AES British Section lectures - October 2006

Audio is done, isn't it?

Keith Howard

www.audiosignal.co.uk October lecturer Keith Howard

On October 17th the UK Section of the AES hosted a lecture titled "Audio is done, isn't it?" by Keith Howard, Hi-Fi News consultant technical editor and Stereophile contributing editor. The talk set out to discuss six of the challenges still to be addressed in the audio industry.

Keith started off with a discussion of multichannel audio, talking about Quadraphonics (what went wrong), the reasons for going multichannel, the types of multichannel sound and issues surrounding soundfield delivery (Ambisonics and Wave Field Synthesis).

Keith then went on to look at Trifield processing and gave a convincing demo showing the benefits including improved central image stability and wider listening area. Keith suggested that one of the reasons behind the small uptake of Trifield processing included the fact that there is only one piece of, expensive, software available to do the processing.

Keith then covered the problems with jitter and how they are effected by modern high definition audio formats. He also discussed the audibility of jitter at different frequencies, gave a demonstration of what jitter sounds like and talked about mechanisms for measuring it.

Keith then went on to talk about some tests he did to look at how the pre-response on down sampling filters effects the output audio. He concluded that it had no audible effect.

Keith discussed the persistent problem of speaker-room interaction. He contended that speaker directivity should be contrived to make it as simple as possible for the ear/brain to recognise and suppress the effect of the room, so speaker directivity must be constant with frequency. Pointing out that omni's were not necessarily the answer, he pointed out the various techniques for directivity control.

Finally, Keith covered the topic of microphone noise and how it interacts with quantisation, effectively removing the need for dither.

He gave an interesting demo which showed only when the microphone noise was very low did the quantisation error signal contain an audible distorted copy of the original music signal.

There followed a lively question and answer session. A recording of the lecture is available from the UK section website at www.aes.org/sections/uk/meetings/a0610.html.