Video - getting frames "really" in? RRS feed

  • Question

  • I am trying to get a video to render into a panorama.

    In the naive case, I just define the start and the end point, then render. This yields the somewhat unsatisfactory first panorama sample as shown below.

    In the "lets get our wits out", I am trying to convince ICE to include more - somewhat successful, see the second panorama sample.

    I was struggling to get as far as the second sample in trying to improve the rendering. The new ability to (re)define video sections in the editing process was of great help, but I see two challenges:

    a) why is there such a discrepancy between the naive approach and my "assisted" approach?

    b) would there be a way to enable better back-and-forth editing by way of including a copy of the latest "previous" stitch in the rather empty right pane of the import tab? For video editing this would provide a reference point where further editing needs to take place.

    Example: Import panorama from video using naive approach -> stitch (renders) -> back to import

    --> reference stitch from previous visible on import page (possibly with: "earlier stitch" label to avoid mistaking this as a preview stitch of the current import process)

    FWIW, I can make the video available for debugging, as a link to cloud storage.

    Monday, February 9, 2015 9:34 PM