Microsoft has unveiled a Speech Devices SDK that allows manufacturers of microphone-enabled devices to integrate their devices with the cloud-based Microsoft Speech service.
The SDK consists of a pre-tuned library that is paired with a microphone-enabled device. It takes advantage of multi-channel audio and is capable of noise suppression, echo cancellation, far-field voice and beamforming. All of these capabilities allow the Speech Devices SDK to support "a variety of voice-enabled scenarios, such as drive-thru ordering systems, in-car or in-home assistants, and smart speakers."
Microsoft points out that the Speech Devices SDK can be used to build ambient devices that activate when a custom wake word is detected. Such devices, which include Amazon's Echo devices and Google Home, are increasingly popular with consumers and in creating the Speech Devices SDK, it is clear that Microsoft is aiming to gain a foothold in this market by helping power the next-generation of speech-enabled devices.
Initially, Microsoft has partnered with ROOBO, an AI solutions provider, which is offering end-to-end system reference designs for smart devices that are integrated with Microsoft Speech using the Speech Devices SDK. These system reference designs are intended to help smart device creators more rapidly develop their devices.