Audio Coding, Pt. 3

(Page 3)

But Layer II is used more widely for satellite program distribution and is an important contender for Digital Audio Broadcast. Why is this?
Layer II is currently the accepted coding scheme for applications with bit rates of 128kbps per audio channel. There is no logic, however, in concluding that because Layer II is preferred for satellite, it should be used for ISDN. Nor is it true that if you use Layer II in one part of your audio chain it is the best choice for all parts of the system.

Some Layer II-only manufacturers have criticisms of Layer III. Can you address them?
Certainly. There have been interviews in several journals with Layer II-only manufacturers who state that Layer III is too complex and too expensive. When you consider Zephyr, it's obvious that these are unsound arguments. Complexity and cost are design hurdles that Telos has overcome to create products that are reasonably priced, easy-to-use, and incorporate the most appropriate technology. There is nothing that prevents the Layer II-only manufacturers from making the same investment in resources and commitment to their customers. They, for whatever reasons, choose not to do so.

Do you have other data that support Layer III's superiority at ISDN bit rates?
Audio Precision, the world's most respected developer of audio test equipment and procedures, recently released a test suite for evaluating codec performance. This protocol, TechNote 14, can be used to demonstrate two deficiencies in Layer II's ISDN bit-rate performance. Layer II is limited to about 10kHz when used on a single 64kbps ISDN "B" channel. In addition, Layer II's joint-stereo mode reduces stereo separation when used on two ISDN "B" channels. Nearly all discrete channel audio information above about 6kHz appears on both channels of the codec's output.

On the other hand, Layer III permits 15kHz mono on a single "B" channel, which is a perfect match for FM broadcast. Layer III also offers clearly superior stereo separation as well as dual-mono on two "B" channels.

What is the history of Layer III and Layer II?
Some years ago, the international standards group ISO/IEC established the ISO/MPEG (Moving Pictures Expert Group) in order to develop a universal standard for encoding moving pictures and associated audio for use with digital storage and transmission media. The standard was finalized in November 1992, with three related algorithms, called Layers, being defined for encoding of audio, taking advantage of psychoacoustic effects. A variation of Layer I is being used for Digital Compact Cassette. Layers II and III have found more widespread use in broadcast and professional audio applications.

Zephyr harnesses the power of both Layer III and Layer II efficiently and economically. Whichever you choose, you can rest assured that you are getting optimal audio performance.

Where does Telos stand on compatibility?
Our industry must avoid proprietary schemes. It is unfair to lock customers into old technology in order to maintain compatibility with their existing equipment, rather than allowing them to take advantage of emerging technologies.

It is also essential for codec manufacturers to comply with the coding algorithms that have been carefully crafted. Our coding is "laboratory-reference," strictly adhering to the algorithms of their developers, Fraunhofer for Layer III and the IRT for Layer II. We have strong working relationships with these laboratories and intend to utilize the best and most powerful coding algorithms regardless of their origins.

Do you see cascading multiple codecs as a problem?
It appears that, in the near future, most broadcast audio signals will be subject to multiple encode/decode cycles before they reach the listener. Unfortunately, there is no completed research into the effects of cascading using "real-world" broadcast audio chain configurations. Our view is that one should code only where the bits are unavailable or expensive, and then use the most bits and the most powerful coding method.

•