I read with interest the article by Doug Irwin in November’s
edition of Radio magazine called High-Performance Audio.
The points raised in this article should allow any self-respecting
engineer to improve their station sound. However, I do believe that one
important aspect of the broadcast chain has been omitted, namely audio
compression and the effects of multiple and tandem passes.
Be it contribution, storage or distribution; audio compression is a
fact of life in the broadcast chain. The greatest challenges facing the
engineer are how the content will sound after each pass has taken
it’s toll and how to overcome the latency introduced by the
algorithms that require a degree of processing power.
In essence there are fundamentally two types of algorithms: time domain
and frequency domain. Frequency domain algorithms (ISO MPEG Layer II,
Layer III, PAC) tend to have higher compression ratios and actively
remove content deemed to be irrelevant to achieve this process. Thus
adversely affecting the content after multiple passes and introducing
an unmanageable latency i.e. 100+ milliseconds. Time Domain algorithms
(apt-X, G.722) tend to have a lower compression ratio (4:1), are less
destructive in their approach to coding techniques and have a
manageable latency of 3 milliseconds.
Each type of algorithm has its place in the broadcast chain when
balancing operational costs versus quality. However, if an engineer
wants to improve sound quality, arguably a revised approach on the
stations compression algorithms could assist in achieving this
Belfast, N. Ireland