On Tue, Jun 11, 2013 at 11:25 AM, John Byrd <jbyrd@giganticsoftware.com>wrote: > Yes, they're easy to implement but priorities only tend to work well for > trivial nodegraphs. Once you get into complicated nodegraphs with > arbitrary node priorities you'll find that bizarre audio dropouts and the > like begin occurring. > OK. Still sounds better than #1 (per-node deadlines don't help if there are too many nodes), or #3 (doing expensive processing when you're already overloaded sounds mad). One way of improving this situation somewhat is to have node priorities > flow downstream with signal flow; however this capability is not > contemplated in the current Web Audio design. > What if you limit the use of priorities to source nodes? It seems to me that would work well for a lot of simple applications, at least. Even for many complicated networks, it seems to me it would work well, since if all the inputs to a node are dropped we can optimize the processing of that node for many types of nodes. Rob -- q“qIqfq qyqoquq qlqoqvqeq qtqhqoqsqeq qwqhqoq qlqoqvqeq qyqoquq,q qwqhqaqtq qcqrqeqdqiqtq qiqsq qtqhqaqtq qtqoq qyqoquq?q qEqvqeqnq qsqiqnqnqeqrqsq qlqoqvqeq qtqhqoqsqeq qwqhqoq qlqoqvqeq qtqhqeqmq.q qAqnqdq qiqfq qyqoquq qdqoq qgqoqoqdq qtqoq qtqhqoqsqeq qwqhqoq qaqrqeq qgqoqoqdq qtqoq qyqoquq,q qwqhqaqtq qcqrqeqdqiqtq qiqsq qtqhqaqtq qtqoq qyqoquq?q qEqvqeqnq qsqiqnqnqeqrqsq qdqoq qtqhqaqtq.q"
Received on Monday, 10 June 2013 23:35:57 UTC