Voice record/playback design help

[nexus_2006]

Hi all, first post.  I need help designing a voice storage system.  I want to store about 5 short messages (less than a minute in most cases), and be able to play them back.  I am trying to find a decent IC that can do this, but am not having any luck finding something readily available.  The goal is to be able to listen to a voice-band radio and store the 5 most recent for playback.  I'm thinking the "click" of the microphone keying up would make a fine "start and stop recording" trigger.  It would also need to be able to throw the oldest message out to free up space for newer transmissions.  This is likely least complicated with a dedicated microcontroller that can be programmed by PC or something, rather than discreet components.  Attached is the concept block diagram.  Any help at all would be appreciated.

View attachment 41382

 

[KMoffett1]

Check these segment addressable recorder/playback chips:
http://www.nuvoton.com/hq/enu/ProductAndSales/ProductLines/ConsumerElectronicsIC/ISDVoiceIC/ISDChipCorder/Pages/default.aspx

Ken

 

[pdp]

In my opinion, it could make some sense if only the goal was to build a special featured device. Othewise it would just have to compete with great amount of those inexpensive factory-made voice recording gadgets that are easily available now.

 

[kxenos]

My suggestion is: Use an ARM uC to capture, encode and store the signal into an external eeprom. Use an open source speech codec like SPEEX which is perfect for encoding speech with excellent quality into bitrates of 2kbps, 4kbps etc. With 8kbps you will have perfect results. With 8kbps audio bitrate, if you want to store 5min of audio you need 5min * 60sec = 300sec. 300sec * 8kbit/sec = 2400kbit or 300KB of data. Take a look at ST micro cortex M3 which is STM32*. There is an application note that does exactly that. It has the audio circuit around the uC for line input and line output. OLIMEX has a dev. kit that has implemented exactly this circuit. So all you have to do is load the code from ST and off you go. You don't even have to do a prototype for proof of concept.
App. note.: http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/APPLICATION_NOTE/CD00204907.pdf
dev. board: STM32-103STK link: http://www.olimex.com/dev/pdf/ARM/ST/STM32-103STK.pdf
You will also have to use an external memory for the messages. The dev board has an SD/MMC connector so you might as well use this resource. Extra feature: Voice messages could be transferred into laptop or other device via sd card if you implement a known file system like FAT16.
Hope I helped. Just registered today, so I'm excused for the delay, right?

 

[ARUN CHOWDARY]

it is a very easy design use a high sensitive microphone and amplify the signal by using a power amplifier store it in a memory device use computer for it or any interface recorder

 

[Herman the German]

Hi nexus_2006,

you might want to try this circuit.

It is built around an ISD2532 (32seconds at 8KHz sampling frequency). If that one won't suffice you might select an ISD2564 (64seconds).

The device can be controlled locally and via a microcontroller. All necessary connections are available at SV2.

Remark: The mic connects via two pads. However if a mic with built in Fet-transistor is used polarity may be changed turning the mic 180degrees.

If you are interested in the Eagle files leave your full name and email address in a PM.

Regards

HtG

View attachment 41499

View attachment 41500