Home
JAQForum Ver 24.01
Log In or Join  
Active Topics
Local Time 14:48 26 Nov 2024 Privacy Policy
Jump to

Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.

Forum Index : Microcontroller and PC projects : Help with voice recognition modules....

Author Message
Grogster

Admin Group

Joined: 31/12/2012
Location: New Zealand
Posts: 9306
Posted: 11:45pm 27 Nov 2023
Copy link to clipboard 
Print this post

Hi all.

I am looking for a programmable voice-recognition module, ideally with an on-board relay output that I can use to switch something off and on etc.

Do any of the members have any experience with any of these kinds of modules?
There are a plethora of different modules available on the likes of AliExpress, from super-cheap to moderately priced, so I am a bit lost as to which one to go for.

If any of the members have played with voice-recognition modules and found good/bad ones, can you please post here with links to any module you might have used.

I really only need ONE voice command, as I just need a simple on/off switch, ideally setup as momentary - if the module recognizes the voice command, it trips the relay for about one second, then releases it and goes back to listening again kind of idea.
Smoke makes things work. When the smoke gets out, it stops!
 
hitsware2

Guru

Joined: 03/08/2019
Location: United States
Posts: 713
Posted: 01:10am 28 Nov 2023
Copy link to clipboard 
Print this post

  Grogster said  
if the module recognizes the voice command, it trips the relay for about one second, then releases it and goes back to listening again kind of idea.

recognize the command word ,
or also the voice of the speaker ?
my site
 
Grogster

Admin Group

Joined: 31/12/2012
Location: New Zealand
Posts: 9306
Posted: 01:28am 28 Nov 2023
Copy link to clipboard 
Print this post

The word would be fine.
I don't really need it to be voice-specific, just need to be able to switch on the command word.

Kinda like the Enterprise computer.
Smoke makes things work. When the smoke gets out, it stops!
 
hitsware2

Guru

Joined: 03/08/2019
Location: United States
Posts: 713
Posted: 03:51am 28 Nov 2023
Copy link to clipboard 
Print this post

You could just whistle  
Should be easy to DIY .....
my site
 
PeteCotton

Guru

Joined: 13/08/2020
Location: Canada
Posts: 368
Posted: 07:04pm 29 Nov 2023
Copy link to clipboard 
Print this post

It might be huge overkill, but I've been dabbling with Alexa, and you can fairly easily hook an ESP-32 up to the Alexa network (and control it by voice). Of course, you still need something to get the audio into Alexa - but that could be any old smart phone you have lying around. The Alexa Dot is currently half price (in Canada) but with a long delivery time after Black Friday.

I know this is a bit of a sledge hammer to crack a nut solution, and people have valid concerns about Alexa listening in, but it's a possible solution.

https://www.youtube.com/watch?v=MsyTQG2ey9s
 
twofingers
Guru

Joined: 02/06/2014
Location: Germany
Posts: 1241
Posted: 11:55pm 09 Feb 2024
Copy link to clipboard 
Print this post

Hi Grogster,

I assume you've already solved the issue with the speech recognition modules. Nevertheless, here's my experience with the Ai-Thinker VC-02 module.

Firstly, I want to clarify: I do NOT recommend the module due to several drawbacks. However, it is inexpensive. It costs just 5 EUR (AE) complete with microphone and speaker. The microphone sensitivity is sufficient, and commands can be easily given and recognized from a distance of 3 meters or more.

The biggest disadvantage lies in the poor and inadequate documentation. It's mostly in Chinese, and unfortunately, many functions are not explained ("automatic study commands"?, I2C ? wake up button ? ...). Another drawback is that compiling the firmware via the web interface typically takes 30 minutes and sometimes unexpectedly fails.




In addition to a poorly or undocumented SDK, there's the option to create firmware for the module using a partially dysfunctional web interface.

Registration is required here:
http://voice.ai-thinker.com/#/





Typically, five GPIO outputs can be linked with a single command (in English or Chinese). A command can control multiple outputs:

GPIO27 = GPIO PIN
GPIOB2 = SCL, Cold light, (GPIO PIN!)
GPIOB3 = SDA, Warm light, (GPIO PIN!)
GPIOB6 = RX1 (GPIO PIN!)
GPIOB7 = TX1 (GPIO PIN!)

Additional pins
GPIO25 = blue light
GPIO26 = "wake up" button

Not accessible from the web interface:
GPIO28 = -- PA chip enable control pin - don't use!
GPIOB0 = - don't use!
GPIOB1 = - don't use!
GPIOB8 = - log ?



It's possible to configure these GPIO pins as high/low or as PWM via the web interface. The commands can be in natural English, as well as the confirmation.
Caution:
Commands must not contain double spaces!

There are two operating modes: Wake state and wake-free operation. To activate the wake state, a keyword must be specified (e.g., "Hello Pudding" or something similar). The wake state lasts for a specific, adjustable time (up to 60 sec) and then switches off. For wake-free operation, up to ten commands can be specified.

The module works well for me now after taking a few days to understand and test everything.
  Grogster said  I really only need ONE voice command, as I just need a simple on/off switch, ideally setup as momentary - if the module recognizes the voice command, it trips the relay for about one second, then releases it and goes back to listening again kind of idea.



I believe it would be suitable for the intended purpose.
The reliability/selectivity is good and I have not encountered any unwanted activations.

-------------------
The module can be programmed directly via the (micro) USB port.

Operation of the burning program (UniOneUpdateTool.exe):

Press the lower button to select the Com port (yellow when the port is online).
Select the file to burn (uni_app_release_update.bin).
Then press the reset button on the vc-02 development board.
The progress of the transfer is displayed in percentages.
-------------------


Here are some additional links:

How to burn the firmware by using the serial port?-VC-02-Kit
https://www.youtube.com/watch?v=yLTr_Xz1-5c

How to train your robot voice module to speak | VC-02 offline voice recognition
https://www.youtube.com/watch?v=rAbNhz6W46g

https://www.youtube.com/watch?v=rAbNhz6W46g&t=45s

https://docs.ai-thinker.com/en/voice_module

https://docs.ai-thinker.com/en/welcome

https://forum.arduino.cc/t/how-to-program-ai-thinker-vc-02/1132521
(only the last post)

------------------

I hope this helps.

Best regards
Michael
causality ≠ correlation ≠ coincidence
 
Grogster

Admin Group

Joined: 31/12/2012
Location: New Zealand
Posts: 9306
Posted: 12:31am 10 Feb 2024
Copy link to clipboard 
Print this post

Hallo.

Thanks for your post.
I think I will check out the module, but yes - I found a solution to this, following the suggestion by hitsware2.
I built a whistle/blow switch.
Adjusted with suitable sensitivity, it can be made to ignore all normal noises like talking and the TV etc, but if you blow/whistle directly at the mic, that crosses the threshold and trips the unit.
So far, that has been working very well, but I will still have a look at this other module you have found.
Smoke makes things work. When the smoke gets out, it stops!
 
Print this page


To reply to this topic, you need to log in.

© JAQ Software 2024