Voice support by BLE statck

[wang]

Dear Nordic Team,

I want to know if we can support the real time talking like phone call between peripheral and central by software device. On application layer, we just enable the this function, voice from mic-->PDM-->PCM-->software device(BLE controller) can get data from PCM buffer and send out to remote devices over air. Receiver --> get data from air-->put into PCM-->speaker; Should disable the retransmission on BLE controller layer.

Is this possible supported on NRF52832?

Thanks for your support.

[endnode]

Yes, people managed to transfer audio/sound over BLE link with nRF5x devices so it is possible. However there is no standard profile for audio transfer over BLE by BT SIG (yet) so there is no such example in Nordic SDK. If you search the forum you will see many thoughts and also maybe few examples on this topic, try it.

[wang]

Dear Nodic Team, I have see some thoughts on the forum based on nordic board, but mostly they just take ble as a transfer tools, data handle in app layer like this: mic-->PDM-->PCM-->BLE app profile-->host stack-->controller--> air to remote receiver --> get data from air-->past to host layer-->past to app layer-->PCM-->speaker In this have two problem, firstly, data double handle by app and host layer, it is ineffective, secondly, for real time talking, we need make the ble link as SCO, discard some packets if environment is not good, this is controlled in software devices that we can't change it... Do nordic have any solution based on software device change to support this feature, just no plan to support this in software devices and just let the customer change the application to make it. Thanks a lot.

[endnode]

But what you describe is completely NOT use case for BLE! Why you don't go with custom HW+SW on both sides of the link and then you can use custom radio protocol/stack with all these features such as "QoS" and flushing the queue if the throughput is lower then required etc.?

[wang]

Hi, So appricate for your comments. Yes, it not use case for BLE, I just talk about the possibility of BLE voice, it is not a requirment, just disscuss on technologic. For my personal opinion, it more make sence to do it in controller than, but as you say, it not in BLE spec, most real time talking scenario using the be/edr SCO. Thanks again.

[endnode]

I'm afraid from BLE point of view it doesn't make sense (to built it some specific privileged profile/use case lower to the stack, that's not how BLE was built).