AC to AC supply

You're right Lawrence, it is all about the sound of a concert in a real hall in front of you. Close miking in studios has a similar effect of being artificial as non-one has their ears right next to a guitar, a drum, or a cello. Instead we hear it in a room.

Jumping back to the CD versus streamer bit above, I found my expensive CD player sounded hard and unnatural even via a Cusat50 uinto my Majestic. The same tracks via my PC and Squeezebox Touch sounded much better in the ways described above.

I did always wonder how the integrity of the analogue o/p of an early CD player (or any other)
could survive the brick wall analogue filtering without compromise to phase - it didn't.

RichW wrote: I used to get very very frustrated with CD sound & more than once threw a jewel case at the wall in anger ...

Better than the CD player I suppose (better for the wall)

But I'm with you on this, and I developed a theory which is put into play in both the Bitzie and the Majestic.

It goes back to a development - a rough and ready prototype - which was tried out on DVV (ex of TNT Audio).

It shifted the phase at high audible frequencies without altering amplitude. It had some limited success. DVV liked it.

The theory goes, if you alter the phase then the brain expects a difference in amplitude. Like I say, it is only a theory. But early CD analogue anti-aliasing filters gave a large phase difference without a corresponding attenuation of the signal. The theory explains that the brain will be subconsciously puzzled by this, which might result in frustration.

The test development shifted the phase even further, and with a particular popular CD player's analogue filter, rotated the phase back to "flat".

But not all CD players use the same filter idea, so I knew it wasn't saleable. But it helped the theory along.

Oversampling digital filters have moved the Nyquist frequency upwards, and so today's analogue filters don't have to be so steep, or they don't have to act so early.

With the help of SPICE modelling it was easy to see where the phase started shifting with respect to frequency, and also see how much attenuation each filter idea would provide at the oversampled Nyquist frequency.

I decided as my (and most adults) hearing started to cut at 13kHz, I would arrange for the phase shift to reach -45 degrees - which in "natural analogue" is -3dB amplitude, to be a little up from 13kHz.

As I could not hear much above that, I reasoned that the phase shift would appear "normal". With most data sheet filters the phase shift is much lower than 13kHz and far more than 45 degrees.

In fact, there is a phase test on most audio analysers and it has to be for more than testing absolute phase, one would think.

In broadcast audio I recall it was often stated that 12 degrees throughout the audible range, was acceptable for most applications.

In which way do my CDs sound coloured? I would say as in not natural sounding. A kind of added sound thing, similar to how I've explained heterodyning, or beat frequencies putting sounds or colorations in where I simply know they should not be.

I am pretty sure it is the method used to produce the S/PDIF output that has something to do with it. This is why I now take a data copy using EAC (of the CDs I have bought) and play them using Foobar2000 (seeing I use Windows).

The bit depth and sampling frequency is the same, but I can now listen and enjoy.