Where Wi-Fi and Hi-Fidelity A/V collide

High fidelity Audio/Video (AV) over Wi-Fi is extraordinarily difficult to achieve. Current technical limitations must yet be overcome. This post explains some of the obstacles, and touches on how embedUR is applying its considerable Wi-Fi networking expertise to overcome them.

The case for Wi-Fi in high-fidelity AV systems

The advantage of wireless is convenience. Traditional analog audio systems require only one cable from the source to the output, but they are difficult (read “inconvenient”) to install. Running wire neatly along complicated ceiling structures or inside finished walls can get expensive and takes many hours.

Standards and solutions have been recently developed for supporting Hi-End AV multi-channel content distribution over wired Ethernet networks, based on Ethernet AVB (Audio Video Bridging) and TSN protocols, to maintain quality metrics based on low loss and low latency.

Moving to an Ethernet AVB solution helps mitigate the analog problem to some extent by moving away from custom wiring towards standard Ethernet wiring. But it’s still not as convenient as Wi-Fi. Wireless speakers are vastly more convenient to install. Paradoxically they too, require a wire, but only to the nearest power outlet. Hence, installation of a wireless speaker takes only minutes.

This makes high-quality audio devices clear candidates for wireless adoption. Micro-sized computing devices, RF innovation, and cloud computing suggest epic opportunity – and epic disruption. Although low-quality wireless audio using Bluetooth has been around for years, high-quality wireless audio has been elusive. Consider the current technology landscape.

Wi-Fi beats Bluetooth hands down for Hi-Fi grade A/V

Bluetooth has become popular – and it’s not going away soon – for its low-power applications in small portable speakers, earbuds and wearables. But its limitations keep it in the low-end of the audio market. For stable connectivity, Bluetooth requires devices be near each other with little chance of interference.

The slightest disruption to the connection, such as when a person stands between the transmitter and the speaker, can introduce a notable disruption. As a sole service for high-fidelity wireless audio, Bluetooth is sub-optimal. In addition, for technical reasons, Bluetooth’s inability to multi-cast makes it ill-suited to multi-channel distributed systems, which is the most common use-case for streaming wireless audio.

Wi-Fi offers greater promise. Wi-Fi can carry data for longer distances and with much greater throughput. No doubt it is for these reasons (and others) that it is the platform of choice for Apple’s AirPlay and Google’s Cast.

Wi-Fi still has technical limitations for audio streaming

That said, current technical limitations make Wi-Fi anything but a slam-dunk for audiophiles and their brands. Analog engineers solved their audio distribution problems years ago. They set the system up to succeed. There are very few problems that arise with this very simple setup:

  • A single distribution point (the amplifier) distributes signals to many outputs over simple wire that carries both power and analog signals.
  • The entire system is fixed and the medium is dedicated, and therefore static. Nothing changes, so there is little risk to system integrity.
  • High-fidelity depends less on the connection and more on the endpoints.
  • Since the signal travels at near the speed of light, the only potential problem is that one or more wires are much longer than others, and in such rare cases, they engineer a signal delay to the nearer ones so that the outputs come out at the same time.

Wireless’ convenience comes with far more variables (read “room for problems”) and for many of the challenges there is no good solution… yet.

  • High end multi-channel wireless audio is a multicast data stream. One or more inputs are broadcast to multiple outputs (speakers) simultaneously.
  • Because the input or source can be mobile the input connection is also wireless. Each movement of the input triggers a recalculation of data rates based on the environmental conditions facing the input device.
  • As the wireless medium is shared and often noisy, wireless high-fidelity audio depends more on signal quality than endpoints.
  • Risks come from multiplying several connectivity factors times the number of devices.

Managing distance and timing factors

Wireless systems must dynamically manage distance and timing for multiple, independent sources and receivers! To serve high-fidelity AV applications, a wireless system needs to maintain very high connectivity and synchronization performance for each device. And because of the nature of wireless, these factors must be managed discretely, for each device has its own environmental factors.

Connectivity challenges: Connectivity is core to Wi-Fi functionality. Consider the variables that could degrade the premium, high-quality AV experience, such as bandwidth, latency, and jitter.

