Voice recognizer

[paparts]

Have you seen the hollow man? Have you noticed how they go up and down the elevator? I want to design how their voice are recognized. Can you give me ideas on how to start? tnx

 

[Papabravo]

I would start with a frequncy analysis of a voice sample. It could be an analog analysis with an array of narrow bandpass filters which measure the sounde energy over a range of frequencies. It could be a digital analysis using a Fast Fourier Transform or related algorithm to digitally extract the sound energy as a function of frequency. Assuming you could account for amplitude variations you could use a goodness of fit criteria to match a sample to stored transforms looking for a match.

BTW I have no idea weather either approach would be successful, it's not my field.

 

[dknguyen]

I think you would use an ADC to digitize the sounds coming in after passing it through some filters. THen you would use some Fourier transform (you would probably need a DSP rather than an MCU), and then maybe normalize the fourier spectra and run it against a library, making sure to take into account possible fuzziness and variation in readings?

I guess you would probably also average it over time in order to ignore sounds that are constantly there, but that's later on.

 

[Electronics4you]

Just use the VR Stamp from Sensory

**broken link removed**
**broken link removed**

 

[Oznog]

There are 3 terms, which often get mixed up in common use:
1. Voice Detection (VOX)- just tell if someone is speaking, used to key a microphone
2. Speech Recognition (Word Recognition)- recognize the words in speech
3. Speaker Recognition- recognize the identity of who is speaking among multiple possible options

 

[Sceadwian]

For the cost that would be required to reliably re-create something like that you could probably higher someone at minimum wage to listen to mass microphone input and push the buttons for them..

 

[Papabravo]

Elbonians would be more than happy to quote your outsourcing requirements. Grin