STM32 Audio and Voice Solutions
STMicroelectronics has been providing solutions for voice recognition and audio processing for years, such as X-CUBE-AUDIO and the Alexa for IoT reference solution, and has built relationships and partnerships with third parties who are leaders in their fields.
STM32 local voice UI solutions
The STM32H5 and STM32H7 families offer solutions for local Voice User Interface (VUI). Here, “local” means “on the edge”, as the speech recognition is performed directly on the STM32 microcontrollers, without any connection to an external host or cloud.
These local voice reference designs integrate Sensory's, advanced technologies to offer the most intuitive and customizable offline device interactions.
The STM32H573-DK and STM32H747-DISCO boards are supported by X-CUBE-LocalVUI and its LocalVUI-Denoised variant to support different solutions.
In both cases, all the audio and voice related software is stored in the embedded flash, ensuring maximal performances with no cache miss impact.
Some examples also use the board’s external memory to store graphical objects.
X-CUBE-LocalVUI includes speech recognition application examples and supports both keyword spotting and intent understanding.
This package supports STM32H573-DK and STM32H747-DISCO boards, utilizing the on-board microphone for audio input.
Designed for use in low-noise environments, X-CUBE-LocalVUI allows for the easy prototyping of custom vocabulary. It is available for download on st.com.
On top of X-CUBE-LocalVUI, LocalVUI Denoised integrates an advanced audio front-end for denoising, capable of handling both low and high levels of noise.
This package supports the STM32H573-DK and STM32H747-DISCO boards, along with the STEVAL-MIC008A microphone extension board. Depending on the other features managed by the application, the keyword spotting solution can be implemented on various STM32 families.
Snapshot of LocalVUI denoised fan demo on STM32H747I-DISCO and STM32H573I-DK
More advanced voice user interface solutions can be used to implement thousands of variants of request forms with Sensory’s TrulyNatural (TNL) technology.
Snapshot of the coffee machine model from LocalVUI Denoised package
Other STM32 microcontrollers can be chosen from the H5 and H7 families, and potentially others depending on the use case.
The STM32MPU families can also support advanced voice recognition.
STM32 audio processing solutions
Partnership with DSP Concepts
ST has had a partnership with DSP Concepts for years, providing audio output processing solutions and voice recognition solutions.
DSP Concepts, is a world-class leader in audio processing and offers solutions on STM32.
X-CUBE-AUDIO-KIT
X-CUBE-AUDIO-Kit offers a royalty-free solution to design, implement, and fine-tune an audio processing flow on supported series of STM32 microcontrollers.
This package eases and accelerates the integration of use cases such as acoustic processing for voice communication or speech recognition, audio conditioning, audio effects, audio output enhancement and any other audio processing use case.
- It is an audio processing ecosystem that integrates:
- A tool to design and tune the audio processing dataflow on STM32
- A growing number of algorithms and examples
- An execution framework
snapshot of X-CUBE-AUDIO-KIT LiveTune
Key Products
| Part Number | Description |
| STM32H5 | High-performance MCUs |
| STM32H7 | High-performance MCUs |
| STM32MP1 | STM32MP1 microprocessor series with dual Arm Cortex-A7 and Cortex-M4 Cores |
| Part Number | Description |
| MP23DB01HP | MEMS audio sensor multi performance mode digital microphone |
| STEVAL-MIC008A | Dual MP23DB01HP MEMS mic daughter board |
| Part Number | Description |
| FDA90xx | Class D digital input automotive power amplifier with load current monitoring, wide voltage operation range for car audio and telematics – slug down version |
