Regarding Halcro amps, the fact that they do very well on a few static distortion tests with sine wave input signals doesn't mean they would have a good null test result.

In a CD of a live recording that was recorded digitally and digitally mastered what is the "input signal" that a DAC has to accurately reproduce?

Not quite so simple is it  .  The input signal to an amp however is always a known quantity.

Coming back to my DAC and why it might sound different to the DAC64, if my DAC measures similarly to the 47 Labs DAC then it does indeed have a different freq. response.

Here's the 47 Labs DAC freq response (taken from http://www.stereophile.com/digitalsourcereviews/800/index5.html):

(the HF boost trace is for a CD with pre-emphasis which the DAC ignores so can be ignored)

...and here's the freq. response graph for the DAC64 (taken from http://www.stereophile.com/digitalsourcereviews/624/index6.html):

(the DAC64, like most DACs does handle pre-emphasised CDs so the de-emph. trace looks just like the normal one).

It's clear the the 47 Labs DAC has some HF rolloff  compared to the pretty much flat response of the DAC64.  I personally doubt whether that is what makes the 47 Labs DAC (or my clone) sound more lifelike.

Michael.