What place does audio have in the web, anyway? The first attempts to integrate sound into the mix were dire, the graves of a thousand midi files testimony to that fact. Is there any hope?
Clearly, people aren’t happy with having strange noises thrust upon them. Browsers lacked sufficiently convenient UI to circumvent the problem so, because of aggrevated users, the web fell silent. Opt-in is the best policy for web audio or, at the very least, players which make it easy to stop the noise.
Alternatively, audio could be used completely separately from traditional web content. Largely static web pages with varying amounts of content are perhaps not an ideal way to use media. We tend to go to a web page, skim it and then move on. With the exception of short, atmospheric background audio this is at odds with how we listen to music (typically 3-4 minutes of continous use). Incompatible? Perhaps.
We need pure audio websites if being used at all. Podcasting and radio broadcasts are becoming more common online and there is a market, however niche it may be at the moment. This is the only real future I see for web audio, beyond the current standard of listening to one or two tracks by an artist. However, we should always head the words of the audioblogging manifesto. A feature for the features sake is usually worthless.