Print This Page

6. Distribution & DSP issues

The core functionality of a networked audio system is determined by the audio network protocol and its hardware to take care of the distribution of audio signals throughout the system, and Digital Signal Processors (DSP) to process (change) the audio signals. This chapter focuses on the core distribution and DSP functionality.

The sound quality of a networked audio system is not affected by the distribution system. A properly set up networked (or any digital) connection will transfer digital audio signals without changing the samples - the only function of a distribution system is to distribute, and not to process. The DSP hardware itself also does not affect the sound quality of a networked audio system - the internal processes in a DSP core can be regarded as distribution processes - moving samples to and from the DSP’s memory.

The DSP algorithms however do affect sound quality by definition - as the default status of a DSP algorithm is to just pass audio information from input to output without any change. Every change a DSP algorithm poses to an audio signal is therefore intended - it’s the most important part of a digital audio system’s Response.

The audio quality of a networked audio system on the other hand is not affected by the DSP algorithm - as all DSP algorithms are fully intended. Instead, only the bit depth and the sample rate of the distribution system and DSP hardware architecture affect the Performance of a system’s core functionality.

6.1 I/O distribution

If the bit depth of a networked audio system’s A/D and D/A converters is 24-bits, then of course the connection to and from the network also has to be at least 24-bits to preserve the 24-bit resolution. Most conventional digital audio formats such as AES3 (AES/EBU), AES10 (MADI), and Ethernet compliant audio network protocols such as EtherSound and CobraNet are 24-bit digital audio formats. All these formats are linear - they transport the digital audio samples without changing them. All digital audio formats therefore ‘sound’ the same. However, formats do differ in latency - with the Ethernet compliant audio protocols such as EtherSound, CobraNet and Dante posing a higher latency than the conventional point-to-point connection formats such as AES3 and AES10. The implications of latency differences in correlated signal paths are described in detail in chapter 5.6.

Return to Top