This page is available in English only

Vintage ring modulator

Diode-based ring modulator as modeled by Julian Parker

(30 Votes)
2.0 (Updated 8 years ago)
69.2kB
May 20, 2013
Reaktor 5 or lower
Effect Other

DESCRIPTION

Reaktor implementation of my friend Julian Parker's model of a diode-based ring modulator (http://recherche.ircam.fr/pub/dafx11/Papers/66_e.pdf).

Sounds much richer than digital multiplication! And you may also change the diode parameters.

In high-quality mode, 8x oversampling is implemented by shaping in parallel all subphase signals of a polyphase decomposition.

v1.2:
Hi2 (CPU-optimized high-quality) mode kindly provided by Donald Philips.

v.1.3
Fixed minor bug in v.1.2.

Future work:
- Review polyphase filters.
- 32 x oversampling, as suggested by the author.
- Optimize Core structures (I'm quite a newbie with that).

COMMENTS  (3)

Robin Parmar
3 years ago
Very nice! Am happy using it without the HQ mode (too greedy). Maybe add a small incremental amount to VL-VB, so it doesn't bottom out at zero.
Louis-Xavier Buffoni
8 years ago
Thanks Donald, I would certainly be interested to know how you made these optimizations! FYI, I used Matlab's FIRPM, which computes the coefficients of an optimal Chebyshev window given a few constraints. I simply chose a transition band with a reasonable amount of coefficients that would yield a "good enough" stop-band attenuation.
salamanderanagram
8 years ago
Hi, nice work. I have done some optimizations if you are interested (I have high-quality mode taking up 5.7% of my CPU, whereas before it took 14.7%). Would also like to hear how you designed the polyphase filters or what your source for that was - is it a windowed sinc filter? E-mail me at salamanderanagram at gmail dot com if interested!
now