5.2.2 Speech to Text
Feature Introduction
This section introduces the basic functionality of Automatic Speech Recognition (ASR) and its example usage. After users input voice through a microphone, the system can automatically recognize and convert it to text.
Project repository: ⭐ Bianbu AI Demo Zoo | NLP
Preparation
Clone Code
git clone https://gitee.com/bianbu/spacemit-demo.git
cd spacemit_demo/examples/NLP
Install Environment Dependencies
It is recommended to use a virtual environment for dependency isolation:
sudo apt install python3-venv
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
Detect System Recording Devices
Refer to the Recording Device Detection section to check the available recording devices in the system.
Execute Example Code
Run the ASR example:
python 03_asr_demo.py
After the program starts, press Enter to begin recording. The integrated VAD functionality will automatically determine if there is human speech and stop recording during silence.
Parameter Description
| Parameter Name | Description | Usage |
|---|---|---|
sld | Silence Duration Threshold (seconds) | Continuous silence time ≥ sld seconds will be considered as end of speech; set to 0 to disable |
max_time | Maximum Recording Time (seconds) | Recording will automatically terminate when this duration is reached to avoid excessively long speech |
channels | Audio Channel Count | Usually set to 1 (mono), mono input is recommended for speech recognition |
rate | Sample Rate (Hz) | Number of samples per second, such as 16000 or 48000, must match model input |
device_index | Input Device Index | Specify recording device, can be obtained through arecord or search_device.py |