“The Atlantis DAC is the result of our most super sophisticated circuitry and feedforward algorithms. A lot more of complexity that also delivered a lot more of performance, not merely a marginal improvement. During the design phase of our new discrete output stage, we obtained some crazy simulation results that yet we have not been able to measure. Two days ago we measured the total audio output with the latest Audio Precision ApX555, currently the World’s most accurate audio measuring device. Our simulated result is actually below the background noise of this instrument, so we cannot really tell where are we. See the picture. First harmonics are at -180dB. So we now know we are at least as low as the lowest noise floor of all available measurement instruments available at any price. We are working on our own measurement device and flow, using advanced statistics and crosscorrelation between different reference set vs the device under test (The Atlantis DAC) to dig deeper into the noise floor.
Beside the singular measurements of the Atlantis DAC measurements are singular, at the same time, I would like to bring up some experiment and experience around this.
On purpose I altered some aspects of the circuitry, minor ones, plus some other changes in the implementation. Curiously, all these changes are not revealed in the typical set of engineering measurements. Of course, all measurement iterations were done following the same measurement procedure. A scientific method. Apparently no differences on measurements were found. However, from a musical perspective, the differences are huge. The capability to create and deliver emotion completely shifted.
Advancing deeper, I created a listener panel, without explaining what was going on. I asked for an A-B preference, asking why they thought it was better. Results written on paper. 4 tracks of 2 minutes, played over 3 times to verify results were consistent. To my surprise.. or not.. the results were deterministic and reliable.
Everyone in the panel agreed on ‘musical playback’ with my chose and the ‘proper’ set of values and settings. But the measurement set is crazy on both.
As I see it, there is something more than just crazy measurements involved in recreating music in its deepest implications. I think that BOTH are needed, or otherwise, we have only part of the solution. An incomplete solution. I remember doing an experiment with my father around 1990 or so. We bought a first generation CD player. One dac multiplexed for 2 channels, 14 bit ENOB (effective number of bits), brickwall filtering, no oversampling. You could hear things that certainly should not be there, but besides this, there was something peculiar on that product. The long term listening was pleasant. Harmonic coherence and structure was good!. We learned a lot from that experience.
As the second derivative of this, there are many audio devices in the industry whose effects rarely can be measured using proper engineering practices. But many of them obviously make a diference and improve the results. Shall we discard something we cannot measure –yet- but that improve our musical experience, that brings us closer into the stage? As far as I am concerned, we should remain open minded, and being thankful and respectful to all those who work applying different mind processes that come down to valid solutions. And if they can prove it using a scientific flow: repetitive experimentation should deliver same positive results.
And, of course, no results.. no qualify” - Javier, Wadax.