See also: IRC log
<trackbot> Date: 16 October 2014
<mdjp> Scribe joe
<mdjp> ScribeNick: joe
<mdjp> Previous meeting minutes -> http://lists.w3.org/Archives/Public/public-audio/2014OctDec/0006.html
matt: most agenda today relate to
TPAC
... TPAC bug review will look through our current list of
issues and try to identify some specific cases...
... some can be closed quickly, others can be tagged as
essential for V1
<mdjp> https://github.com/WebAudio/web-audio-api/issues
matt: let's look in advance and flag bugs with the tag "TPAC 2014" so that we avoid looking through entire database
shepazu: probably a good thing to
approach bugs in this way. also a newbie mistake to concentrate
on just "your issues" at TPAC
... also important to establish liaisons with other groups at
TPAC
... want to make sure we don't lose the opportunity to have
contact with other groups
... I'm double booked on Tuesday with the Web Annotations
group
... but will be able to be pinged via IRC all all day
<cwilso_> I'd note (as previously noted) I'll be MIA Tuesday morning, but back with audio on Tuesday afternoon
shepazu: Chris Lilley will also be able to help with connections to other groups and to help the WG
matt: We want to get a grasp on
the current state of testing and udnerstand where we are and
where we might be going
... we want to understand how this work can be fanned out
between the WG and browser implementors if possible
... who in the WG has a grasp on the testing process so
far?
cwilso: Chris Lowis had some understanding
matt: have reached out to him
<cwilso_> He wrote a post about writing web audio tests, too: http://blog.chrislowis.co.uk/2014/04/30/testing-web-audio.html
shepazu: we should identify every
testable assertion in the spec, identify interactions, identify
tests that exercise assertions and their interactions
... typically tests are written as part of the process of
implementing a spec and those are submitted to the WG
... often there is a "test lead" role for each spec
... they organize the testing specs and cajole others into
writing tests
... the function of the Annotation functionality is intended to
help in identifying and notating testable assertions. I'll demo
that at TPAC
matt: I don't think we have a
history in this WG of implementors submitting tests to the
group.
... Is this something we can change in the future, i.e. can we
take advantage of browser implementor tests
cwilso: Chris Rogers submitted all the tests at the time of the original spec creation
shepazu: Would be sensible to
talk to all the implementors: Mozilla, Google, Apple, MS
... get them to review their internal tests and see which are
suitable for submission to W3C
matt: let's take stock of this at TPAC and find a way forwards
<Zakim> cwilso_, you wanted to 1) suggest "ready to spec" flag and 2) taking ownership of tpac2014 issues.
cwilso: it would be good for
someone to take the lead for each tagged bug to be prepared to
summarize the bug
... I have to admit that I'm losing track of the set of issues
that are cued up for me to find the time to update the spec,
and ones that I want others to look at
joe: let's have a tag that pre-identifies bugs as "in the spec editors inbox" i.e. there's a clear resolution just waiting for application
<mdjp> https://github.com/WebAudio/web-audio-api/issues/368
<cwilso> https://github.com/WebAudio/web-audio-api/issues/368
cwilso: this stated with an issue
on mobile devices. Default mode is HRTF, which requires loading
in a database that has a significant memory and bandwidth
footprint
... so this isn't a very friendly default. Finally HRTF is
really only useful for headphones
... I looked at the PannerNode and most people simply use this
for equal-power L/R panning
... even that is a confusing interface because one has to work
with X/Y parameters and it's not easy
... We see a case for two simpler nodes: a spatialization node
and a "simple pan" node that just does L/R panning
... obviously this is a highly incompatible change
matt: I agree that this is rather confusing; panner does not equal what this node does at the moment
cwilso: suggest that we break this into two issues: 1) changing the default, 2) a backwards-compatible proposal to introduce new nodes and deprecate the old
shepazu: newbie question
...[self-deprecating remarks]... I tried to add a panning
function to move the sounds from L to R and couldn't do
it
... would the design of a simple pan be made easier:
cwilso: yes, if there was a dedicated equal-power panner node.
<cwilso> Added issue for splitting the nodes: https://github.com/WebAudio/web-audio-api/issues/372
shepazu: I really like the idea of a simpler setup, with -1 to 1 left-to-right panning
matt: action is for Chris to propose how we will split the functionality of this node and how we'll deprecate
<mdjp> https://github.com/WebAudio/web-audio-api/issues/361
cwilso: I had an action a while
back to raise this with the TAG and see what they thought of
the pattern
... they thought it was fine. also we resolved some question
about layering architecture. Other feedback: close() a better
name than release()
matt: then do any real issues remain?
cwilso: I don't think so
... the power consumption of nonstop audio engine operation has
caused problems for some of our partners
shepazu: webplatform docs is
meant to be an unusual resources in that it's aligned to
standards as well as impls. While specifically aimed at devs
the aim is to get them involved in the standardization
process
... we're writing a script that watches for specs published at
w3c and sends a notification to the webplatform docs
pages
... this at least flags the fact that the page you're looking
at is out of date
... we're trying to keep things up to date but... you're right
that impls can vary. I think the best approach would be that if
there's something that's different between specs and impl, just
document that change.
... want to also have a caniuse feature based on W3C
tests
... short answer: we can talk about this offline
... let's aim for a statement on webplatform docs that
clarifies the situation and doesn't mislead devs and maybe
deprioritize the goal of complete docs on webplatform right
now
matt: we do need a place that can
serve as a repository of "blessed" examples of Web Audio
code
... let's discuss this too at TPAC
... AOB?
... guess not. Joe and I will send around a draft agenda at
TPAC
<mdjp> rssagent, make minutes
<mdjp> trackbot, end meeting
This is scribe.perl Revision: 1.138 of Date: 2013-04-25 13:59:11 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/today/all day/ Succeeded: s/of a simpler setup/of a simpler setup, with -1 to 1 left-to-right panning/ Found ScribeNick: joe Inferring Scribes: joe Default Present: jdsmith, +1.650.253.aaaa, mdjp, +1.415.645.aabb, +1.617.455.aacc, joe, BillHofmann, rtoyg, Doug_Schepers, gmandyam Present: jdsmith +1.650.253.aaaa mdjp +1.415.645.aabb +1.617.455.aacc joe BillHofmann rtoyg Doug_Schepers gmandyam Found Date: 16 Oct 2014 Guessing minutes URL: http://www.w3.org/2014/10/16-audio-minutes.html People with action items:[End of scribe.perl diagnostic output]