Re: AudioBufferSourceNode.gain

On Mon, Apr 15, 2013 at 8:43 AM, Ehsan Akhgari <ehsan.akhgari@gmail.com>wrote:

> WebKit and Blink implement a gain attribute on the AudioBufferSourceNode,
> and looking at the implementation, it works exactly as if the output of
> AudioBufferSourceNode was directly fed into a GainNode.  However, this
> attribute is not spec'ed, and therefore Gecko doesn't implement it.
>
> Over in https://bugzilla.mozilla.org/show_bug.cgi?id=861591, we have a
> game which is relying on AudioBufferSourceNode.gain and breaks in Firefox
> Nightly because that attribute does not exist.  We can do either of the
> following two thing:
>
> 1. The right thing to do: continue to not spec AudioBufferSourceNode.gain
> and remove it from both WebKit and Blink.  I doubt that there is a lot of
> content out there using this attribute, given how late this bug was found
> in Gecko.
>

Unfortunately there's enough content out there using it that I cannot
recommend removing it from WebKit/Blink.  One notable one is
http://playcanvas.com/ which is a game library.


>
> 2. If WebKit/Blink is not willing to remove this non-standard attribute,
> then we should perhaps spec it.  This will be sad, since this attribute
> doesn't really serve any purpose and it's unusual for a node to support the
> functionality of another node in this way.
>

I agree that it's redundant, and slightly ugly, but I think we'll need to
spec it.  Luckily, it's very easy to implement and its meaning is clear.


>
> Regardless of the outcome of this discussion, I will try to reach out to
> the author of that game and ask them to change their code based on the Web
> Audio spec.
>
> Thanks!
> --
> Ehsan
> <http://ehsanakhgari.org/>
>