AES British Section lectures - November 2007
Accurate sound reproduction from two loudspeakers in a living room
Siegfried Linkwitz
www.linkwitzlab.com
On the 13th of November the UK section of the Audio Engineering Society hosted the lecture "Accurate sound reproduction from two loudspeakers in a living room" by Siegfried Linkwitz. Siegfried is well known for his loudspeaker and crossover designs.
Siegfried started by stating that he comes to loudspeaker design as a hobby. His goal has been to have a satisfying listening experience in his living room, without special treatment to the room.
Siegfried made two claims in his lecture. The first was that by playing audio over two loudspeakers we can only hope to get an illusion of original event. The second was that there are certain steps we can take to minimise the confusing auditory cues which destroy this illusion.
Siegfried went on to describe various cues which can give away the illusion being created. These included on axis frequency response variation, resonances in the loudspeaker, non-linear distortion, cabinet edge diffraction, the modes of the room and the room reflections. He also discussed a couple more questionable cues: off-axis frequency response and floor reflections.
The steps Siegfried suggests to minimise these cues included trying to achieve symmetry of reflections, a minimum delay of greater than 6ms for the first reflection and ensuring the spectrum of the reflections are the same as the direct sound.
Siegfried described his living room and his comparison of dipole and monopole loudspeakers. He observed that even though the measured response of the two loudspeakers was quite different, they sounded almost identical in spectral balance and clarity. Siegfried then related this finding back to the evolution of spatial hearing. He concluded that two channel playback in a normal living room can provide a satisfying listening experience, but you have to have loudspeakers with uniform polar response and proper placement.
Siegfried then looked at the various design issues when building omnis, dipoles and cardiod loudspeakers. He finished with a look at the current practices in the audio system design and why they were wrong.
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/a0711.html.
Report by Tim Harris