Adaptive Filter | ADPCM | ADSP | ADSP-2181 | Aliasing | AMR | Anti-Aliasing | ARMA | Autocorrelation | AutoCovariance | Beamforming | Bessel | Blackfin | Butterworth | C6713 | CCS | Chebyshev | CIC Filter | Circular Convolution | Code Composer Studio | Comb Filter | Compression | Convolution | Cross Correlation | DCT | Decimation | Deconvolution | Demodulation | DM642 | DSP Boards | DSP/BIOS | DTMF | Echo Cancellation | Equalization | Equalizer | ETSI | EZLITE (Ez-kit Lite) | FFT | FFTW | FIR Filter | Fixed Point | FSK | G.711 | G.723 | G.729 | Gaussian Noise | Goertzel | GPIO | Hilbert Transform | IFFT | IIR Filter | Interpolation | Invariance | JTAG | Kalman | Laplace Transform | Levinson | LPC | McBSP | MIPS | Modulation | MPEG | Multirate | Notch Filter | Nyquist | OFDM | Oversampling | Pink Noise | Pitch | PLL | Polyphase | QAM | QDMA | Quantization | Quantizer | Radar | Random Noise | Reed Solomon | Remez | Resampling | RTDX | Sampling | Sharc | TI C6711 | Undersampling | Viterbi | Wavelets | White Noise | Wiener Filter | Windowing | XDS510PP | Z Transform

Hi,
We're implementing a VoIP-device for a university project. To handle the 
sound part of the device we're using a Blackfin 532 from Analog Devices 
  As a start we want to provide G.711a- and ulaw for the user and later 
if time permits a lower bitrate codec such as g.726.

Our problem is that we're stuck in selecting an external audio codec.

What we would like to see in an external audio codec is:

- should run on 3.3v (at least this is preferred but if needed we will 
add a 5v supply on our PCB)

- configureable through the Blackfin SPORT (others seem to use SPI, but 
the SPI on the blackfin is allready running in slave mode connected to 
our master MCU)


We have been looking at ADI's selection of external audio codecs and so 
far we've found the "AD1981BL - Low Voltage AC'97 SoundMAX Codec"
http://www.analog.com/en/prod/0%2C2934%2CAD1981BL%2C00.html
which appears to meet our requirements.

However as we're both very new to the world of DSP we want to make sure 
that the AD1981BL will be able to do the G.711a- and ulaw sampling 
before we design our development PCB. As I understand G.711 encoding is 
converting 13 bits(A-law)/14 bits(U-law)linear PCM samples to 8 bits 
compressed samples and vice versa for decoding. I would just assume that 
the external codec should then just be able to sample signals at 8000 
samples/second with at least 14 bits resolution.

The documentation for the BL1981L states that it will do Full-Duplex 
sample rates from 7040 hz to 48000hz in 1 hz steps at 16-bit resolution 
so to me this seems ok.


But is that all I have to take into account?

Are there any other important parameters that I must consider?

Any pointers, pro/cons on choosing this codec for a VoIP solution?

Best regards
Nikolaj

______________________________
What are the most bookmarked web sites about "linux tutorial"? Find out here.

You are using a Blackfin just to do G.711? Just how big a hammer do you 
use to crack nuts? :-)

Most cheap telephony codecs actually output G.711. Most times in VoIP a 
DSP is decoding G.711 from a codec so it can re-encode in something more 
sophisticated.

Steve


Nikolaj K. wrote:

> Hi,
> We're implementing a VoIP-device for a university project. To handle 
> the sound part of the device we're using a Blackfin 532 from Analog 
> Devices  As a start we want to provide G.711a- and ulaw for the user 
> and later if time permits a lower bitrate codec such as g.726.

______________________________
Are you a member of DSPRelated.com? If you are, you can now access a powerful tool to discover new web sites. Details here.

Steve Underwood wrote:
> You are using a Blackfin just to do G.711? Just how big a hammer do you 
> use to crack nuts? :-)

Bigger is better ;-). No seriously as we are limited in time for this 
project we focus more on the networking area of the WLAN VoIP phone - 
that is developing a stable IP-stack and SIP-stack and a WLAN-driver. 
However at a later point some other students at the university with more 
DSP knowledge than us might want to take over the project and implement 
Codecs like iLBC, Speex, G.72x - perhaps even MP3/WMA for streaming 
netradio - which can't be done by a simple telephony codec. Therefore a 
DSP like Blackfin is put in between the external codec and our master 
processor.

