Record Voice Recognition Module DD501 【RP4】


1. Specification

Based on the neural network processor NPU, the module realizes off-line local speech recognition through off-line neural network algorithm.Using this module for off-line speech recognition does not depend on the network, has small delay and high performance, can achieve a high recognition rate of more than 95%, the recognition distance can be up to 5-10 meters, and the shortest response time is 0.2 seconds.



Parameter item

Detailed description


Identification type

Local speech recognition(Offline , without network)


Recognition distance

Effective distance 5 ~ 10m in a quiet environment


Recognition accuracy

Recognition accuracy > 95% in 5m distance


Response speed

< 0.2 seconds


Sound input

1-channel microphone input  


Working parameters

DC 3.6V-5.0V, current < 50mA


Work environment

-40~85℃,0~85%RH


Pad Port

Pin hole, pad spacing 2.0mm


Product size

18x13x3mm





2. Record other Command (unlimited language) to control

If you want to use voice for control, you need to first record your voice(Can be in any language)

Using voice to control IO output requires recording voice before use.

The operating steps are as follows:

(1)Power the product with a DC power supply of 3.6-5.0v and connect a microphone

Connect a button to GND through pin P5;

Single Click the key: Record to the command in order;

Press and hold for 6 seconds to delete the recorded command。

(2) Recording instructions:

The LED indicator light ned to be connected to the P1 pin, and press the key(connect to Pin P5)to record 4 instructions in order ,

When recording the first instruction, the light flashes once,

When recording the second instruction,the light flashes twice ,

When recording the third instruction, the light flashes three times,

When recording the fourth instruction, the light flashes four times,

After successful recording, the light flashes twice.


(3)Output signal: (Standby status: High-level)

After recognizing the first recorded instruction, P1 outputs a 100ms low-level pulse,

After recognizing the second recorded instruction, P2 outputs a 100ms low-level pulse,

After recognizing the third recorded instruction, P3 outputs a 100ms low-level pulse,

After recognizing the fourth recorded instruction, P4 outputs a 100ms low-level pulse,


Attention: 2~4 words are the best record command, command word duration less than 3s.








3. Wiring instructions:

VCC and GND are connected to 3.6-5V power supply,

MIC+and MIC - are connected to the positive and negative electrodes of the microphone, respectively,

Control output using pins P1, P2, P3, P4; and the LED indicator light need to be connected to the P1 pin

Pin P5: Button input pin, used for recording (effective at low level).

Based on the neural network processor NPU, the module realizes off-line local speech recognition through off-line neural network algorithm.Using this module for off-line speech recognition does not depend on the network, has small delay and high performance, can achieve a high recognition rate of more than 95%, the recognition distance can be up to 5-10 meters, and the shortest response time is 0.2 seconds. If you want to use voice for control, you need to first record your voice(Can be in any language) The LED indicator light ned to be connected to the P1 pin, and press the key(connect to Pin P5)to record 4 instructions in order , MIC+and MIC - are connected to the positive and negative electrodes of the microphone, respectively, Control output using pins P1, P2, P3, P4; and the LED indic