[Minutes] Teleconference Thurs 24 Sept 2015 12 noon EDT

Please excuse the rough form of these minutes as they're cut from an IRC
window.

[11:06] <BillHofmann> Joe: let's start with review of AudioWorker
[11:07] <padenot> I'm here
[11:07] <padenot> but you can't hear me
[11:07] <padenot> one sec
[11:08] <BillHofmann> padenot: now taking his notes and converting to
formal spec text
[11:09] <BillHofmann> padenot: creating section 3 - processing module - how
rendering thread/control thread works, how you sendMessages, etc - need to
push into his spec (3.4 partially done, 3.3 done, ...)
[11:10] <BillHofmann> padenot: explains how webaudio works internally - it
explains how it should work, functionally
[11:11] <BillHofmann> billhofmann: is the text normative?
[11:11] <BillHofmann> padenot: there will be *some* normative text, but
much is informative
[11:12] <BillHofmann> padenot: need to specify how you order nodes in
rendering - it's observable, so should be specified
[11:12] == jdsmith [~jdsmith@public.cloak] has joined #audio
[11:12] <jdsmith> present+ jdsmith
[11:13] <BillHofmann> padenot: will spec how nodes are executed - won't
change much from cwilso's original spec
[11:13] <ToddG> present+ ToddG
[11:13] <BillHofmann> padenot: important how changes from control thread
are reflected in rendering thread
[11:14] <BillHofmann> billhofmann: do we think some kind of nightly version
is possible by TPAC
[11:14] <rtoyg_m> We won't have an implementation ready by then either.
[11:14] <BillHofmann> padenot: no.  will have spec ready for detailed review
[11:16] <BillHofmann> joe: ready for review is great - would be good to let
group know beforehand so we can review prior to TPAC
[11:18] <BillHofmann> joe: asks if people are good with this approach
(spec'ing the processing model, building audioworker atop)
[11:18] * cwilso afk:brb
[11:18] <BillHofmann> <general agreement>
[11:18] <BillHofmann> joe: review of progress on issue resolution...
[11:19] <BillHofmann> joe: made a lot of progress... 15 issues closed, lots
happened in the last 2 weeks
[11:19] <joe>
https://github.com/WebAudio/web-audio-api/issues?q=is%3Aopen+is%3Aissue+label%3A%22Needs+WG+review%22
[11:20] <BillHofmann> Review of #560 - clarify computedPlaybackRate
[11:21] <BillHofmann> joe: move to ready for editing
[11:21] <BillHofmann> Review of #541
[11:22] <joe> https://github.com/WebAudio/web-audio-api/issues/541
[11:23] * cwilso back
[11:23] <BillHofmann> joe: ray/paul agree, move to ready for edit
[11:25] <joe> https://github.com/WebAudio/web-audio-api/issues/535
[11:25] <BillHofmann> cwilso: haven't discussed with TAG yet.
[11:26] <joe> https://github.com/WebAudio/web-audio-api/issues/532
[11:26] <BillHofmann> joe: this is part of padenot's work
[11:26] <BillHofmann> billhofmann: should we close?
[11:27] <BillHofmann> joe: no - describes his work
[11:27] <joe> https://github.com/WebAudio/web-audio-api/issues/503
[11:29] <BillHofmann> cwilso: changes what happens if you've called
setTargetValue, then set value, breaks an incorrect use case
[11:30] <BillHofmann> joe: have a pull, not in any urgency to merge, would
rather have the discussion and resolve properly
[11:30] <joe> https://github.com/WebAudio/web-audio-api/pull/583
[11:31] <joe>
https://github.com/WebAudio/web-audio-api/pull/583#issuecomment-135505186
[11:32] <BillHofmann> <people on call review the threads...>
[11:36] <BillHofmann> joe: can hoch split the issues?
[11:36] <BillHofmann> hoch: yes
[11:38] <BillHofmann> rtoyg_m: noting that at least in Chrome, once you
start an automation, the code path is different than if you do a set value
[11:38] <BillHofmann> rtoyg_m: addressing this will cause an observable
change
[11:39] <BillHofmann> rtoyg_m: we should do the #503 change, it clarifies
[11:39] <BillHofmann> joe: we'll merge PR583, close #128 and #583, file
cancelScheduledValues as a separate issue
[11:40] <BillHofmann> hoch: will write up proposed solution and distribute
[11:41] <joe> https://github.com/WebAudio/web-audio-api/issues/436
[11:44] <BillHofmann> joe: general consensus that 3 nodes would be impacted
with this... long discussion in bug.
[11:47] <BillHofmann> cwilso: need to define what the observable behavior,
not clear why we need to throw.
[11:47] <BillHofmann> joe: propose we use time at TPAC to figure this out
properly, unless someone wants to step up...
[11:49] <BillHofmann> jdsmith: is reuse tied to nulling?
[11:49] <BillHofmann> cwilso, joe: yes, perhaps (per joe) a degenerate
case, but hard to separate
[11:51] <joe> billhofmann: jdsmith, joe and I met yesterday to resolve some
gaps in the device output APIs. we are trying to propose a companion API to
getUserMedia() which is for the input side
[11:52] <joe> billhofmann: we want to present a proposal to WebRTC @ TPAC
describing a proposal to obtain a device with constraints, examine the
capabilities of output devices, and create AudioContexts that direct output
to those devices.

-- 
.            .       .    .  . ...Joe

*Joe Berkovitz*
President

*Noteflight LLC*
49R Day Street / Somerville, MA 02144 / USA
phone: +1 978 314 6271
www.noteflight.com
"Your music, everywhere"