Ok, I found some time to analyze Yannick's test-runs.

I took the files and extracted the impulse tests. Or more accurately: the DA - cable - AD impulse response. For my tests, I decided to use an impulse length of 6404 (it's much larger than really required in our case).

All impulses where scaled to reach 0dB at exactly 5kHz and have been windowed with a Kaiser window (alpha = 3).

Here follows all 3 tests results:

**AES 01**
Impulse response (only the central ~100 taps are shown here)

Frequency response +/- 1dB, linear frequency scale

Frequency response +/- 1dB, log frequency scale

**AES 02**
Impulse response (only the central ~100 taps are shown here)

Frequency response +/- 1dB, linear frequency scale

Frequency response +/- 1dB, log frequency scale

**Velox 01**
Impulse response (only the central ~100 taps are shown here)

Frequency response +/- 1dB, linear frequency scale

Frequency response +/- 1dB, log frequency scale

**First conclusion**
It is clearly visible that we most of all measured the DA and AD converter's response. One can easily spot the modern ADA converter's linear phase Nyquist filter section tuned right above 20kHz + the subtle phase shift of the analogue Nyquist filter (in fact, it's just its "tail", the analogue Nyquist filter operates at a very high frequency, far outside human hearing limits).

So far, no surprises.

All three files have almost the same impulse responses and thus almost exactly the same frequency and phase response. The differences are extremely small and have typical white-noise properties (i.e. it is most probably noise).

All three cables have almost absolute transparent linear behaviour, if any!

Sine-sweep and IMD tests show absolutely no measurable traces of non-linearity (just open a spectrum analyser, set it to a large window size and look for harmonics/IMG products). Even the most minimal, inaudible form of non-linearity becomes very apparent on the scope, but these 3 files show nothing! (A good sign for Benchmark and a very bad one for cable voodoo)

I can share the impulse response if anyone is interested.