Buriedcode
Full Member level 6
Hi,
I posted several questions in this forum about a project of mine that I've been working on, on and off for about 2 years. Its a wireless digital audio system for musical instruments. Heres the rough specs..
resolution: 16 bit. CS5330A ADC, stereo, only using left channel.
Sampling rate: 32kHz
Radio band: 2.4Ghz, simplex.
FEC: undecided, but I really want error correction.
Other coding: Cannot use compression, or source coding, due to latency and the signal being sent.
Objective: To get wireless, high quality audio
I've posted in several different areas of eda, CPLD, DSP, etc.. and I've had some responses, just thought the people in here could give some advice.
Right, since this is a 'real-time' transmission, the latency between samping at the ADC in the Tx, and writing to the DAC must be below 2ms. I've researched so many different areas of channel coding, both block and convolutional (golay,hamming - Viterbi etc.). It would take hours for me to explain every idea I've had, and the experiments I've done.
In the latest version:the 16-bit word is split into 12 and 4. Thefirst is encoded using the golay algorithm, the last 4 bits (LSB) with a 8,4 hamming. Since the most significant bits are a lot more 'significant' in audio.
First version was this: ADC -> Hamming (13,8 on high nibble and low nibble) -> Packeting with a 8 bit preamble -> Manchester encoding.
All using the MACH4-64/32 CPLD.
Its simple but it does the job, although I haven't tried it with a radio yet, which brings me to the point of this post.
Considering I'm looking at a data rate of 48*32000 = 1.536MB/s, with manchester encoding thats, 3.072MB/s, is there a 2.4Ghz solution that can cope with this? I know all about FHSS and DSSS, but they either require a rigid packet structure, as in FHSS, or are too specifically designed for Wireless networks (802.11a/b/g). Maybe those 2.4Ghz AV modules would do, seeing as the bandwidth of a video signal is about 5Mhz. The project can be simple, but its easy to start 'over-complicating' it with ultra modern comms technology, like QPSK, DSSS, scrambling, turbocodes. Any input would be helpful, I'm completely stuck.
Also, one more thing
I know this isn't aspecific question, and its vague, but I really need some guidance. I don't want to end up woth something thats unreliable, or too expensive. So maybe I should add/remove things from the design.
I posted several questions in this forum about a project of mine that I've been working on, on and off for about 2 years. Its a wireless digital audio system for musical instruments. Heres the rough specs..
resolution: 16 bit. CS5330A ADC, stereo, only using left channel.
Sampling rate: 32kHz
Radio band: 2.4Ghz, simplex.
FEC: undecided, but I really want error correction.
Other coding: Cannot use compression, or source coding, due to latency and the signal being sent.
Objective: To get wireless, high quality audio
I've posted in several different areas of eda, CPLD, DSP, etc.. and I've had some responses, just thought the people in here could give some advice.
Right, since this is a 'real-time' transmission, the latency between samping at the ADC in the Tx, and writing to the DAC must be below 2ms. I've researched so many different areas of channel coding, both block and convolutional (golay,hamming - Viterbi etc.). It would take hours for me to explain every idea I've had, and the experiments I've done.
In the latest version:the 16-bit word is split into 12 and 4. Thefirst is encoded using the golay algorithm, the last 4 bits (LSB) with a 8,4 hamming. Since the most significant bits are a lot more 'significant' in audio.
First version was this: ADC -> Hamming (13,8 on high nibble and low nibble) -> Packeting with a 8 bit preamble -> Manchester encoding.
All using the MACH4-64/32 CPLD.
Its simple but it does the job, although I haven't tried it with a radio yet, which brings me to the point of this post.
Considering I'm looking at a data rate of 48*32000 = 1.536MB/s, with manchester encoding thats, 3.072MB/s, is there a 2.4Ghz solution that can cope with this? I know all about FHSS and DSSS, but they either require a rigid packet structure, as in FHSS, or are too specifically designed for Wireless networks (802.11a/b/g). Maybe those 2.4Ghz AV modules would do, seeing as the bandwidth of a video signal is about 5Mhz. The project can be simple, but its easy to start 'over-complicating' it with ultra modern comms technology, like QPSK, DSSS, scrambling, turbocodes. Any input would be helpful, I'm completely stuck.
Also, one more thing
I know this isn't aspecific question, and its vague, but I really need some guidance. I don't want to end up woth something thats unreliable, or too expensive. So maybe I should add/remove things from the design.