I ordered a looking glass yesterday, I obviously don’t received it yet but I observed the file HoloPlay.js to see how it works. I think I understood the main idea but I would like to know more about the native resolution of each layer used in the looking glass.
In the code, I saw that rendering could be done using a 2048x2048 or 4096x4096 texture.
With a 2048 texture, the resolution of each view is 512x256
With a 4096 texture, the resolution of each view is 819 x 455
We know that the resolution of the device is 2560 x 1600 but I don’t know how to understand this value : a resolution of a view in high resolution (820x455) is very far from 2560x1600 , and it’s impossible to fill 2560x1600 pixels with 45 images made of 820x455 pixels.
What I really want to know is the native resolution of each “layer” and if it’s possible to read a kind of movie (not interactive) made of 45 view of 2560x1600 or maybe “only” in 1280x720 or something like that. ; what is the maximum data I can send to the device for each frame and for each view.
I also saw that you can ajust the amount of view , I thought it was defined by the amount of physical “layers” but maybe it’s not. Can you give me more information about the minimum & maximum of view we can use at the same time.
Thank you for reading
EDIT : Sorry I have another question
Let say I want to render a rotating plane in 3D with the better quality possible.
The classical way to do it imply to use a certain amount of views with the same resolution for each, but if I do that, most of my views will fill be filled by transparent pixels then it’s really easy to work with but not very efficient.
An optimized way to process could be to send multiples images for each view corresponding to the bounding box of each “painted areas” and some basic infos (x,y,width,height) for each image to allow the device to draw only what need to be draw for each view. It would be not only faster to render but it also allow to use a much better quality because if you remove the empty space of each view you can transfer much more data by frame.
So my question is "would it be possible to do that in the future ? - if its not already possible - "