Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

lm386 and speech filter circuit

Status
Not open for further replies.

Mayank Jain

Newbie level 4
Joined
Mar 1, 2014
Messages
6
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
36
I am using lm386 for signal amplification recievied from microphone and then send them to analog pin of arduino..do i need a speech filter circuit also between lm386 and arduino or lm386 would limit the frequency response to 300-3400hz (the speech audio spectrum range)??....
{ basically i am trying to make a audio recorder..the arduino will store the data in a wav file and transfer it to sd-card for playback }..
 

An LM386 is a power amplifier designed to drive a speaker. Its output noise (hiss) is probably too high for it to be a mic preamp. Use a low noise audio opamp instead.

A telephone has a narrow bandwidth because 138 years ago that was the best they could do. It cuts the natural low frequencies of an adult male voice (my voice is not deep but it goes down to 80Hz when I speak) and it cuts all the important consonants sounds of speech (s, f, th etc) that extend to at least 14kHz.
Many words get confused when consonant sounds of speech are missing. "I am failing", "I am sailing" and "I am ailing" sound the same without consonant sounds.
 
if i use lm386 followed by a second order speech filter....will the problem be solved and un-necessary noise will be rejected/reduced

- - - Updated - - -

will this work for reducing hiss???

 

An LM386 is a power amplifier designed to drive a speaker. Its output noise (hiss) is probably too high for it to be a mic preamp. Use a low noise audio opamp instead.
.

Quite so.

Depending on your mic, I would suggest a bandpass filter both to cut unwanted sounds and keep the sampling rate and so resulting wav file reasonable
 

will this work for reducing hiss???

Hiss has the same frequencies as the important consonant sounds in speech. If you filter out the hiss then you are also filtering out the intelligibility of speech.

Your filter cuts radio frequencies higher than the LM386 can amplify. 10pF in series with 10k ohms cuts the radio frequency of 1.6MHz a little.
The LM386 already cuts 20kHz a little and cuts 100kHz to half when its gain is 200 times as shown on a graph on its datasheet.

Why are you using a noisy LM386 power amplifier instead of a low noise opamp?

I worked for years with sound systems for concert halls and live theater. Microphones were used. Speech filters and LM386 power amps were never used.
The preamps were low noise opamps and there was no hiss. Background noises were prevented from happening by closing the doors.
The sound was REAL and ALIVE, not canned and muffled like a telephone.
 

Omit R1 and replace C3 with 3.3nF. Upper frequency limit will be about 3.4kHz. Check for possible unstability.
 



If i use this second order speech filter(pass 300 to 3400 hz) after lm386..will this work??
but the only problem with this op-amp is supply of +9and -9V so i am also searching for the alternative of this opamp if the above solution works
 

OP inputs are mixed-up in the schematic. Active filters can be designed for single supply, but the present circuit isn't suited for it, apart from the said fault.

It would be reasonable to tell your filter specification instead of posting arbitrary dimensioned circuits with unknwon behaviour.

- - - Updated - - -

I see it's roughly a butterworth filter. The circuit can be changed into a single supply filter by using an unity gain sallen-key filter design and replacing R6 by a Vcc/2 voltage divider. Or keep the gain and adjust the voltage divider for second stage output bias of Vcc/2, input bias must be about Vcc/6 in this case. And correct OP input polarity, as said.
 

Omit R1 and replace C3 with 3.3nF. Upper frequency limit will be about 3.4kHz. Check for possible unstability.

I got your part but i thought using simply lm386 omiting the above circuit then using a second order speech filter(that pass 300 to 3400hz)...will this make any difference than your solution
 

any idea what mic you will use? high impedance would work better with the filter you attached


this has some ways to use an op-amp with single-ended power supply
**broken link removed**
 

i thought using simply lm386 omiting the above circuit then using a second order speech filter(that pass 300 to 3400hz)...will this make any difference than your solution
An LM386 cannot be used as the Sallen-Key filter you showed that uses an opamp.

I simulated your circuit using ordinary opamps. Its peak gain is 4.0dB. It cuts deep adult male voices (-3dB at 295Hz) and cuts all the consonant sounds of speech (-3db at 3200Hz) something like a telephone.

You need a mic preamp circuit that amplifies the mic 10mV level up to about 640mV then the preamp feeds the filter.
 

Attachments

  • Sallen-Key voice bandpass.png
    Sallen-Key voice bandpass.png
    30.3 KB · Views: 184
Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top