It has been known for a long time that Sonos is working with high pressure to make the Multiroom speakers fit for the language control by Amazon Alexa and Co. About a year and a half ago the founder and then CEO John MacFarlane wrote an article in which he wrote On the one hand, identified the streaming as the future of the music consumption, and on the other hand showed itself enthusiastic about Amazons echo speakers and the connected speech control.
The Play: 5 speakers from Sonos was then the first product of the manufacturer, which was equipped with two microphones, but remained silent until further notice. Later it was said that the microphones were primarily intended to enable the truplay technology that the loudspeakers measures and thus optimizes the sound to the location of the terminal. However, since this is done with an iOS app (the Android version is still waiting for you), you should rather rely on the microphones of iPhone and iPad. Ergo was the new Playbase without the integrated microphones and concentrated on the development of a skill for Amazons Alexa. Their start is now imminent. Sonos has changed privacy regulations and explicitly mentions Amazon and Alexa. Sonos explains:
Most importantly for you is that Sonos does not record your voice data. The data goes directly to the provider of the speech control service (eg Amazon), which you have activated on your Sonos system. We have also included information on how Trueplay technology analyzes your space to optimize the sound quality of your product. (…) We only pass on data to our partners, which are necessary to provide the service on Sonos at all and in the best possible quality. Data will only be disclosed in connection with products or services that you have requested or authorized.
In concrete terms, this means:
That your voice-supported Sonos product buffers and recolves the ambient noise, but at a local level without transmitting or storing this information until it detects a word or phrase that causes the device to start an active recording. If the product does not recognize the wake-up command, the recording is repeated in endless loop every few seconds and the previous recording is overwritten. These operations are only performed locally on your device and are not sent to Sonos or third parties. When the wake-up command is detected, the product begins recording. This means that the product does not record, store or transmit any audio data before “waking up”, a visual signal, such as a light on the product, indicates a running recording The voice command will be sent to the partner for voice control that you have authorized to receive such recordings (for example, Amazon), and Sonos will not record the voice recording.
This is initially pleasingly transparent. But also interesting in other matters. Because so far, Sonos products beyond small status LEDs have no way to display visual feedback. The wording “such as a light on the product” could be an indication that there will soon be new additions in the loudspeaker family which can precisely translate the activation of the voice control into a clearly visible manner. Wait. First, the Alexa-Skill must be published and tested for practicality.