BLE Audio – the real state of play
Anyone following the development of Bluetooth Low Energy will have heard about Bluetooth Low Energy Audio. But what is it, exactly?
Is it the same as Bluetooth 5.2? And does that mean Bluetooth 5.2 compliant devices support BLE Audio ? There are a lot of misconceptions around what BLE Audio is, and what is available today. This article by Dr Nick Wood, Director, Sales & Marketing, Insight SiP explains the current level of progress in the technology, and what is needed to implement it successfully.
Read the article on Electronic Specifier
Bluetooth “Classic” Audio and BLE
Let’s start with why Audio over BLE has not been available so far. Bluetooth Low Energy was conceived to meet IOT use cases where there was relatively infrequent data transfer, but a requirement for low power. “Classic” Bluetooth didn’t offer an optimal solution to these as it required a continuous connection. Because of this, devices had to be in the “Radio On” state whilst connected. BLE allows devices to sleep between connections, therefore saving power.
Bringing Low Energy and Classic Bluetooth together
Clearly the Audio case is somewhat different to the above. You generally need a continuous and relatively high data rate to stream audio. So converging these two strands of Bluetooth was not straightforward and took a number of distinct steps which are outlined below.
For starters, you need to support streaming. This is implemented in the 5.2 revision of Bluetooth. One misconception is that Bluetooth 5.2 (BLE 5.2) is the same thing as BLE Audio. In fact, Bluetooth 5.2 provides an implementation of some necessary underlying technologies for Audio as well as some other new unrelated ones, but it is not in itself “BLE Audio”.
In a standard data transfer protocol, in the case of a packet failure, normally the transmission will be retried a few times until it’s successful. For a streaming protocol, what you want is that data is “time-stamped”, and any dropped packets should be simply discarded. This way, whilst one might experience the odd glitch in an audio stream, that stream won’t stutter and break timing. This is implemented in BLE 5.2, under the name of “Isochronous Channels”.
BLE 5.2 takes streaming one step further than Bluetooth classic, by introducing the ability to stream multiple synchronous channels. Current generation stereo Bluetooth headsets must stream the stereo signal to one earpiece, then send the relevant mono channel to the other one. This is inefficient and limiting. By contrast, under Bluetooth 5.2, it is possible to stream many channels from a single device, for instance to generate 5 channels “home cinema” surround sound.
Another aspect that needed to be tackled was the data rate limitations of BLE. Classic Bluetooth EDR (Enhanced Data Rate) offers theoretical speeds up to 3 Mbits/s. Bluetooth Low Energy is limited to 2 Mbs. As above, some design engineers are also looking to perhaps incorporate multiple channels. They need to bear in mind theoretical speeds are just that: in the real world environment of 2.4Ghz radio frequency signals existing around the body, actual speeds are likely to be much lower.
It is claimed that in blind listening tests, the LC3 offers better perceived audio quality than existing Codecs (for example, SBC, which stands for “Low-complexity subband codec) at half the data rate. So even at 192 kBit/s, it is possible to transmit high quality audio signals, better than a 384 kBit/s SBC stream. This represents a major advantage and allows high quality audio to be transmitted despite BLE’s more limited throughput. (Note that this codec not specifically part of BLE Audio, although it was clearly designed with it in mind. It can be used on any kind of audio transmission.)
That covers the core technology, but what about devices? There are BLE 5.2 compliant devices available now, so does that mean a user can buy one, plug it in, and start streaming sound? Unfortunately not! Firstly, being “Bluetooth 5.2-compliant” does not mean all Bluetooth 5.2 features are implemented. So check that the full set of features, in particular Isochronous Channels, are implemented and available.
Secondly, there has to be a LC3 implementation available. Note these are not part of BLE Audio as such. You also need to ensure that the device had sufficient power to run the codec. The technology is both memory and processor-hungry. So only the more powerful devices have the capacity to run them, or you may require a second processor/DSP (digital signal processor) to do so. The user must also run an independent codec concurrently for each channel for stereo sound.
Lastly, it is unlikely that your typical BLE device has an analogue-to-digital convertor of sufficient quality to capture and/or output an analogue signal suitable for high quality audio, let alone implement advanced features such as noise cancelling and so on. So in most cases, a high quality analogue-to-digital DSP is required to generate/process the digital signal carried across the Bluetooth link.
In practical terms, the latest generation of Bluetooth silicon – often with dual core, high-specification processors (such as the ARM M33 core), with large flash and RAM capacity target this kind of application. However, one has to be cautious about older generation devices that have been upgraded to “Bluetooth 5.2 compliance”. Many of these will remain insufficient for the task.
True state of play
So back to the core question: what is the true state of play in BLE Audio?
The standards are in place, and the new audio standard implements a clear improvement on the existing Bluetooth audio implementation with multi-channel streaming. A new codec offers high quality audio at lower BLE-friendly data rates.
Next generation, highly capable devices are available that have capacity to implement the core BLE Audio. However, additional components are still likely to be required. As ever, the last piece of the puzzle to fall in place are the software implementations. This will vary by vendor, but 2022 should be the year Bluetooth Low Energy Audio implementations become truly available.