Need advice for a handmade Wi-Fi speaker!

Hi all,

I am a new member here and new to microcontroller too. After trying to simulate some simple applications with PIC, I get excited with MCU and want to do a real project. OK, this is a school project and part of it is to design a real-time Wi-Fi speaker. After looking around, I intend to use the following main components a Microphone + Router + a Wi-Fi module (the Wi-Em module of DigiConnect) + PIC16F886 + VS1011.

My plan is that the Microphone will receive and convert analog voice signal to digital signal; and then a software will be used to code the digitilized signal in MP3 format. The Router will play as a base station and transmit the coded information to the Wi-Fi module which then transfer data to PIC for processing. I am stuck at this point as I don't know how the PIC will do with the coded data. Can it just pump out anything it get from the Wi-Fi module to the chip VS1011 to produce sound or do I need to used an extra memory device to store the data before pumping to VS1011 for processing?

Do you guys have any idea for the feasibility of my project?

(I am not sure about my plan as this is my first time to do an electronic project)

Thank in advance!

The theory is OK but do some research on data rates before going any further. In particular look at the bit rates and check the wireless link can handle them and the processor can unload the data fast enough. A good place to start would be to take the size of a typical MP3 file and divide it by it's length in seconds to discover how many bit/s are needed to carry it. Then add on an additional margin for the transfer protocol and possibly error correction. You will probably find the PIC is unable to offload the data fast enough to work in real-time. I have not done any real calculations, the result depends on many factors, in particular the MP3 sampling rate you use, but hopefully I have given you some pointers to research topics.

Brian.

Thank you for your reply Brian.

As your suggestions, I did some calculations and refered to datasheets for checking. I find that a MP3 file that is 32 kbps or 48 kbps can be handled by the copomnents involved as they can support up to 112 kpbs (but 57.6 kpbs is safer). The sampling rate that fast is good enough for the project.

The thing that I worry the most is the error correction scheme, is it easy to happen in this kind of project?

Ngoc Anh.

Errors will almost certainly occur, either because of momentary loss of signal or interference from other devices on or close to the same frequency.
You can deal with them in three ways:
1. Ignore them, but risk crackles and pops from the loudspeakers when they occur.
2. Add some correction bits (like FEC) to the data to allow the receiver to correct small errors, this increases the data rate somewhat which in itself might be counter productive.
3. Use a "resend bad data" protocol to request for data with errors to be resent. This requires a bi-directional wireless link and additional buffering at the receiver to maintain audio while a resend is in progress.

If you are confident the link will be clean and uninterrupted, you can probably use the first strategy. If it doesn't perform you can always fall back on the other methods instead.

Brian.

i think do no encode it in mp3.bcoz,mp3 coding & decoding takes much time so not realtime...
use analog modulation.......i think it's the best way......u can use 433 MHz RF Receiver and Transmiter
link: 433 MHz RF Receiver and Transmiter - Rs. 250.0 : Probots - India's Electronics and Robotics Superstore