Voice-controlled soundboard

Hi everyone!

I’m new here and looking to build a soundboard that we can control by voice.

A voice-controlled soundboard is an application or device that allows users to trigger and play various sounds or audio clips using voice commands. It combines speech recognition capabilities with a collection of audio files to create an interactive and hands-free experience for playing sound effects, music snippets, or pre-recorded sounds.

I was attempting to create a demo version using Porcupine until I realized that I can create only 3 models within a month. This is a huge blocker since I needed the soundboard to detect 9 distinct words and play a sound clip paired with a given word. I’m curious if mycroft-precise works as an alternative in this case.

Thanks for any suggestions!
Michi

see Introducing precise-lite-trainer

2 Likes

I am interested in your experience.
In my experience, first I ask chatGPT.
Just I used ChatGPT so I knew these things.
Mycroft Precise is a wake word listener and speech command recognition engine that can be used as an alternative to Porcupine for your voice-controlled soundboard project. While Porcupine has limitations on the number of custom wake words that can be created within a month, Mycroft Precise offers more flexibility in terms of training and recognizing multiple wake words or phrases.
This means you can potentially train it to detect and recognize the 9 distinct words you mentioned and associate each word with a specific sound clip for your soundboard.


So I think you have to collect sample dataset and identity start and end points of each word in the audio samples.
And also it has to label the training data.
Next, you should train Mycrosoft Precise model with labeled training data and integrated soundboard.
I hope this answers your question.
If you have any more problem, feel free to ask!

1 Like