- From: Raymond Toy <rtoy@google.com>
- Date: Thu, 25 Mar 2021 11:21:21 -0700
- To: "public-audio@w3.org Group" <public-audio@w3.org>, public-audio-comgp@w3.org
- Message-ID: <CAE3TgXGLXijxLe1SSGiLJJNXbFP+GX9CYNRn0m4AmMsvzynf3Q@mail.gmail.com>
Attendees: Philippe Milot, Jack Schaedler, Matthew Paradis, Paul Adenot,
Raymond Toy, Hongchan Choi, Christoph Guttandin, Ruth John.
Minutes:
- Admin
- Daylight savings in the US starts Mar 14, so the meetings will be
9am PDT. I think that means it will start one hour earlier if
you're still
on standard time.
- Poll Results
- V1
- CR and Rec Status
- Raymond: 4 testimonials submitted. Will you write one Paul?
- Paul: I'll look into it; maybe have games or emscripten people
can look into it.
- Matt: Will you write one Raymond as chair, as Chris suggested?
- Raymond: Yeah, I'll write one.
- New issues
- Output channels of OscillatorNode
<https://github.com/WebAudio/web-audio-api/issues/2309>
- Paul: we can just add a sentence
- Raymond: Yeah, add one sentence like ConstantSource
- 4 of 8 Outstanding issues
- Output channels of OscillatorNode
<https://github.com/WebAudio/web-audio-api/issues/2309>
- Improve definition of regular tasks processing.
<https://github.com/WebAudio/web-audio-api/issues/2302>
- Paul: I'll take a look. It will express the same thing in a
different way. No change to implementations.
- Inconsistencies with "control thread state"
<https://github.com/WebAudio/web-audio-api/issues/2280>
- Paul: I can take this.
- Specify which task source is used
<https://github.com/WebAudio/web-audio-api/issues/2095>
- Raymond: PR in progress
- Hongchan: Is this implemented in FF?
- [Raymond failed to capture, but generally FF can do something
to make it work].
- Hongchan: Has no plans on fixing it in Chrome.
- Raymond: This is really good work. It's the closest we've been
to Recommedation.
- Raymond: And this is the 10th year since the first public draft
in 2011-12-15.
- Paul: I remember looking this all up when doing a WebAudio
retrospective for the Web Audio Conference.
- 2 PRs
- Replace "queue a task" with "queue a media element task"
<https://github.com/WebAudio/web-audio-api/pull/2307>
- Expose AudioBuffer to DedicatedWorker
<https://github.com/WebAudio/web-audio-api/pull/2289>
- Raymond: One question
- PauL: Met with Chris C and soundtrap about WebCodecs and
encoding. Doing heavy encoding and decoding. Want less in-memory
footprint of AudioBuffers (aka float32)
- Raymond: PR is v1, but issues is in V2.
- Paul: Yeah, we should move to v2.
- Hongchan: Is this blocking them?
- Paul: Yeah, but we're still figuring out what's needed.
- Raymond: Make issue priority 1 on v2.
- Hongchan: Is this about exposing AudioBuffer on workers?
- Paul: It's complicated. There's more work needed to work
with webcodecs.
- Raymond: So prio 1 on v2 because of api changes.
- Hongchan: Want comment from Chris on PR about the future
before we close it.
- V2
- Priority-1 issues
<https://github.com/WebAudio/web-audio-api-v2/issues?q=is%3Aopen+is%3Aissue+label%3Apriority-1>
- Render capacity
<https://github.com/WebAudio/web-audio-api-v2/issues/40>
- Hongchan: No updates
- Raymond: We should work out the API more completely,
defining how it behaves; we shouldn't be completely
blocked by security
issues yet.
- Paul: We could return an array of the last N values to the
user so they can do their own analysis.
- Jack: That would be nice, because then people could do
averaging and pick out min/max values over a time window
that they like.
- Raymond: Just need something concrete for devs to look over
and see if it works for them.
- Render quantum size
<https://github.com/WebAudio/web-audio-api-v2/issues/13>
- Raymond: API mostly done, just need baselatency vs size
- Raymond: FF always returns 0 for baseLatency?
- Paul: Yes. outputLatency reflects what's connected.
- Raymond: Just need to specify the interaction then. Probably
something fairly loose as long as the values (in Chrome)
are consistent in
some way.
