TOPIC:

Hi,
I don't know if this has been done before, but I wondered if I could use the FIR filters in the 2x4 HD to do the RIAA equlization and use it as a phono preamp. So I wrote a Matlab script which takes the RIAA curve and translates it into FIR coefficients.
This is what I want the curve to look like, taken from www.bonavolta.ch/hobby/en/audio/riaa.htm :



I generated filters with 1023, 1363 and 2041 taps for 4, 3 and 2 channels. Imported to the plugin they look like this:



It seems to work. With the vinyl player connected to the analog inputs and the input and output gains bumped up a bit I can listen to my LPs, though I haven't compared it to a dedicated phone preamp yet. Fewer coefficients lead to less gain in the low frequencies which might be an audible problem.
I also attached the matlab script and text files with the coefficients if someone wants to try this. Let me know what you think.

Woww.. Very neat indeed! Let us put a sticky on this one...

We hope to hear back from the community on this! Thanks again for sharing your great work!
DevTeam

Hi

As soon as I stumbled onto the minidsp site my first thought was about using it as a digital phono stage. So this is great! You probably need to build a simple gain stage to bring the output of the cartridge up to at least .25 or .5 volts, maybe full line level given that cartridge outputs are in the low mv at best. The MiniDSP guys can probably advise you on that. Also directly into the miniDSP you are almost certainly not loading the cartridge properly. Though come to think of it maybe it doesnt matter in this circumstance, not sure.

I would try it with a simple opamp stage as can be found here in figure 3 with the eq removed.
sound.whsites.net/project25.htm

Later If you want to get ultra fi with it I would suggest adding a balanced output jack to your turntable and then building a differential low noise input stage as that has great CMRR and will cancel out a lot of clicks pops, and even some surface noise.

Hi,

Very neat indeed!

Thank you!

You probably need to build a simple gain stage to bring the output of the cartridge up to at least .25 or .5 volts, maybe full line level given that cartridge outputs are in the low mv at best. The MiniDSP guys can probably advise you on that. Also directly into the miniDSP you are almost certainly not loading the cartridge properly.

You are right, the signal should be buffered correctly. I will look into building one of the circuits suggested on the page you linked. Thanks.


In addition I looked into doing the RIIA eq with the PEQ of the 2x4hd. This has been done before and looks like it only takes one biquad: www.musicdsp.org/showone.php?id=122. Though the author writes "These coefficients were found by matching the magnitude response of the s-domain transfer function using some proprietary Matlab scripts", so I don't really know what exactly happened to get the coefficients. So I did it myself in Matlab using its bilinear transform function, also referencing www.hagtech.com/pdf/riaa.pdf for the Laplace transform and time constants.

These are the coefficients I get:They only differ from the ones for 96KHz at musicdsp.org because I normalized the 1KHz point to 0dB.

This solution is cheaper and it also matches the RIIA-curve in the low frequencies. Therefore you don't get a constant group delay.
Magnitude and phase of the IIR filter:

Group delay:


In comparison the FIR (The phase just does not look linear beacuse of the logarithmic frequency scale):

Group delay:


I also attached the Matlab script if someone is interested.
Also when importing the coefficients to the plugin it took me way too long to realise a1 and a2 are already negated, although it's obvious in the 2x4hd manual under 5.4.1

Luddi wrote:

Hi,

Very neat indeed!

Thank you!

You probably need to build a simple gain stage to bring the output of the cartridge up to at least .25 or .5 volts, maybe full line level given that cartridge outputs are in the low mv at best. The MiniDSP guys can probably advise you on that. Also directly into the miniDSP you are almost certainly not loading the cartridge properly.

You are right, the signal should be buffered correctly. I will look into building one of the circuits suggested on the page you linked. Thanks.

I also attached the Matlab script if someone is interested.
Also when importing the coefficients to the plugin it took me way too long to realise a1 and a2 are already negated, although it's obvious in the 2x4hd manual under 5.4.1

And... did you get it to work? how did you solve the gain problem?

Your matlab script works GREAT!  Thanks so much.

-Jim