Audio Data API

Programmable audio APIs will do for <audio> what <canvas> did for <img>. Prior to the canvas tag, images on the Web were largely opaque to scripts. Image creation and manipulation had to occur on the sidelines—namely, on the server. Now, there are tools for creating and manipulating visual media based on the canvas element.
Similarly, audio data APIs will enable music creation in HTML 5 applications. This will help round out the content-creation capabilities available to web applications and move us closer to a self-hosting world of tools to create media on and for the Web. Imagine editing audio on the Web without having to leave the browser. Simple playback of sounds can be done with the <audio> element. However, any application that manipulates, analyzes, or generates sound on the fly needs a lower-level API. Text-to-speech, speech-tospeech, synthesizers, and music visualization aren’t possible without access to audio data.

We can expect the standard audio API to work well with microphone input from the data element as well as files included with audio tags. With <device> and an audio data API, you may be able to make an HTML 5 application that allows users to record and edit sound from within a page. Audio clips will be able to be stored in local browser storage and reused and combined with canvas-based editing tools. Presently, Mozilla has an experimental implementation available in nightly builds. The Mozilla audio data API could act as a starting point for standard cross-browser audio programming capabilities.