Go Down

Topic: How to store a voice signal in Arduino Mega (Read 117 times) previous topic - next topic

B-Vikas-Nanthan

Aug 07, 2015, 07:09 pm Last Edit: Aug 09, 2015, 07:50 am by B-Vikas-Nanthan
I'm making a robocar using voice control,but don't know how to store a voice signal in arduino and compare the signal with the input voice signal.Does it need any extra peripherals. Kindly reply.

DVDdoug

#1
Aug 08, 2015, 11:57 pm Last Edit: Aug 08, 2015, 11:59 pm by DVDdoug
I don't know how voice recognition/speech recognition works, but I know it's not  as simple as comparing audio "data".*

You'll need some external memory if you want to store sound.  Maybe, an audio shield.   i.e.  A "CD quality" uncompressed recording has a sample rate of 44.1kHz, which is 16-bits per sample and requires 88,200 bytes for one second of mono sound.

I suggest you search for "Arduino speech recognition".   I've seen something for doing that, but I don't have a link and I don't know it's limitations.




* For example - You can record yourself saying "hello", and if you subtract a copy of the file from itself (which you can do with an audio editor, like Audacity) you'll get dead silence showing that the two files match and there's no difference.

However, if you record yourself saying "hello" twice, you'll find the difference-file sounds exactly like addition (audio mixing)...  It sounds like you and your twin are both saying "hello" together.   This is because at the sample-level the two files are completely different and uncorrelated. 

Grumpy_Mike

#2
Aug 09, 2015, 06:51 pm Last Edit: Aug 09, 2015, 06:52 pm by Grumpy_Mike
To use voice recognition you need another processor not an Arduino. There are shields with these processors on that have varying degrees of success, from poor to crap.

Go Up