agn@cmu-cs-k.ARPA (Andreas Nowatzyk) (03/31/85)
After receiving Sony's D-5 schematic, I was pleased to find that
it is easy to extract the digital sound. A buffer chip (a 74ls07
in a flat package) needs to be added to buffer the desired 5
signals (you don't want to kill the EDC chip when touching the
digital output connector while not being grounded). BTW: adding the
connector is the most difficult part (Warranty? What warranty!).
The 5 signals are:
1. A serial data output
2. A corresponding clock
3. A word clock (indicates which bits are in a word).
4. A L/R signal
5. A muting signal
This opens the path to more ambitious home-audio projects. What I
have in mind is a programmable audio-processor that can do things
like equilization (including phase), 'holography', etc. The preliminary
design below is subject to discussions (and change without notice).
Based on AMD's 29510 16bit multiply accumulator, each channel should
be able to perform 256 *+ operations on each sample (about 85 ns cycle).
A fast, static memory is organized as 4 shift-registers (1024 deep,
16 bit wide). The L/R inputs are fed into 2 of the shifters (LI,RI).
The result of the accumulator (after 256 mpy/add steps and after
rounding to 16 bit) is the digital output of the unit. It is also
fed into the other 2 shifter (LO,RO) to allow recursive filters.
Programming the unit is done by loading a table with the address
of a tab to a particular shifter and the corresponding weight:
Example:
result = a*LI[0] + b*LI[15] + c*RI[22] + d*LO[2] + ...
for some 16bit constants a,b,c,... . LI[15] would mean: take the
value of the left channel after a delay of 15 * 22 usec. This
allows a fairly general class of filters to be programmed. It
also allows delays of up to 22 msec (that is about 7 meter
of acustical distance, far more than necessary for 'holography').
The access to the other channel is required for certain operations
and the recursive part makes it easier to cope with low frequencies.
Some kind of PC will be required to download the tables and there
is a need for signal-processing software to compute the tables for
a desired frequency response.
The digital output requires an audio DAC unit with the usual LP filters.
This could be Sony's DAS-702ES DAC unit (not yet on the market).
Is someone on the net is planing to build a DAC 'right' (that is with
digital filter, 4x sampling rate and *16* bit DAC's)?
I plan to complete the design as a low priority project, so it
will take several+ months. The components are not too cheap and a
multilayer PC-board is necessary (85 nsec cycle !). Any ideas
concerning alternate designs or the software part are welcome.
-- Andreas Usenet: ...!seismo!cmu-cs-k!agn
Arpa: Andreas.Nowatzyk@cmu-cs-k.ARPA