This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Sir,
I am a member of TI INDIA ADC Contest.I would like to know, how to interface apr9600 voice recorder and playback kit with MSP 430. I would like to know is there any requirement of coding and test vectors.And how to get started with this interfacing.
Regards
A.Nikilesh Reddy
Sir, I didn't get ur exact sense.Apr9600 it does everything, in what sense do u mean it sir,means whether itself interface with msp430 or information is converted into digital mode by itself.Can i use apr 33a3 as voice recorder and playback kit.
nikilesh reddy said:Sir, I didn't get ur exact sense.Apr9600 it does everything, in what sense do u mean
I mean that to operate, apr9600 does not need any microcontroller. That's why I asked you - what will be msp430 function? Please describe your project in more details, because "interface apr9600 voice recorder and playback kit with MSP 430" is not that much.
Sir,
These are my project details, we are creating VOICE CONTROLLED CALIPERS.In our project we will be first recording the voice with apr9600 and passing it through the MSP430G2553IN20 and the signal will be processed in the MSP430G2553IN20 and we will be matching the recorded voice with the test vectors. If it is matched with the Test vector then it is passed through the encoder and encoded and with the help of transmitter it is transmitted through the receiver section and it is decoded and again processed through MSP430G2553IN20 and it is sent to the L293D driving circuit and the instrument is driven. These are the details of our project.So,kindly help us in the interfacing part of apr9600 and MSP430.Please suggest me some test vectors required for the process.
Regards,
A.Nikilesh Reddy
nikilesh reddy said:In our project we will be first recording the voice with apr9600 and passing it through the MSP430G2553IN20 and the signal will be processed in the MSP430G2553IN20 and we will be matching the recorded voice with the test vectors.
Well... thing is that apr9600 is just voice recorder which is aimed to record voice and play back later. To do voice recognition you don't need to **record** voice in tape or voice recorder, you need that your microcontroller is able to record voice and do voice recognition alone. Actually I highly doubt that msp430 is good for your application because voice recognition is CPU-intensive task. Msp430 is not that powerful uC. Also as you use L293D it seems that you don't need low power features of msp430 either. Try to consider Tiva C microcontroller or maybe even some DSP from TI.
Ilmars said:I highly doubt that msp430 is good for your application because voice recognition is CPU-intensive task. Msp430 is not that powerful uC
I would agree with that.
nikilesh reddy: What research/study have you done into the techniques of voice recognition to determine which one(s) may be suitable to your application?
Once you've identified suitable technique (or techniques), then you can estimate its/their computational requirements - and pick a processor appropriate to that...
I would like to know whether i can use C 2000 microcontroller can be used for voice ***recognition*** and for the processing of the recognised voice.Please clear my doubt.Or can i use any other controller.
Why do you ask C2000 question in msp430 forum? Hint:
http://e2e.ti.com/support/microcontrollers/c2000/default.aspx
Andy Neil said:What research/study have you done into the techniques of voice recognition to determine which one(s) may be suitable to your application?
Once you've identified suitable technique (or techniques), then you can estimate its/their computational requirements - and pick a processor appropriate to that...
I once had a voice recognition and speech processor system for my C64. In 64k ram (including program space and phonetic speech sample storage), only 1MHz CPU speed, and a 4-bit flash ADC, it was capable of recognizing 32 trained words in realtime. It was fun playing poker with the C64 by just saying 'draw'. 'fold' and so on, and also getting a spoken answer.
So it shouldn't be a problem with the MSP too. ADC10 or ADC12, even in 8 bit mode, are more than enough. And the CPU power should be plenty. If training data is stored in flash, then I don't see a problem why not using an MSP like the 5437A or so.
Jens-Michael Gross said:So it shouldn't be a problem with the MSP
Sure you can do seemingly impossible tasks with msp430. Question is - what kind of performance you will get in result. Your example was not voice recognition but rather sound pattern matching. Yes, it is possible to do something like that on msp430 but then please don't call it voice recognition.
Agreed.Ilmars said:not voice recognition but rather sound pattern matching.
Well, even leading systems like Dragon Natural Speaking require quite some training. So where do draw the border? It just seems to be a question of the database size and how fast you can search it.Ilmars said:please don't call it voice recognition.
And from semantics, 'voice recognition' means detecting that someone says something (but not what he says) :)
Hi nikilesh reddy,
You don't need to go for APR9600, if you plan to use MSP430FG4618 kit,
It has internally option to store voice.
-karthik
**Attention** This is a public forum