Agree. In particular, with native client and webgl (and I hope some, at least, half decent thought given to audio), something the under the ground is moving.
Mozilla (specifically Vladimir Vukićević) has been the driving force behind WebGL from the start, and they're also leading the way in audio: https://wiki.mozilla.org/Audio_Data_API#Authors
Just curious if anyone is actively developing, or has plans to develop, any applications or services targeting WebGL in a business context?
We've run some initial tests and its a very promising visualization surface. The best place to start are the lessons on Giles Thomas' up-to-date and instructive site:
Everything is pretty straigtforward if you are familiar with OpenGL ES 2.0.
N.B. I've had better luck with Firefox 4 Minefield than the Chromium builds, and Mozilla also has its own WebGL tutorial site but its not as current as Learning WebGL:
I'm currently building a visualization app(with medical data) using WebGL (and google's O3D framework) and it's working great. I strongly suggest looking at O3D even if it's a little rough right now since they transitioned from a plugin to webgl. The doc hasn't yet been updated.
We already have users using it in work settings using firefox 4 beta and chrome builds.
The worse problem I'm hitting with this stuff is getting the browser to fetch the data needed quickly and efficiently. I'm trying to figure out if I can use the BinaryBlob interface in the new javascript FileAPI spec to see if that could be used. Right now I'm parsing plain text data (gzipped server side) and it's slow.
3D medical data sets, from PET scans and such, are indeed huge in size. Typically GB's of voxel data. The sort of thing that requires an overnight download unfortunately.
One possible optimization is to recreate the 3D vertex set from a series of 2D texture images, by interpolating vertex positions between frames. Surprisingly only a small subset of images is needed to recreate the resulting 3D geometry, but the tessellation calculations are pretty math intensive.
Wow, I didn't know someone at Robarts was doing this stuff( I work for McGill another University in Canada ;p)
I'll have to communicate with him.
I've managed to get some full volumes loaded in my app. It was a scan of my brain. That work ok (slow and takes a few minutes to load) but when I loaded the brain or a rat, my app totally crashed. To be fair the rats brain was at a higher resolution :)
12 comments
[ 3.1 ms ] story [ 38.9 ms ] threadhttps://sites.google.com/a/chromium.org/dev/developers/demos...
We've run some initial tests and its a very promising visualization surface. The best place to start are the lessons on Giles Thomas' up-to-date and instructive site:
http://learningwebgl.com/blog/
Then delve into the draft Khronos WebGL spec:
https://cvs.khronos.org/svn/repos/registry/trunk/public/webg...
Everything is pretty straigtforward if you are familiar with OpenGL ES 2.0.
N.B. I've had better luck with Firefox 4 Minefield than the Chromium builds, and Mozilla also has its own WebGL tutorial site but its not as current as Learning WebGL:
https://developer.mozilla.org/en/WebGL
We already have users using it in work settings using firefox 4 beta and chrome builds.
The worse problem I'm hitting with this stuff is getting the browser to fetch the data needed quickly and efficiently. I'm trying to figure out if I can use the BinaryBlob interface in the new javascript FileAPI spec to see if that could be used. Right now I'm parsing plain text data (gzipped server side) and it's slow.
3D medical data sets, from PET scans and such, are indeed huge in size. Typically GB's of voxel data. The sort of thing that requires an overnight download unfortunately.
One possible optimization is to recreate the 3D vertex set from a series of 2D texture images, by interpolating vertex positions between frames. Surprisingly only a small subset of images is needed to recreate the resulting 3D geometry, but the tessellation calculations are pretty math intensive.
I'll have to communicate with him.
I've managed to get some full volumes loaded in my app. It was a scan of my brain. That work ok (slow and takes a few minutes to load) but when I loaded the brain or a rat, my app totally crashed. To be fair the rats brain was at a higher resolution :)