A while back, I teamed up with David Yeh and worked on a software model of the SP12. The model was intended to emulate the sampling engine and low pass filter in the SP12. We had no intention of modeling the SP12’s sequencer. The goal was to build a software model that could input an audio file, modify the file so that it sounds exactly like it was sampled by the SP12, and then save the affected file in wav format. The user would have the option of detuning the sample (which creates the classic ringing sound of the SP12), and turning on/off the SP12’s low pass filter. The tuning would exactly match the 32 tuning steps available in the SP12. The project was very successful. David’s paper on the subject is available
More recently, I took some time out and rebuilt the tuning of the algorithm to model my SP1200. I thought it was perfect, so I had the new tunings tested side by side with the SP1200 by an industry leading golden ears plug-in developer. We tested several sample sets by sampling into the SP1200 and running the same sample at the same tuning through our modeling algorithm. The two output files are indistinguishable.
The algorithm was coded in Matlab by David. This made it difficult (i.e. expensive) to port to a platform that could be easily run on Windows and Mac. I think I found a way to create a simple GUI that can be used on any machine without the need for Matlab. It won’t be pretty, but it will sound very pretty. I’ll post the program if I work it out.
One important thing to take away from this project: The low sample rate causes a digital sampling error called aliasing. This is what causes the classic SP12 ringing sound. When aliasing occurs, frequencies appear in the output sound that were not present in the original sound. From a mixing/mastering point of view, this is particularly useful for drums. When the drums contain a wide range of frequencies, the can be heard on any speaker system. It doesn’t matter if your system has no bass, because the presence of the kick can still be heard in higher frequency ranges due to the aliasing errors. This concept is the basis for the Flesh Modulation effect (it adds clean analog broadband noise to drum samples by using FM radio). The drums can be heard on any system.
Read more..