Issues with mapping video to jit.gl.gridshape for camera control
The goal of this patch is to map a movie to a gridshape so it can be positioned with a camera, while preserving the correct aspect ratio. Even though I am using transform_reset 2 in the videoplane, the aspect ratio is squeezed. If I use transform_reset 2 in the gridshape or the node objects, I have the correct aspect ratio but the camera no longer controls the image.
This is a real puzzle!
Thanks, Randall
Yeah, but you're using a square plane to project the rectangular movie to. Adjust the x-scale for the gridshape to 1.25 or something (probably not the right ratio but it looks ok) and all will be well...
Edit: 133 is the correct scale, see patch
Thanks Luvulongtime, you are absolutely right about scaling the plane. However, there are a bunch of other issues I have been trying to solve around this idea of mapping a video to a gridshape while preserving aspect ratio AND maintaining camera control. There were a whole bunch of little quirky things to figure out , such as doing the scaling with jit.gl.node so as to preserve the scale and aspect ratio of both the gridshape and the videoplane, otherwise all kinds of crazy things happen (at least for me). I have updated the patch for anyone who is interested in doing something like this.
Any feedback/suggestions would be welcome!!
Glad to help, what other issues are you having then? The patch works fine now doesn't it?
My original question was whether or not sending a video to a gridshape key matte is the best/only solution for achieving camera control. Secondly, I originally used the node object to place various effects between node and videoplane, but it seems certain effects work better if they are placed between the video and the jit.gl.gridshape object. I am going to expand this demo patch (the one I am working on is much more extensive) to include things like cropping, luma key, noise displacement, and other ISF shaders. Just trying to get input on the various strategies towards the end goal situating a video (in my case a live streaming video of a live performer) as an animated object inside of a 3D virtual set, which is why the camera control is so important. Not sure if this makes any sense but I'll continue to upload more examples.
Yeah, you're about three "levels" of manipulation down now. Don't know if that makes sense but yeah, it depends on which level you want/need to do what kind of manipulation. In the case of video it should be done first.