Table 5. Full list of files in HyperFlash filesystem (continued)
Name
Start address
End address
Description
Bank A Signing Certificate
0x61D0_0000
0x61D3_FFFF
Used for validating the signature of Bank A
Bank B Signing Certificate
0x61D4_0000
0x61D7_FFFF
Used for validating the signature of Bank B which
is written during first OTA
Bootloader Signing Certificate
0x61D8_0000
0x61DB_FFFF
Use for validating the signature of bootloader
3.5 Audio application architecture
The audio capture application is implemented as a pipeline shown in
. The application can be configured to capture either
two or three microphones (two by default). Because the i.MX RT106S device has enough processing power, the core is used to
process the entire chain from the PDM decimation to the Automatic Speech Recognition (ASR) engine. Every 10 milliseconds,
the DMA moves raw PDM data from each microphone. This data is fed into the NXP Solutions PDM decimation software IP to
convert the audio into 16-bit, 16-kHz PCM data. When it comes out of the decimation block, it is fed to the Audio Front End (AFE)
to perform beamforming and acoustic echo cancellation. At this point, it is a single 16-bit, 16-kHz mono audio signal.
Figure 4. Audio application pipeline
Although the ASR works on multiples of 10 ms of audio data, a 30-ms data block is recommended for the input. Thus, the
audio_processing_task
accumulates 30 ms worth of processed audio before sending it to the ASR for processing.
3.6 ASR application
The AFE output signal is transferred to the ASR, where the wake word engine waits for a wake word. If a wake word is detected,
the same language’s command engine is loaded to process the voice control commands. Developers can also implement multiple
groups of commands sequentially to create a dialog-style voice-control application.
NXP implemented the following three types of baseline demos:
• LED voice control demo
— English
— Two-stage (wake word and command) ASR
• IoT/elevator/audio/washing machine voice control demo
— Selectable combinations of English, Chinese, German, and French
— Two-stage ASR
• Oven voice control demo
NXP Semiconductors
Introduction
SLN-LOCAL2-IOT Developer’s Guide, Rev. 0, 19 April 2021
User's Guide
15 / 87
