What about digital RIAA?

Jump to Latest
#241
This to and fro is a pity, achieving nothing more than moving lots of air in the room ... Aleksunder, you hear digital sound not being right much of the time, and you're correct in finding it so - but consider the possibility that there may be other reasons for that being the case, other than the ones you're putting forward ... at least a few people have gone down other lines of investigation into what's going on, with excellent results.
 
#258
scott wurcer said:
The output impedance is listed at 100 Ohms the 10K is probably an arbitrary value chosen to measure the specifications. The difference in output level between a 100 Ohm and a 4.3 kOhm vs 10 kOhm divider would be trivial.
Click to expand...


re. the Creek OBH8-SE into the Essence STX;

I've had serious overloading issues into the Essence from the Creek OBH8-SE with several MM cartridges. I regularly see the levels going into the red, (ASIO, always) which means the Essence is seeing peak voltages from the Creek of well over 2V.

Ended up having to use a Rotel RA921 integrated as a 'buffer' - CD-input, ('proper' 47Kohm input impedance) and tape-out.

I enquired with Creek and got a reply from Mike Creek's son (no less) that the Essences's 4.3Kohm input impedance wouldn't be a problem, but clearly it is.

Probably, the Creek uses a "purist" single-ended or 'class-A' architecture (it does sound very good, very good indeed), hence the notice to "load to 10K ohm or more" label on the back of the device.

Oh well.
 
#259
This is exactly what TC Electronic has done with their Konnekt Live audio interface.
It is Firewire based and has some of the best (low jitter) digital


The high quality (and high gain) microphone preamps are used to connect the phono cartridge. It has a build-in RIAA correction plugin and can even work stand-alone (so no computer host needed) once properly configured.

It is discontinued now but you might be able to find one used for an interesting price.
 
#260
An externally hosted image should be here but it was not working when we last tested it.


I'm releasing soon free 'demo' (fully working, no restrictions from me) of this beast done using Cycling74's Max/MSP (this needs their runtime library for v. 6 to run).

What all it does?
Well, behind the features seen in main screen there are just couple math expressions for to calculate the biquad coefficients for selected samplerate in real time using the time constants of each pole/zero and samplerate ofcourse.
With output level adjusting pot, showing there in GUI, I plan to adjust the eq filter gain coefficients so there's going to be enough gain reserve for setups not having suitable pre-amplifier in path (input +18dB, filter +n dB, output +30dB).

Here's an example code for coefficient calculation in C++:
Code: 
  double a0, a1, a2, b0, b1, b2;
  double fs = 96000.0;
//timeconstants (case RIAA):
// frequency -> time conversion 1/(2*pi*fc) (= R*C)
//poles
   double p1 = 3180e-6; // 1/(2*pi*50.05Hz)
   double p2 = 75e-6;    // 2212Hz
//zeros
   double z1 = 318e-6;  // 500.5Hz
   double z2 = 0.0;       // use 3.18e-6 for Neumann pole (50kHz)

double pole1= exp(-1.0/(fs*p1)); 
double pole2 = exp(-1.0/(fs*p2)); 
double zero1= exp(-1.0/(fs*z1));
double zero2 = exp(-1.0/(fs*z2));

a0 = 1.0;                    // 1.0
a1 = -pole1 - pole2;    // -0.967774
a2 = pole1 * pole2;    // 0.0
b0 = 1.0;                   // 1.0
b1 = -zero1 - zero2;   // -1.867057
b2 = zero1 * zero2;   // 0.867481

Example 2:
Code: 
 double a0, a1, a2, b0, b1, b2;
 double fs = 44100.0;
//timeconstants case Columbia N78:
//poles
   double p1 = 159155e-6;   // 1Hz
   double p2 = 100e-6;        // 1590Hz
//zeros
   double z1 = 530.5e-6;     // 300Hz
   double z2 = 0.0;             // - 

double pole1= exp(-1.0/(fs*p1)); 
double pole2 = exp(-1.0/(fs*p2)); 
double zero1= exp(-1.0/(fs*z1));
double zero2 = exp(-1.0/(fs*z2));

a0 = 1.0;                    // 1.0
a1 = -pole1 - pole2;    // -0.958195
a2 = pole1 * pole2;     // 0.0
b0 = 1.0;                   // 1.0
b1 = -zero1 - zero2;   // -1.795148
b2 = zero1 * zero2;    // 0.795177


I have not checked the accuracy of the resulting curve for these coefficients got using above calculation method but AFAIK, these are not as accurate as what Robert Orban and Scott Wurcer has released, especially at lower samplerates.
 
You must log in or register to reply here.
Top Bottom