Audio CODEC
I searched OBEX but not found a CODEC for telephone communication, something like speex...
I will try to interconnect the propeller to this board :
http://www.semiconductorstore.com/pdf/newsite/silvertelecom/Ag1171v1-3.pdf
Wich is a cheap and simple subscriber line interface.
While at it i looked at the reference circuit at propeller manual and found audio in and out.
The pins connected to audio in/out are analog ones ? Or some kind of PWM/Comparator is being used ?
I will try to interconnect the propeller to this board :
http://www.semiconductorstore.com/pdf/newsite/silvertelecom/Ag1171v1-3.pdf
Wich is a cheap and simple subscriber line interface.
While at it i looked at the reference circuit at propeller manual and found audio in and out.
The pins connected to audio in/out are analog ones ? Or some kind of PWM/Comparator is being used ?
Comments
Look into this for examples: http://www.parallax.com/Portals/0/Downloads/appnt/prop/AN001-PropellerCountersv1.2.zip
Similarly, an analog audio output can be built with a resistor ladder DAC or even an RC circuit, but other specialized devices will provide better quality.
I ended up creating this schematic that i want to share :
analog inout.pdf
Is this ok ?
Im still afraid because prolly the line will output 0 to 1v peak audio, and this schematic will output 0 to 3.3v... (i need to read more about the AD7303 to see if the internal VRef can be programmed to 1V).
Well, generally, is this setup ok ?
Edit1: Found the first error, ADC0832 is a 5V only part...
You could get started with "microphone to headphone" object.
I thought there was also DTMF or touch tone object somewhere.
Perry
Looking at the datasheet, the "codec" they are indicating is simply the audio input & output. So this could be a normal mic/speaker (for analog) or an ADC/DAC (for digital).
The sample "microphone to headphone" object uses a counter and two pins to create a delta/sigma ADC then uses a second counter to output PWM to two pins (normal & inverted). Crude but functional.
AFAIK no-one has implemented a true audio codec for the Propeller. Audio processing does have the advantage of (relatively) low sample rates, but the codecs often require multiplications and floating point - neither of which the Propeller has.
tel interface.pdf
After thinking a lot about this i think that due to the capacitive coupling i will have no problems (i can simple increase/decrease output/input volume in software if needed).
Now i will do the PCB.
I decided to do DTMF decoding in software.
I'm going for a software PLL approach.
Edit:
Theres a much simpler FFT algorithm here :
http://propeller.wikispaces.com/FFT
And i would really need filtering the audio input to the ADC (as least a high pass filter to the ground).