Frequently asked questions

Why is having a low and deterministic latency important in Live Sound?

SuperMAC and HyperMAC provide extremely low-latency transmission paths – 62.50 μs and 41.66 μs per link respectively at 96 kHz operation. This is possible because of the TDM nature of the formats, and cannot be achieved by any of the IP-based audio transmission systems for an equally large number of audio channels. Why is this so important, when link latencies of 1 ms or so are typically regarded as “acceptable”? The problem comes when the user starts to build genuine networked systems as opposed to simple point-to-point digital snakes. If a Cat 5/Cat 5e audio network is used for ALL the connections in an audio system, as opposed to a single path, then the latencies start to add up pretty quickly.

Keeping transmission latency to an absolute minimum is critical with the increased adoption of in-ear monitoring, the delay for example that a vocalist experiences between singing a note and hearing it in their earpieces cannot be perceptibly greater than the time it takes for the sound to be physically transmitted from their mouth to their ears. Even for latencies below which a time delay cannot be consciously identified, the effect on artists can be very disturbing and their performance impaired. They will be aware that something is wrong even if they cannot identify the cause.

In a Midas XL8 network system, for example, the signal path might be microphone → splitter system → router → DSP → router → output box → in-ear monitor. This involves four SuperMAC network links, so including the additional samples required for converting the received data into an internal data format such as I²S or similar:-

Splitter System → Router	6 + 2 samples	83.33 μs
Router → DSP	6 + 2 samples	83.33 μs
DSP → Router	6 + 2 Samples	83.33 μs
Router → Output Box	6 + 2 samples	83.33 μs
	Total	333.33 μs

In-ear monitoring is about the most latency-critical application in live sound applications, and to minimise the additional system latencies Midas digital consoles use premium low latency analogue-to-digital and digital-to analogue converters that add a further 141.67 μs combined latency. The internal processing latency from input to auxiliary output on an XL8 is 106 samples, or 1.10 ms. The total system latency in a Midas XL8 digital console system from microphone to in-ear monitor is less than 1.60 ms.

If a network was used with 1 ms latency per link, the cumulative network latency would be 4 ms, and again the internal processing latency and the combined converter latency also needs to be added to this figure to give an overall system latency – the combined converter latency can be in the region of 2 ms at 48 kHz operation (as used by the majority of digital consoles in live sound applications) for many industry-standard converters. The resulting total system latency would probably be unacceptable for in-ear monitoring.

The importance of having a completely deterministic transmission latency is that precise phase alignment between many different channels can be guaranteed, so that when a large number of microphones are being used on stage in close proximity to each other, such as on a drum kit where a certain amount of spill is inevitable, a clear and coherent stereo image is presented in the Front Of House mix that the audience hears.

< Back to frequently asked questions