Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Help! Don't know how to build a speech-to-text system

Status
Not open for further replies.

urangeloflove

New Member
I have just been given a college project to work on and it involves building a basic speech-to-text system. All I have been able to find out is that I will need to design the mic pick-up part and the Analog-to-Digital Converter part. But I don't know what else to do? Please guys I need your help. Kindly guide me or give me a web link to get a practical idea of what to do. Thanks guys!!!
 
So what do you have to do? Have you just been assigned the mic amplifier and ADC selection?
 
The speech to text part is very complicated (Hidden Markov model type of algorithms). Win XP onwards supports this so I assume it's an academic exercise.
You need an electret mike, an amp with at least 50dB of voltage gain (preferably with soft limiting/compressing), anti-aliasing filter and an ADC to sample at least 11025 Hz at probably 10+bits for speech. Easily done with modern codecs, but I suspect you need a broken down system with independant copmponents. You can combine the amplifier/limiter and anti-aliasing filter by using a low grade audio op-amp with a bit of high frequency roll off in the feedback circuit .
 
The speech to text part is very complicated (Hidden Markov model type of algorithms). Win XP onwards supports this so I assume it's an academic exercise.
You need an electret mike, an amp with at least 50dB of voltage gain (preferably with soft limiting/compressing), anti-aliasing filter and an ADC to sample at least 11025 Hz at probably 10+bits for speech. Easily done with modern codecs, but I suspect you need a broken down system with independant copmponents. You can combine the amplifier/limiter and anti-aliasing filter by using a low grade audio op-amp with a bit of high frequency roll off in the feedback circuit .


Thanks I kind of understand what you advised but can I be guided more practically? Thanks again...
 
I meant are you only responsible for finding a solution for the mic amp and ADC?

ie is this a group project that has had the responsibilities apportioned to each member?
 
Right. It's a pretty big project for you then.

As you've mentioned, you'll need an amp & ADC; perhaps the SSM2166 (https://www.electro-tech-online.com/custompdfs/2012/02/SSM2166_a.pdf) would be good as the amp. You can use the ADC internal to the processor or an external one and interface via I2S; perhaps WM8783 (https://www.electro-tech-online.com/custompdfs/2012/02/WM8783_1.pdf) would be suitable - it has the anti-aliasing filter as well as the ADCs.

If you don't have to roll your own speech recognition engine, then perhaps have a look at the microchip library (**broken link removed**) which boasts a 100 word vocab.
 
I dont want to open a new thread, Another Similar Thread: https://www.electro-tech-online.com/threads/converting-speech-to-text.41354/

Speech Recognition seemed to be a very difficult job even to Microsoft office (speech feature), i found that there was lots of mistakes that used to occur using it. But, very recently i saw a student project wherein a persons voice message is converted to a text message, and that text is displayed using a LED Matrix. It was not possible for me to get into detail of the project. But AFAIK, the project used MATLAB code, which enabled to convert the voice to text in the computer and displayed that through a LED/LCD Display. The MATLAB Code will be similar to something here: https://www.mathworks.com/matlabcentral/fileexchange/11297-text2speech-tts .

And, it works well and good. I wonder why ms office had the problem a problem with this.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top