Design Article
Wireless audio distribution - the last 30 meters
Brent Allen, SMSC
9/21/2010 9:21 PM EDT
The vast majority of consumer music today is on CD and therefore audio quality is often measured by the quality of CD music. Music is stored digitally on a CD with the left and right channels both encoded using 16-bit samples and a sampling rate of 44.1k samples/second. This results in a total data rate of 1.411Mb/s.
Of course, music is increasingly being distributed electronically. Electronically distributed music is CD quality music (i.e. 16-bits, 44.1ks/s) that has been compressed using high quality compression algorithms that can reduce the digital file size and the download data rate (e.g. 192kb/s or 256kb/s) without a detectable loss of quality.
Well-known compression formats include MP3, WMA, AAC, and others. These formats use lossy compression that targets a specific compression ratio (e.g. 4:1, 10:1, etc.) or data rate (e.g. 128kb/s, 192kb/s, etc.) and eliminates as much information as is necessary to achieve that amount of compression. Perceptual encoding is used to minimize the audibility of the information loss. Relatively high compression ratios can be achieved at the expense of audio quality and at the cost of complexity and power consumption of the circuitry performing the encoding and decoding.
Bluetooth is an example of a wireless technology that uses lossy compression to transmit audio and SMSC's Kleer technology is an example of a wireless technology that transmits audio losslessly. The figure below shows the impact on audio quality of the two methods. Bluetooth compression results in a noise floor that is at least 40dB higher than lossless, with a corresponding reduction in SNR and dynamic range.
An argument often used to rationalize the use of lossy compression in digital wireless audio applications is that the music itself is usually stored using lossy compression and therefore the wireless transmission compression is not causing any additional reduction in audio quality. There are several problems with this argument.
The first problem is that the audio quality associated with compression formats used for wireless transmission is not as good as that associated with digital storage compression. This is because the compression used for wireless transmission must take into consideration the power consumption and latency associated with real-time compression and decompression.
The second problem with this argument is that combining digital storage compression in series with wireless transmission compression (sometimes referred to as transcoding) may reduce the audio quality more than either compression method individually. The third problem with the argument is that portable audio devices are evolving towards using lossless storage compression as the cost of storage comes down.
Portable audio devices are becoming the primary storage vehicle for entire music collections that are then played in the home and car, as well as on the portable device. This is driving a demand for higher quality (lossless) storage, in which case lossy wireless transmission will be the limiting factor in audio quality.
Next: Latency
|
|
|
|
|
Frank Eory
9/23/2010 4:07 PM EDT
Great article, nicely covering the major issues of wireless digital audio distribution. I'm looking forward to trying out an SMSC-powered set of wireless speakers.
Sign in to Reply
t.alex
9/25/2010 8:12 PM EDT
Besides the issues covered by the article, I think the number of audio streams supported is important as well. In home theatre system, the left and the right speak play different contents.
Sign in to Reply
WKetel
9/29/2010 3:15 PM EDT
While there might actually be some value in wireless audio, using it just to avoid running wires is a terrible waste of spectrum. Remember that there is only so much spectrum, and when it is full there is no more room, and something loses. So yes, it is neat to be able to do this, but it does constitute a waste of spectrum.
Sign in to Reply
t.alex
10/1/2010 8:24 PM EDT
Wketel, if the spectrum is available, why not use it?
Sign in to Reply
WKetel
10/6/2010 6:34 PM EDT
The reason for not using up all of the spectrum today is that tomorrow something much more useful may arrive that would benefit us a lot more than not having to run speaker wires. That is one reason. Next, consider that these wireless systems offer no protection from interference by other systems using the same frequencies. Beyond that, there is an interesting realm of copying concern, if we have a digital transmission of some copy protected program, and the transmitted version is not copy protected, we have just defeated the "copy police". I admit that is a stretch, but it could get some folks excited. To repeat myself, when the spectrum is all used up, it is gone, and they aren't making spectrum any more.
Sign in to Reply
Frank Eory
10/13/2010 1:37 PM EDT
As the author points out, it's much easier to find 3 MHz of available spectrum in an ISM band than it is to find 20+ MHz. Yet even with the 20+ MHz needed for 802.11g or the 40 MHz needed for 802.11n, hundreds of millions of WiFi users are able to make their WiFi devices work, despite sharing the same spectrum -- with other WiFi users as well as with Bluetooth, Zigbee, cordless phones, baby monitors, etc.
The argument that tomorrow may bring a wireless application that is more beneficial is, in my opinion, a good argument for allocating even more spectrum for unlicensed ISM use -- not an argument for limiting ISM use to only those standards that are already established (WiFi, Bluetooth, etc.).
As for copy protection concerns, that is a good subject for an entirely different discussion about analog vs. unencrypted digital vs. encrypted digital content and the DMCA law...and why copying for personal use is permitted for the first two types of content, but not for the third (encrypted).
Sign in to Reply