need help about realtime digital audio microcontroller

Thread Starter

angelita

Joined May 29, 2009
25
hi i had a project to build an automatic audio systems for some place

and i need to use realtime microcontroller and i dont know how to search

please i need information about the best
digital audio processing microcontroller

i need a microcontroler to get rid from noise


is there any way to get rid from echo using microcontroller
books ..sites ..suggestion data sheet plz help
 

Papabravo

Joined Feb 24, 2006
21,225
I believe the 8-bit AVR family is mostly unsuitable for a digital audio application. The AVR32 has a subfamily called the UC3 which has a greater chance of being suitable for such applications. Neither one of those is a true DSP.

If I were you I would examine offerings from Freescale (563xx) as they have specific peripheral devices that make processing digital audio streams much easier. Next would be Analog Devices with the SHARC and Blackfin series of processors. Then TI which has an assortment of fixed and floating point parts. A dark horse would be XMOS that has a scaleable multi-threaded processor that runs faster than a scalded duck.

Our legacy products use the 56004 from Freescale which is now well past EOL. We're deciding which way to go next.

I completely agree with Bill. This is not a project for beginners or even most graduate engineers. Believe me when I say that it is really tough to do it right.
 

Papabravo

Joined Feb 24, 2006
21,225
The 8051 variants are even more limited than the AVR

What can you do with a DSP?

  1. Sample rate conversion
  2. MP3, WMA, and FLAC decoding
  3. FIR and IIR filters
  4. Format conversion
  5. Apodizing filters
  6. Pre-emphasis and de-emphasis
Just for starters
 

retched

Joined Dec 5, 2009
5,207
You DONT want to do Digital Signal Processing of any kind of complex signal (Over 4 bit) without a DSP uC.

They are designed to use the power that is fed to them as efficently as possible for DSP tasks. A standart PIC or AVR is not.

They are more 'generic' so-to-speak. They pass around the power pretty equally. You dont want that. You want to process signals...digitally ;)

You CAN use 2 uCs if you are so inclined. You can use a DSP for the signal processing and a run-of-the-mill 8bit PIC for other functions, like user interfacing.

If you want to play music in the background while the user searched through menus and the such, this will work. The 2 uCs can communicate to each other cia SPI or whatever and you can allow one to INTERRUPT the other when data needs to be passed.
 

Thread Starter

angelita

Joined May 29, 2009
25
do i still need to use acoustic echo cancelation even the microphone too far from speakers

and

after the sound processing can i distribute the sound equally over speakers using dsp microcontroller ..
 
Here's a bit of advice.

Do your project on a PC before you attempt it on an embedded platform. If you can't get your algorithms working on a PC with a soundcard / microphone you have zero chance of getting it working on an embedded design.
 

Audioguru

Joined Dec 20, 2007
11,248
I worked with very high quality boardroom tele-conferencing systems. They used digital echo cancellers so that they could transmit and receive at the same time (full duplex).

At the beginning of each tele-conference the unit was "trained" with pink noise from the speakers so it knew the room echoes and phases which are different with the number of people present and if a door was open or closed. If something changed the echoes during a teleconference like a door opening or closing then the digital echo canceller would get confused and cause acoustical feedback howling for a few seconds.
 
Top