ and FIJI that are NIH, these are open source scientific package of soft. One of these soft plungin, working in FIJI named mosaicj is doing about the same job as PS photomerge but with less options. In particular, PS has the great feature of letting you choose to fuse or not your images mosaicj do not. Up to now I tried photomerge with 15 images with success. But it seems to failed with ... 300 images. well I understand there a limit, but what is this limit ? in original images numbers ? in total pixel in one dimension ? in size of total addition weight of images in ko ? instead of running trial failures that would take me a lot of time, I should definitely appreciate to know what is the actual limit of photomerge. Thanks for your help.

            Jean-Michel Hermel

            • 3. Re: Upper limit for photomerge?
              D Fosse Adobe Community Professional & MVP

              If the process fails, resulting in disjointed segments, the problem is poor overlap and not too many frames. There has to be around 40% good overlap - meaning areas with enough matching structure that the algorithm has something to work with. Something like for instance a night sky with a few stars won't work here.


              Another common reason for failure is "false overlap" - shifting water surface, rapidly drifting clouds, that sort of thing.


              I don't know if there's a hard limit somewhere. I've seen the process run for 20 minutes or so, before finishing successfully, and I suspect it can go on for a lot longer. If you run out of machine resources you would likely get an "out of memory"-message.