Streaming has become the dominant source format for most listeners, and with it has come a new set of variables — network architecture, streaming service quality, file format, and clocking at the receiving device — that determine the quality of what arrives at the DAC. Understanding these variables allows the streaming source to be optimised rather than simply accepted as given.
The first and most fundamental variable is the quality of the audio data being streamed. Streaming services vary significantly in the bitrate and format of the audio they deliver. Lossy compression formats — AAC and MP3 at various bitrates — discard audio data that psychoacoustic models predict the listener will not notice. At high bitrates (320 kbps and above), the losses are small and often inaudible on typical systems. At lower bitrates, compression artefacts — a characteristic smoothing of transients, a loss of fine spatial detail, and a slightly processed quality to the sound — can be identified on resolving systems.
Lossless streaming — CD-quality FLAC or ALAC at 16-bit/44.1 kHz — retains all the data of the original recording without compression. High-resolution lossless streaming at 24-bit and sample rates above 44.1 kHz carries additional data beyond what the CD format contains. Whether this additional resolution is audible under realistic listening conditions is a subject of ongoing research, but the engineering argument for at least the lossless base level is clear: retaining rather than discarding data is the correct starting point.
The network environment introduces variables that are specific to streaming and have no analogue in physical disc playback. Network noise — the electrical interference associated with Ethernet and Wi-Fi communication — can enter the audio signal path through the streaming device’s power supply and data connections. Galvanic isolation between the network connection and the audio circuitry, or the use of a dedicated network switch with optical isolation, reduces this contamination.
Clocking at the DAC is as significant in streaming as in any other digital source. When a streaming device delivers data to an external DAC, the timing of that data delivery is subject to network jitter — variations introduced by the packet-based nature of network communication. A DAC with a large input buffer and a high-quality local clock can reclock the incoming data independently of the network timing, eliminating network jitter from the conversion process. This architecture — buffered reclocking at the DAC — is one of the most important features to look for in a streaming-capable DAC.