Hello All
A question about how to improve the performance of rendering a lot of small pieces of text.
For example, I am rendering a point cloud of 5014 points (rendered as small spheres). On my machine, this takes about 0.86 seconds. Very acceptable. But when I add a text object to note the point number (for each of the 5014 points), the runtime shoots up to 26.03 seconds. That was a bit surprising.
While the initial rendering calculations are slow, once that's done though, manipulating the model (zooming, panning, rotating) works beautifully and is very responsive.
I am running OSG 3.6.5 on a Windows 10 machine using MS VS 2019 (Latest Version) on a 3.6Ghz Intel i9 with 32 Gb of memory and an NVIDIA GeForce GTX 1080 card.
Font file was: arial.ttf
I've tried several things to speed up the initial calculating process (all to no avail):
1. Using simpler fonts, i.e., simplex.ttf
2. Using other font formats like dosapp.fon (seemed to ignore that)
3. Rendering in either SCREEN_COORDS and OJBECT_COORDS
4. Turning ON or OFF the option: AutoRotateToScreen
5. Setting data variance to STATIC
Nothing seems to make much noticable difference.
I am wondering if there is some switch I am not setting or some procedure that I am not implementing that would help me out. (Perhaps a noob goof?)
Failing that, I was wondering if it would be possible to kick off the rendering of the text on a separate thread so that the point cloud could be viewed quickly and the text could become visible later whenever its rendering finished.
Anyone have any idea(s) that I might could try out?
Thanks in advance,
Bob Kiser