The Challenges of Digital Audio (Part 4)
By Alan Ruberg, Systems Architect
This is Part 4 of a 4 part series; here is Part 1.
In my last blog post (Part 3), I discussed the move from 2.4 GHz to the 5 GHz U-NII band. The image below shows wireless traffic at a trade show. A 7.1 speaker system using WiSA-compliant technology inhabits channel 140. This entire 5GHz band, although set aside for wireless networks, still conflicts with government and weather radar, to which interference is prohibited. This is easy enough at channels 36-48 (low power) and 149 to 165 (ISM band), but radar monitoring and Dynamic Frequency Selection (DFS) is required on channels 52 through 140 to avoid conflicts.
Simply put, if a radar signature is recognized, then network communications must move immediately to a different frequency. In order to detect a radar signal, the entire network must be quiet for a period of time or the system will “blind” itself. Since radar signals are pulses, monitoring needs to be repeated frequently because a single sample is likely to miss detection.
Before starting to use a DFS frequency, it must be clear of radar for one minute. When a radar conflict is detected, the frequency cannot be re-used for 30 minutes. The system can either switch immediately to a non-DFS frequency and/or it has to monitor a new DFS frequency for one minute before using it. In a computer network, this process can be quite cumbersome with all of the interruptions and waiting around. IEEE Standard 802.11h specifies this process for WiFi networks.
A WiSA-compliant transmitter deals with DFS requirements quite naturally. I mention in Part 2 that network audio is transmitted in packets. Uncompressed audio, as used in WiSA-approved technology sends these packets at a constant rate. Since not all of the bandwidth is used, radio monitoring can be hidden between packets and has no impact on performance (space, speed, or time) as shown below.
As for the one minute rule, the technology has an extra receiver that monitors the next channel so that a new DFS channel is known by the entire network (transmitters and receivers). If a channel change is required, then the switch to the next DFS channel can be performed without losing a single audio sample. This way, the audio network remains free of congestion. As a bonus, any type of interference, not just radar, is avoided by the technology to provide the best quality. There are plenty of DFS channels, so there is room for multiple systems. If convenient or necessary, the system may still use non-DFS frequencies; however DFS rules are still used to achieve interference-free operation.
Wireless audio must sound great and be robust. It must sound as good as or better than speaker wire. WiSA-certified technology achieves this by making the right assumptions for audio, not for computer networking. The first, and best, line of defense is picking a channel free of interference. This is done by aggressively operating in the U-NII, and specifically, in the DFS band. The second line of defense is when error correction and concealment are used to maintain a great sound when packet loss inevitably occurs.