about MSP430FG4618 project

[abdulmeeravali]

sir, i want to develop the automatic speech recognition using MSP430FG4618 so pls tell me the pre-requisite things need to know...
and help me in giving some tips for programming..............

 

[pashok84]

you which development board, build on your own or brought from TI or softbaugh or from someother company?

simple and general list
you need a development environment IAR or MSPGCC
programmer

 

[poorchava]

I think this MCU is WAY to slow to do speech recognition at reasonable rate. I've seen it done on dsPIC30f5013 but it operated at around 30 MIPS with DSP core involved.

In most cases speech recognition is realized with Hidden Markov Models (HMM). In the past main method was Dynamic Time Warping (DTW).

Generally speech recognition involves analog signal frontend conditioning signal from either microphone or line input (amplification, shifting, filtering). Next that signal is sampled by the ADC to get digitalized version of sound waveform. And then the magic happens. Software inside microcontroller has to recognize speech from general noise, then separate particular words or woves from speech. Then you need an algorithm oto classify the words and act accordingly (e.g choose some option in a menu).

This task is very hard to accomplish, because it onvolves pattern classification which is always hard for computers.