1. 19
  1. 3

    Uncovering the secrets of how the chip operates may help build more accurate DX7 emulators.

    This is one of the weird things about emulation, especially audio: the bugs, quirks and limitations of the original become desirable later on. There’s nothing to be learned about FM synthesis by studying the DX7 hardware, since the algorithms are well known. There are countless newer implementations of it. What the hardware can reveal is how the DX7 deviated from the ideal through tricks like the way it handled exponentials. I have no doubt that this affected its sound slightly in ways that purists might recognize.

    (I can quibble a bit with the facts in the article. The DX7 didn’t democratize synths in general, there were cheaper ones, especially on the used market. But it was affordable for a polyphonic synth, which made it appealing to mainstream pop/rock keyboardists who liked boring mainstream stuff like chords. The DX7’s own popularity kind of turned against it later on — those 32 presets became so overused because programming your own sounds was beyond anyone but Brian Eno, and people got thoroughly sick of them. By the late 80s the cutting edge of electronic music was moving back to analog synths like Roland’s iconic TB-303 and Jupiter, which were super affordable to starving artists once the 80s musicians had sold/pawned them to buy DX7s. The result was acid house, techno, hardcore… These genres did use a lot of digital gear too, but it was mainly samplers, not synths. And when digital synths came back in the 2000s, most of the innovation was in emulating analog sounds, not FM.)