  • Since Wi-Fi is a shared medium, other Wi-Fi usage introduces risk of congestion, and neighbors and non-Wi-Fi devices (including Bluetooth and Cellular repeaters) can cause interference, all of which lead to collision and retransmission problems. Already we have raised the specter of a poor high-quality audio experience, and we haven’t even addressed the impact perimeter devices will have on Wi-Fi data frame rates.
  • Low latency and jitter rates are critical to Hi-Fi sound output because audio data, more than other streams, is time sensitive. Audio outputs establish rhythm; poor latency can kill it. Likewise, low latency is essential to avoid video / audio lip-sync problems. Multi-room configurations need speedy data rates, but walls, doors and RF interference put real-world constraints to bandwidth, and cause higher latency and higher jitter. Not only that, but congestion from other Wi-Fi users, improper queuing or playback buffer management can cause rapid latency degradation. Some premium audio technology companies stipulate that the latency must not exceed 20ms.
  • Current Wi-Fi protocols introduce buffers to solve the common latency and jitter brought by connectivity problems. Making the playback buffer larger isn’t a good solution for two reasons. It introduces design challenges because a larger playback buffer requires more storage, or more storage at the cost of other embedded resources, and perhaps more important to high-quality audio, it makes for a poor user experience because it introduces longer delays between when playback is initiated and when it is heard. This is particularly impactful for multi-channel streams, as latency and synchronization across multiple devices needs to be maintained.
  • Packet loss is also common to Wi-Fi networks, and a severe problem to audio data transmission. Packet rebroadcasts will probably happen too late for audio output. The technical achievement of an eventual packet delivery takes a poor second to the lost sound beat.

Synchronization challenges: While analog speakers require little if any synchronization, wireless devices require persistent synchronization, demanding extensive computational resources. Speakers must have a control mechanism to begin playback of the same data frame at the same time. If they don’t, the user experience is poor. The sound source is unclear, echoes reverberate around the room, and sound fidelity is muddy.

Using the current standard, Network Time Protocol (NTP) isn’t good enough. For one thing typical implementations are simply not granular enough and are accurate to only milliseconds. For another, NTP traffic can overload a network already burdened by high bandwidth and connectivity demands. In addition, the computational resources needed to synchronize many devices in real-time exceed those of most wireless devices.

Mobility challenges: Most of the challenges discussed so far have been about receiving a signal at the speakers. But what about the input? This is another rather big challenge, and the one on which embedUR is focused. For example, say that a wireless audio stream is initiated to all speakers in “Party Mode.” In this scenario, there is a multicast stream from the Wireless AP to all the speakers. But what if the source is a mobile phone, and the user is roaming around the building.

Now the system must also preserve the multicast session even as the client roams from one AP to another which takes about 50ms - best case - with fast roaming (.11r) if that feature is even implemented. Many low-end home Wi-Fi mesh solutions do not support it. You can see, this is starting to get rather complicated. Along with all the other variables relating to input signal quality, one must also deal with 50ms or more of “radio silence” as the client generating the input signal roams from one AP to the next.

What Wi-Fi needs for Hi-Fi is connectivity innovation

Some of these challenges can be overcome (and have been) in small residential “Lo-Fi” applications simply by ignoring them! But for product designers and audio engineers pursuing wireless Hi-Fi, there is an urgent need for better solutions. The standards and solutions for Hi-Fi AV applications that have been developed for the Ethernet AVB domain need to be re-engineered and reapplied for the Wireless domain.

No single technology or platform currently exists that fully solves the modern problem of delivering Hi-Fi multi-channel audio/video content over Wi-Fi. But this is soon to change. Along with a few others, embedUR is applying its extensive experience to these challenges, in an effort to meet the market’s demand for wireless audio convenience without sacrificing audio quality.

What does this have to do with embedUR?

For more than 12 years what has got our ~120 Wi-Fi engineers up in the morning, has been to solve some of Wi-Fi’s toughest problems, such as these outlined above. Today we are working with a leading Wi-Fi chipset provider specifically to address the issues related to supporting low-loss and low-latency multi-channel AV content across a system of wireless sources and destination devices. This involves extending and adapting the same AVB and TSN standards and protocols being used on Ethernet, to now work over the Wi-Fi medium in order to maintain quality metrics based on low loss and low latency.

Talk to embedUR about your Hi-Fi and Wi-Fi product innovation challenges.