Searching \ for 'PIC doing Sound Recognition?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: piclist.com/techref/microchip/ios.htm?key=sound
Search entire site for: 'PIC doing Sound Recognition?'.

Truncated match.
PICList Thread
'PIC doing Sound Recognition?'
1999\09\28@142828 by Wagner Lipnharski

picon face
Is this a nice thread to talk how to develop a SRM using a fast PIC and
an ADC?

My suggestion is using four digital (software) filters, 200Hz, 450Hz,
1.1kHz and 2.5kHz, or something similar (not harmonics relative):

a) Scan the incomming sound signal in modules of 100ms.

b) Account how long / amplitude each filter identified that frequency
band during each module in an accumulative 16 bits register. As example,
a pure 200Hz tone of zero dB during 100ms would generate a 16 bits value
of 0FFFFh at its correspondent filter register, while a complete lack of
some frequency band will generate value of 0000h at that particular
frequency filter register.

c) Store those 4 x 2 bytes x 10 modules (80 bytes) per second.

You will have *some* representation of the sound. Any small changes in
speed, frequency or intensity would change small numbers at the 16 bits
filter registers.  An aproximation comparative software can locate
similar pre-scaned sounds within a pre-determined percentage of error in
value and in time.  It will not discriminate between "Six", "Sex",
"Sax", "Sucks" or "Socks" (well, who does?), but for "Stop", "Go",
"Pause", "Left", "Right", "Up" or "Down" it should do it.

Increasing the quantity of filters or reducing the time per module
should increase resolution. Probably eliminating the amplitude sampling
and accounting just frequency could turn the device recognition less
critical.

The device could recognize some "clank-dee-clank" machines sounds too,
or even an "Ouchh" meaning "STOP".

What kind of processor and speed, a device like that would requires?
Probably a PIC per digital filter, parallel processing?

Do you guys think it is a valid thread to keep on going?

Wagner

1999\09\28@211045 by Russell McMahon

picon face
I have (somewhere) a very old article using a Motorola 6800 which did
similar to this (a few more filters than this I think) with creditable
results. Certainly a possibility.

RM

_____________________________
What can one man do?
Help the hungry at no cost to yourself!
at  http://www.thehungersite.com/

From: Wagner Lipnharski <spam_OUTwagnerlTakeThisOuTspamEARTHLINK.NET>

>Is this a nice thread to talk how to develop a SRM using a fast PIC and
>an ADC?
>
>My suggestion is using four digital (software) filters, 200Hz, 450Hz,
>1.1kHz and 2.5kHz, or something similar (not harmonics relative):
>
>Do you guys think it is a valid thread to keep on going?
>

1999\09\28@223527 by Richard Prosser

flavicon
face
I can feel the possibility of a voice operated TV remote control arising -
was it that creditable?

Richard P

> {Original Message removed}

1999\09\28@230936 by Wagner Lipnharski

picon face
Richard Prosser wrote:
>
> I can feel the possibility of a voice operated TV remote control arising -
> was it that creditable?

The interesting point here is that you could control several things at
your home via a simple phone call, whit your answering machine loud
speaking your voice commands.

Of course your TV voice remote could not be programmed with trivial
words that could come from the TV speakers, as OFF, ON, but instead you
could use RICKYON, RICKYOFF... :) isn't nice?

I have a Magnavox SmartTalk voice remote control, that can recognize all
the TV and VCR commands for 4 different persons, works somehow nice, but
it drains 4 AA cells every two weeks... not nice at all, and you need to
press a button to activate the voice recognition system.

More... (looser matching)
- Last day of these posts
- In 1999 , 2000 only
- Today
- New search...