Comments on: CanvasArrayBuffer and Canvas*Array

By: Ilmari Heikkinen

Ilmari Heikkinen — Tue, 17 Nov 2009 15:24:50 +0000

The example about overlapping arrays to a single buffer makes a case for having a byteStride for the WebGLArrays. I mean, what you’d like to write is “points[i*3] = foo; normals[i*3] = bar; texcoords[i*2] = qux;” where points, normals and texcoords all point to the same ArrayBuffer.

Well, what I’d actually like to write is “points[i] = vec3(1,2,3); normals[i] = vec3(1,2,3); texcoords[i] = vec2(1,2);”, but that’s orthogonal to byteStride.

By: James Boston

James Boston — Sat, 14 Nov 2009 20:09:42 +0000

This will be a great help to c3dl developers. Thanks.

By: Two big retrospective changes to the lessons - Learning WebGL

Two big retrospective changes to the lessons - Learning WebGL — Sat, 14 Nov 2009 03:48:47 +0000

[...] was odd, given that the classes weren’t really canvas-related, but as Vladimir Vukićević pointed out the other day, the classes could be useful for JavaScript code that has nothing to do with WebGL. [...]

By: Jason Oster

Jason Oster — Mon, 09 Nov 2009 19:07:44 +0000

I actually like the sound of this. One of my problems with the Canvas API (as a video game level editor author) is that using a canvas ImageData / CanvasPixelArray to fill a large pixel buffer with random pixels (like a complete SNES level will do) is much slower than it ought to be. Filling up a “CanvasUnsignedWordArray” or whatever the unsigned 32-bit data type is, would be much faster than writing 4 bytes to a CanvasPixelArray per pixel.

Wherever the CanvasArrayBuffer API goes, I hope (please! please!) that it will also be usable directly with putImageData() and friends.

By: WebGL around the net, 9 Nov 2009 - Learning WebGL

WebGL around the net, 9 Nov 2009 - Learning WebGL — Mon, 09 Nov 2009 13:47:28 +0000

[...] Vukićević has posted a great article clarifying what CanvasArrayBuffers and the associated CanvasFloatArray, CanvasByteArray, and so on [...]

By: Giles

Giles — Mon, 09 Nov 2009 12:50:03 +0000

Vlad,

Many thanks for the explanation, that clears up some discussion we’ve been having on the WebGL forums: http://www.khronos.org/message_boards/viewtopic.php?f=45&t=2163&p=5404

However, as one of the guys in the forum said, it looks like there’s still a little way to go in Minefield; if you run:

var f2 = new CanvasFloatArray([12.3, 23.4, 34.5]);
alert(f2[2]);

…the alert box says “undefined”. It works as you’d expect in Chrome.

(Happy to put in a bug report if that would help.)

Cheers,

Giles