Still voice recognition ....

Go To Last Post
16 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Did anybody test this system ? It is mentionned in relation with Arduino ... And does anybody know where one could find more informations about it ? The target of the link at the page's bottom is ... in Chinese ! Thanks for all answers ...

This topic has a solution.

Last Edited: Mon. Aug 30, 2021 - 04:34 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

alberich wrote:
And does anybody know where one could find more informations about it

How many google hits do you get when you google the system?

As with most things Arduino, there will be lots of hits, and will take some time to sort through all the noise.

Good luck with your project.

Jim

 

 

Keys to wealth:

Invest for cash flow, not capital gains!

Wealth is attracted, not chased! 

Income is proportional to how many you serve!

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Finally I tested almost ALL available systems :

 

- Elechouse, both SD and SI, but they don't have an adjustable level of sensitivity (too many problems with ambient noise) and work rather bad with "other" microphones.

- Fortebit/Sensory : serial communication with uP very difficult, supplementary delays must be included in many places of prog ; and level of sensitivity very finicky to adjust, between too many rejects and too many false recognitions.

- Mikroe SpeakUp : impossible to get this module recognized with Win7 x64 for vocabulary training (I saw on www that I am not alone with this issue) ; and no training "in situ" available ...

 

Thanks for "any other" suggestions ! sad

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

alberich wrote:
Thanks for "any other" suggestions !
trying to do recognition in the micro itself is so "last millennium". Maybe look at the way it's now done in 2021...

 

https://cloudcompiled.com/blog/transcription-api-comparison/

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

But I am looking for "offline" solutions, for several reasons ...

Last Edited: Sat. Aug 28, 2021 - 10:20 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

Then don't expect to achieve Alexa/Google/Siri levels of operation. The reason all those systems send the sound sample off to some insanely powerful server farm to do the actual recognition is because that is what quality recognition needs. Otherwise all your Amazon Echos and so on would be doing it locally on their (already powerful) micro rather requiring a live server connection.

 

(but it is true that our house comes to a bit of a standstill if all the echo's lose Internet connection!)

 

Because distant recognition is the usual method used these days there's not so much development of local solutions so you may be hard pushed to find anything that offers anything close to what is now considered acceptable recognition rates. People are too spoiled by their Siri/Google phones/watches/etc. 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

clawson wrote:
trying to do recognition in the micro itself is so "last millennium".
Yet still alive; IIRC, mid-90s PDA (200MHz[+/-] ARM) could run Sphinx.

CMU Sphinx Downloads – CMUSphinx Open Source Speech Recognition

...

  • Pocketsphinx — recognizer library written in C.

...

edit :

Teensy 4.0 and Voice Recognition (PJRC)

 

"Dare to be naïve." - Buckminster Fuller

Last Edited: Sat. Aug 28, 2021 - 01:56 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

... and I hoped to find a ready-to-use module !

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

A thought occurred to me. Have you considered using a Raspberry Pi as a "module". There's bound to be open solutions for voice recognition in Linux and some of the latest Pi boards have huge amounts of CPU power like quad core 1.4GHz which should be enough for some pretty impressive performance. 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks "clawson", but I am afraid that my project becomes a "gaseous plant" !

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

in #1, alberich wrote:
Thanks for all answers ...

but in #5, alberich wrote:
But I am looking for "offline" solutions, for several reasons ...

and in #10, alberich wrote:
but I am afraid that my project becomes a "gaseous plant" 

 

So you were actually looking for something rather more specific than "all" answers ... ?

 

frown

Top Tips:

  1. How to properly post source code - see: https://www.avrfreaks.net/comment... - also how to properly include images/pictures
  2. "Garbage" characters on a serial terminal are (almost?) invariably due to wrong baud rate - see: https://learn.sparkfun.com/tutorials/serial-communication
  3. Wrong baud rate is usually due to not running at the speed you thought; check by blinking a LED to see if you get the speed you expected
  4. Difference between a crystal, and a crystal oscillatorhttps://www.avrfreaks.net/comment...
  5. When your question is resolved, mark the solution: https://www.avrfreaks.net/comment...
  6. Beginner's "Getting Started" tips: https://www.avrfreaks.net/comment...
This reply has been marked as the solution. 
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

My ultimate hope was a LD3320 based module (on AliExpress) but the only example of software that somebody reported to me in another thread :

1) Was written in C, which I don't use ... (I program in GCBasic).

2) And perhaps I would have been able to understand the code, but the comments were written ... IN CHINESE ! sad (I speak French, German and English, but NO Chinese ! )

Maybe I will test again the SimpleVR ... Perhaps that, with limited vocabularies (using a tree structure) and a cardioid mike ...

Last Edited: Mon. Aug 30, 2021 - 04:35 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

alberich wrote:
the comments were written ... IN CHINESE ! sad (I speak French, German and English, but NO Chinese ! )
Google is pretty good at translating.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Have you considered using a Raspberry Pi 

  my project becomes a "gaseous plant" !

I have no idea what that means...

 

The project linked to HERE, on the Ada Fruit site, uses an RPi, Voice2JSON, and PocketSphinx to do stand-alone voice recognition.

 

(I program in GCBasic)

Although I definitely feel that I am in the minority, Basic is my preferred language, also!

 

JC

 

 

Last Edited: Sun. Sep 12, 2021 - 05:25 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

DocJC wrote:
I have no idea what that means...
Tail wagging the dog perhaps?

 

You have a beast of an Rpi just acting as a "module" to a simple 8 bit micro or something.

 

Having said that decent recognition (to match the quality people expect these days from Siri, Google-Home, Alexa) is going to take buckets of MIPs so there's bound to be some powerful silicon involved whether it be CPU in RPi or DSP or whatever.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Have you guys read about this kind of voice recognition?  I think you can read 10 stories per month at the NYTimes for free without subscribing so most of you will be able to open it.  This one is worth reading.

 

Opinion | Hear That? It’s Your Voice Being Taken for Profit. - The New York Times (nytimes.com)

 

Alexa?  NO THANKS!