Re: JavaScriptNode bufferSize

Hi Jussi,

Can you make sure we have the issue documented in our tracker?
https://www.w3.org/2011/audio/track/issues/new

Our agenda for this week's call (tomorrow) is not set yet, but it seems like this issues, as well as the one Matt raised last week, could be a good basis for discussion on JavaScriptNode buffer size and resulting delay.

Olivier

On 7 May 2012, at 13:06, Jussi Kalliokoski wrote:

> Hello group!
> 
> We discussed with Chris (off-list) the recent change from numberOfOutputs to numberOfChannels in the createJavaScriptNode(), because I noticed my sink.js was broken in the latest Chrome Canary.
> 
> I suggested that the first parameter (bufferSize) be optional as well, reverting to default when passed undefined. In an advanced implementation, this default could be adjusted to a value that the implementation deems a good tradeoff between safety (avoiding underruns) and latency on that system and setup. I believe Chris thinks this is a good idea, so he suggested I bring this to discussion in the group so we can change the spec if we agree on it.
> 
> Thoughts?
> 
> Cheers,
> Jussi