- From: Marcus Geelnard <mage@opera.com>
- Date: Fri, 10 Aug 2012 07:17:53 +0000
- To: Chris Rogers <crogers@google.com>
- Cc: olivier Thereaux <olivier.thereaux@bbc.co.uk>, Mark Boas <markb@happyworm.com>, Jussi Kalliokoski <jussi.kalliokoski@gmail.com>, James Wei <james.wei@intel.com>, Stéphane Letz <letz@grame.fr>, Audio Working Group <public-audio@w3.org>, Matthew Paradis <matthew.paradis@bbc.co.uk>, Christopher Lowis <Chris.Lowis@bbc.co.uk>
Citerar Chris Rogers <crogers@google.com>: > On Thu, Aug 9, 2012 at 1:16 PM, Marcus Geelnard <mage@opera.com> wrote: > >> I hope that we can have a serious discussion with the starting point of >> making custom processing a first class citizen of the Web Audio API, >> including identifying ways to make performance and latency as optimal as >> possible. If we all conclude that some of the required changes are not >> feasible at this point in time, we'll have to settle for something less, >> but let's not close the debate prematurely based on assumptions such as >> "It's complicated", or "Very few developers will use it anyway". >> > > Hi Marcus, > > So far, we've been talking about allowing JavaScriptAudioNode to run in a > web worker, adding the potential for multiple inputs/outputs, and also > potentially adding tighter integration with AudioParam. I know you've also > been working on a math library, which looks like something which could be > designed in parallel as a general-purpose API and that a > JavaScriptAudioNode could call these functions. > > What are your thoughts here? Hi Chris, I think we're heading in the right direction, for sure. - JSAudioNode interface (AudioParam support etc): I think "potentially" is a bit vague - we need some sort of resolution here. IMO it's important that you can mimic native nodes using custom nodes. It's useful for all kinds of purposes (customization/extensions, testing, future compatibility shims etc), but perhaps most importantly it will make sure that native and custom nodes are "equals". What are your thoughts? - DSP library: My experiments so far show that you can implement most/all of the native nodes efficiently using the DSP library, plus you have the ability to do more interesting things that I think wouldn't be natural to do using native nodes in the Web Audio API (for instance, dynamic frequency domain effects and synthesis). I think it would be the perfect companion to the Web Audio API (I'm obviously a bit biased here ;) ). - Web workers: Perhaps this is the area that needs to be explored a bit deeper. What are the implications of different solutions? What can we do to get latency-free processing or at least reduce the latency? I don't have any conclusive answers, so ideas are welcome. /Marcus
Received on Friday, 10 August 2012 07:18:39 UTC