W3C home > Mailing lists > Public > public-audio@w3.org > January to March 2012

Re: Audio-ISSUE-8 (output-rt-analyser): about numberOfOutputs of RealtimeAnalyserNode [Web Audio API]

From: Jussi Kalliokoski <jussi.kalliokoski@gmail.com>
Date: Fri, 23 Mar 2012 11:38:10 +0200
Message-ID: <CAJhzemWjwO0hB3Opcdg_Nn9DgfZOVVK-m+nm12zDzNxBYNWgKA@mail.gmail.com>
To: Audio Working Group <public-audio@w3.org>
Hmm. This is interesting, because FFT definitely has outputs. How many,
that depends on how you look at it, it might be buffer size divided by two,
or it might be one, containing the same amount of channels as the input. If
you think about the latter, there are interesting use cases for it. Let's
take a vocoder and pretend that the RealTimeAnalyserNode has a parameter
called "reverse":

Carrier Signal -> FFT -v
                                      ----> JavaScriptProcessingNode -> FFT
Vocal Data ------> FFT -^

Here the JavaScriptProcessingNode would pretty much just throw a window
function (or many) at the data and multiply the signals in the frequency
domain, and so forth (I can't remember the specifics), and that's it, we'd
have a working vocoder. Sounds pretty neat to me, what do you think?


On Fri, Mar 23, 2012 at 2:56 AM, Audio Working Group Issue Tracker <
sysbot+tracker@w3.org> wrote:

> Audio-ISSUE-8 (output-rt-analyser): about numberOfOutputs of
> RealtimeAnalyserNode [Web Audio API]
> http://www.w3.org/2011/audio/track/issues/8
> Raised by: James Wei
> On product: Web Audio API
> In the spec of Web Audio API, for RealtimeAnalyserNode, there is a note
> for output number:
>   numberOfOutputs : 1    Note: it has been suggested to have no outputs
> here - waiting for people's opinions
> The idea behind it is that It will be good that some inspector Node like
> RealtimeAnalyserNode can pull without connect to downstream node which
> ultimately connect to an audio destination.
> Currently we have to connect RealtimeAnalyserNode to a GainNode with gain
> as 0 and then connect to destination to analyze the stream and avoid to
> impact the audio playing.
> Another potential solution is to add multiple destination support in Web
> Audio API and add a mock destination for the inspector purpose and this
> destination is synced with the real destination.
> http://lists.w3.org/Archives/Public/public-audio/2012JanMar/0450.html
Received on Friday, 23 March 2012 09:38:43 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:49:58 UTC