Most audiophiles seem to live in the frequency domain. Mid-range this, bass that, treble something else. But music exists in the time domain, as a waveform. Our ears have evolved (as our primary danger sensors) to be incredibly sensitive to the arrival time of sound waves. Our pinnae are shaped to give us a 3D aural image giving the height and orientation of the source sound, even if it behind our heads.
There is a Chesky recoding of a repeated chirrup which on a good 2-channel system appears to rise vertically from one speaker before moving in an arc across the ceiling and descending to the other speaker.
I cannot think of much in nature that works in the frequency domain, apart from resonances and ears. The tuned hairs in our snail-shaped cochlea fire nerves when they resonate. The sequence in which they fire feeds our neural network which processes in the time domain.
Our systems on the other hand typically take the waveform and decompose it into frequency bands before feeding each band to a dynamic driver - tweeter, mid-range, bass etc.. We then desperately try to time-align these drivers to get back the original waveform, but physics gets in the way. Unless the drivers act as a point source (not a line source) reflections from walls, floors and ceiling will not be time aligned at the listening position.
Time domain waveforms can be converted to frequency domain using Fourier transforms, and vice versa. Start with a square wave, which contains all higher harmonics of the base frequency, and produce the frequency spectrum using a Fourier transform. Take that spectrum and Fourier transform it, hoping to get the original square wave back. Well, you get a square wave but it has a pronounced leading spike, mathematically and practically.
Linn was all about trying to avoid spurious resonances in the table / arm / cartridge source, with the subjective outcome that listeners were more likely to tap in time with the music.
The most coherent loudspeaker was the Quad electrostatic ESL-63, designed in 1963 to emulate a point source of sound about a foot behind the flat panel. It took a further 18 years of development before the ESL-63 was offered to the market. The final test of each production speaker was to compare it to a reference speaker with a microphone exactly equidistant. A square wave was played to the reference speaker and the same wave with opposite polarity went to the speaker under test. If the microphone gave no output, the speaker passed. These speakers and their descendants have specifications that read like amplifier specifications. Peter Walker said of them "if you don't like what comes out, pay more attention to what goes in".
Try some. If you don't get PRaT, you probably never will!