As I understand the Blackfin SPORT will do G.711 in hardware so at this 
point where no sofisticated codecs has been implemented the Blackfin is 
merely a bridge from the external Codec to my master-processor.

However I'm still in doubt if the AD1981BL will meet the requirements 
that might exist for delivering samples to the Blackfin SPORT for G.711 A/U.


/Nikolaj

______________________________
What are the most bookmarked web sites about "fpga"? Find out here.

Hi,

I'm pretty sure that the SPORTs on the Blackfin are able to receive/transmit
8bit uLaw/ALaw encoded data. So the DSP would be used for something where
you really need the computational power.

For instance, since the ol' times of the 2181, in the EzKit lite, the AD1847
could deliver and receive 8 bit aLaw/uLaw data, and the 2181's SPORTs could
handle it. IOther AC'97 codecs used in ADI's EzKits could do the same: the
AD1819, the AD1881A, etc.

Anyway, it seems that you want a codec that could provide enough bits and
adequate sampling rates for speech processing... G.711 seems to be just the
starting point, and you want the same hardware setup for later moving to
lower bitrate speech processing/compression algos. Am I right? If that's the
case, I think that any current audio/speech codec could suit your needs.
AC'97 codecs are interesting in the sense that they can be configured from
the serial ports, no need for external control ports like I2C or SPI. The
AD1881A could do the job, and is a 3.3V part, at least for the digital
section. It seems that the AD1981 could do it too (I haven't looked at the
datasheets, but I had experience with the AD1881A), and both seem to have
the same price tag, at least at Digikey, around US$7.9 in unit quantities.

Regards,

--
Jaime Andrés Aranguren Cardona
j...@nospam.sanjaac.com
SanJaaC Electronics
Soluciones en DSP
www.sanjaac.com

(Remove "nospam" from e-mail address)

"Steve Underwood" <s...@dis.org> escribió en el mensaje
news:cmmlg1$m7k$1...@home.itg.ti.com...
> You are using a Blackfin just to do G.711? Just how big a hammer do you
> use to crack nuts? :-)
>
> Most cheap telephony codecs actually output G.711. Most times in VoIP a
> DSP is decoding G.711 from a codec so it can re-encode in something more
> sophisticated.
>
> Steve
>
>
> Nikolaj K. wrote:
>
> > Hi,
> > We're implementing a VoIP-device for a university project. To handle
> > the sound part of the device we're using a Blackfin 532 from Analog
> > Devices  As a start we want to provide G.711a- and ulaw for the user
> > and later if time permits a lower bitrate codec such as g.726.
>

______________________________
Are you a member of DSPRelated.com? If you are, you can now access a powerful tool to discover new web sites. Details here.

Jaime Andrés Aranguren Cardona wrote:
> Hi,
> 
> I'm pretty sure that the SPORTs on the Blackfin are able to receive/transmit
> 8bit uLaw/ALaw encoded data. So the DSP would be used for something where
> you really need the computational power.
> 
Correct - at least according to the datasheet. I haven't done real life 
testing with SPORT hardware G.711 companding - I've only done G.711 in 
software using the BF533 om my EZKit.


> Anyway, it seems that you want a codec that could provide enough bits and
> adequate sampling rates for speech processing... G.711 seems to be just the
> starting point, and you want the same hardware setup for later moving to
> lower bitrate speech processing/compression algos. Am I right?

Yes this is indeed the case.

> If that's the
> case, I think that any current audio/speech codec could suit your needs.
> AC'97 codecs are interesting in the sense that they can be configured from
> the serial ports, no need for external control ports like I2C or SPI.

The abillity to configure the codec through the SPORT is very important 
to us which was also why I started to explore the AC'97 codecs.

> The
> AD1881A could do the job, and is a 3.3V part, at least for the digital
> section. It seems that the AD1981 could do it too (I haven't looked at the
> datasheets, but I had experience with the AD1881A), 

Yes, but as you say only the digital supply is 3.3v - the analog supply 
requires 5V according to the datasheet. This is however not a big issue 
as we could just put at 5v regulator on the board, but since the 
AD1981BL is a true 3.3V part I've decided to take my chances with it and 
save the PCB space.


> Regards,

Thank you very much for your input. I'm now more confident in 
implementing an AC'97 based codec for my design.

Best regards
Nikolaj

______________________________
What are the most bookmarked web sites about "fpga"? Find out here.