The first and grandest innovation hides in the processor. RixVox Audio Streamer Amadeus employs 1 GHz dual-core Cortex-A7 processor, that operates asynchronously. The tasks that the processor executes are not shared between the cores, which is the conventional approach, but veritably split between them, according to function. One of the cores manages all of the board’s periphery and the other core takes care exclusively of digital audio data. RixVox team succeeded in reprogramming the processor on low level. Not a small feat in itself, as the IT wisdom said – it could not be done, but most importantly splitting the processor core functionality allows to avoid the whole class of errors that arise from processor switching and sharing tasks between cores (re-prioritisation).

Another class of distortions and errors in the digital signal path arise from improper clocking and bus clocking conflicts. These are the main sources of “jitter” in digital audio, which means bits of data being served with the wrong timing. To counter it, RixVox Audio Streamer Amadeus has a sub-board of a separate set of clocks, running at 44.1 kHz to 352.8 kHz and 48 kHz to 384 kHz respectively. These ultra-precise clocks (with errors of less than 1 ppm) govern everything concerning digital audio data. Thereafter bit-perfection of the streamed data is guaranteed by ultra–precise external clocking. This is the second innovation*.
* for even more demanding customers RixVox Audio Streamer Amadeus has the special option of connecting a separate atomic clock with clocking errors far below the measurable level.

Separate stabilised power supplies for each of the board’s main components is the third innovation. It addresses the class of distortions and errors that arise from parts of the board being momentarily under-powered. In digital as in analog circuits “under-powered” typically means voltage drop. Where analog circuits tolerate voltage drops rather well, digital circuits are highly intolerant to any voltage drop, they may generate switching and calculation errors, hang-up, or initiate restarts. The common solution of just taking large enough power supply does not address this properly, as it creates its own distortions.

CONVERT FROM TWO INPUTS TO ONE OUTPUT.

Specification:
> Sample Rates: 44.1 to 384kHz;
> Input interface 1: I2S (TTL/CMOS);
> Input interface 2: I2S (differental);
> Output interface: I2S (TTL/CMOS);
> Power supply: +5VDC/100 mA;
> Board dimensions: 50 x 42 x 10mm.

CONVERT FROM DIFFERENTIAL I2S TO I2S (TTL/CMOS).

Specification:
> Sample Rates: 44.1 to 384kHz;
> Input interface: I2S (differental);
> Output interface: I2S (TTL/CMOS);
> Power supply: +5VDC/100 mA;
> Board dimensions: 70 x 30 x 8mm.

CONVERT FROM I2S (TTL/CMOS) TO DIFFERENTIAL I2S.

Specification:
> Sample Rates: 44.1 to 384kHz;
> Input interface: I2S (TTL/CMOS);
> Output interface: I2S (differental);
> Power supply: +5VDC/100 mA;
> Board dimensions: 60 x 30 x 8mm.