Jeff Bier’s Impulse Response—Sound from the Cloud

Submitted by Jeff Bier on Tue, 02/12/2013 - 21:01

In my December column, I wrote about how smartphones and tablets are subsuming some categories of consumer electronics, such as MP3 players and networked home audio players. Because smartphones and tablets are network-centric devices, their growing use as media players is contributing to another important trend: multimedia content is increasingly being delivered on-demand via the Internet. These days, I get nearly all of my new audio entertainment content, such as podcasts and streaming music, via Internet streams. I even listen to conventional broadcast radio programs via the Internet because, ironically, I get better reception of my local radio stations in my car via the cellular network than via conventional broadcasts.

Broadcast radio, podcasts, and music services like Spotify are a natural fit with Internet streaming. But what about accessing your own (purchased or self-created) audio library? At an Audio Engineering Society conference 10 years ago, I predicted that we'd all soon have access to our entire personal music collections anywhere, anytime, thanks to inexpensive, high-capacity flash memory. The prediction has proven true; I now have my entire music collection in my tablet. Increasingly, though, audio content is not stored on local devices, but rather in the cloud. Services like Amazon's recently announced "AutoRip" encourage consumers to store their own purchased audio content in the cloud, for "anytime, anywhere" access, as long as you have an Internet connection.

The growing use of the cloud for storing and distributing audio content is leading to two kinds of new opportunities related to digital audio processing. First, the resources of the cloud enable types of functionality that weren't previously possible. Take, for example, Shazam, an awesome mobile application that "listens" to music through the microphone of your mobile device and then identifies the music. Shazam wouldn't be possible without a massive database in the cloud, correspondingly massive computation power, and ubiquitous connectivity. Along these lines, an application I haven't yet seen, but would really like is a Pandora-like service that creates a personalized audio stream featuring artists who will be performing in my local area in the next few months, and whose music shares characteristics with the types of music I typically listen to. Thanks to the cloud, this is doable.

Second, even in the realm of established audio functions, the cloud offers interesting new opportunities. For example, when audio content is streamed from the Internet, audio post-processing functions (such as bass boost or dynamic range compression) can be performed in the cloud, rather than in the mobile device. The virtually unlimited processing power of the cloud may open up possibilities for using more sophisticated (and processing-intensive) algorithms than would be practical if all digital audio processing were done in mobile devices. When I download or stream a movie from Netflix to watch on my laptop, the Netflix servers could process the audio track to compensate for the rather poor speakers in the laptop.

To me, these trends and possibilities reinforce something I've been observing for a long time: digital signal processing is becoming increasingly ubiquitous. In the early days of digital audio processing, most such processing was done on large computers in labs (I dimly recall implementing audio effects on VAX minicomputers in the '80s). Gradually, digital audio become common in large embedded systems, such as digital recording consoles. Later, PCs became the prevalent platform for many kinds of digital audio processing. Today, smartphones and tablets do quite a bit of local digital audio processing. Going forward, I think that we can continue expect all of these types of systems - PCs, smartphones, tablets, and many embedded systems - to do more digital audio processing. And, increasingly, they'll be joined by servers in the cloud.

Jeff Bier is president of BDTI and founder of the Embedded Vision Alliance. Post a comment here or send him your feedback at

Add new comment

Log in or register to post comments