- Paul: Stefan Letz has a nice document on how to do buffering:
https://hal.archives-ouvertes.fr/hal-02158912/file/CallbackAdaptation.pdf
- Headphone detection
<https://github.com/WebAudio/web-audio-api-v2/issues/89>
- Paul: Still investigating. When it works, it's nice, but
it's inconsistent
- Hongchan: Do some work and all platforms have some api to
expose info.
- Paul: Main problem is when people plug in speakers into
headphone jacks.
- Honghcan: CrOS can detect this via impedance
- Paul: Experience with windows is complicated. Get an overlay
asking for what's plugged in. Answers change all kinds of
things including
processing and latency.
- Hongchan: we can at least expose what the platform provides
- Input only / muted AudioContext
<https://github.com/WebAudio/web-audio-api-v2/issues/83>
- Raymond: I think we were going to change priority and forgot
to do that.
- Paul: No news.
- Hard-sync of two oscillators
<https://github.com/WebAudio/web-audio-api-v2/issues/1>
- Raymond: Ruth did some work on it last week.
- Ruth: No updates.
- Others?
- Paul: New issue on raw audio recording. SPN, AW can do it,
WebCodecs will have something.
- Jack: I think wav is pretty well handled with what we've got
- Raymond: What about FLAC?
- Paul: Yeah, that's in scope for webcodecs
- Hongchan: What does FF support?
- Paul: Opus,
- Christoph: You can do webm with pcm inside using appropriate
mimetype.
- Paul: INteresting. Violates mediarecorder specs. :-)
- Paul: webm with riff is valid.
- Christoph: that's my code to unwrap the webm container
https://github.com/chrisguttandin/extendable-media-recorder/blob/master/src/factories/webm-pcm-media-recorder.ts
- Jack: I've used it already, but haven't set aside the time to
package up the results into a valid file. In general I
find the API makes
sense
- AOB
- Raymond: Everyone should be on Daylight Time next week, so meetings
back to their normal times.
- Raymond: Poll to remain open for another week or so; we'll take a
vote in the next meeting or two.
- Paul: anyone who wants to talk about webcodecs, can contact me or
Chris C.
- Jack: (I'll add you to cc on future emails with Chris)
See you all next week at the normal times!
On Thu, Mar 25, 2021 at 8:00 AM Raymond Toy <rtoy@google.com> wrote:
> Here's the agenda for our WG/CG meeting.
>
> Meeting venue https://meet.google.com/wgg-wjtd-kdu
>
> Agenda:
>
> - Admin
> - Daylight savings in the US starts Mar 14, so the meetings will be
> 9am PDT. I think that means it will start one hour earlier if you're still
> on standard time.
> - Poll Results
> - V1
> - CR and Rec Status
> - New issues
> - Output channels of OscillatorNode
> <https://github.com/WebAudio/web-audio-api/issues/2309>
> - 4 of 8 Outstanding issues
> - Output channels of OscillatorNode
> <https://github.com/WebAudio/web-audio-api/issues/2309>
> - Improve definition of regular tasks processing.
> <https://github.com/WebAudio/web-audio-api/issues/2302>
> - Inconsistencies with "control thread state"
> <https://github.com/WebAudio/web-audio-api/issues/2280>
> - Specify which task source is used
> <https://github.com/WebAudio/web-audio-api/issues/2095>
> - 2 PRs
> - Replace "queue a task" with "queue a media element task"
> <https://github.com/WebAudio/web-audio-api/pull/2307>
> - Expose AudioBuffer to DedicatedWorker
> <https://github.com/WebAudio/web-audio-api/pull/2289>
> - V2
> - Priority-1 issues
> <https://github.com/WebAudio/web-audio-api-v2/issues?q=is%3Aopen+is%3Aissue+label%3Apriority-1>
> - Render capacity
> <https://github.com/WebAudio/web-audio-api-v2/issues/40>
> - Render quantum size
> <https://github.com/WebAudio/web-audio-api-v2/issues/13>
> - Headphone detection
> <https://github.com/WebAudio/web-audio-api-v2/issues/89>
> - Input only / muted AudioContext
> <https://github.com/WebAudio/web-audio-api-v2/issues/83>
> - Hard-sync of two oscillators
> <https://github.com/WebAudio/web-audio-api-v2/issues/1>
> - Others?
> - AOB
>
>
Received on Thursday, 25 March 2021 18:21:48 UTC