Sandwich Imagineer at Twinbeard. Made Frog Fractions. May or may not have already made Frog Fractions 2 through 5.


This was a Twitter thread from a few years back that ought to have been a blog post, so I'm putting it here. Probably I'll do more of these in the future.

I think of myself as a pretty thrifty person. I lived on $30k/year in the Bay Area for years because I don't need much to be happy. But whenever I look at music gear, I rediscover what I have in common with people who are super into extremely fancy cars. Especially nowadays when there is a glut of really fun music toys in the $200 range. For example, Behringer makes a TB-303 clone, the TD-3, which to my ears is indistinguishable from the original. You can get it for just over $100 new. Can I afford not to??

For context, in 1981 the TB-303 was meant meant to fill in as the bass player when you didn't have one handy. It does an okay bass line, but the squelchy, shrieky sound that it can also produce is what made it a beloved machine in the techno community.


Its unique sound comes from a combination of factors. For example, it uses a three-pole lowpass filter rather than the more common two- or four-pole, and multiple accented notes in a row get progressively more intense. But the most interesting to me is the predictive slides. When a bass player slides from one note to the next, they start sliding early, so they arrive at the new note when the new note is supposed to start. The TB-303 does this too, but this kind of slide is basically unheard of in synthesizers of the past few decades, because synthesizers are now built on the MIDI protocol.

MIDI was originally designed for live performance, so only knows what notes are playing right now, and doesn't know/care about what notes are coming up. So slides can't end on the new note, they can only begin on it. Synths with built-in sequencers could theoretically implement predictive slides, but if you're already implementing MIDI, at that point you'd be doing the extra work of implementing two separate note pipelines, and also you'd have to explain to customers why predictive slides don't work with their hardware configuration.

(Last GDC I ran into the "Chairman of Pitch & Tuning" for the upcoming MIDI 2.0 spec and according to him this is being addressed. Someday this spec will be finalized and some day after that hardware that implements it will be on the market and some day after that you'll start to hear synthpop with better solos and bass lines.)

The TB-303 was a commercial failure, discontinued in 1984. Only 10,000 were made. Later, when it became the foundation of entire subgenres of techno, the rarity and desirability made it extremely expensive. You can easily spend $4000 on the secondhand market. There are quite a few knockoffs trying to sell to that audience. What makes the Behringer TD-3 interesting is the extremely low price and also that for a little extra they're now selling a knockoff of the Devilfish mod.

The Devilfish mod was designed by Robin Whittle in 1993. It's pricey, but with the cost of a baseline 303, what's an extra $2000? This mod takes what the 303 was accidentally good at and leans into it, re-tuning parameter ranges and adding extra controls.

Robin Whittle is still making/installing the mod today. Behringer tried to make a deal with him to make an official Devilfish TD-3. The deal fell through, but they went to manufacturing anyway. You can buy a "TD-3-MO" for about $250, plus the moral cost. (That's what "MO" stands for.)

There's the sense here of a big company screwing over an individual. But also you could think of it as being like republishing an out-of-copyright book. Patent/copyright is a compromise -- the creator benefits from the work during its lifetime, and the world benefits afterwards. I feel very strongly that IP should expire. Robin Whittle didn't in fact patent the Devilfish mod, but if he had, it would've expired over a decade ago. (Under US law. He's in Australia. Not sure how that math works out.)

But all that's not why I'm not buying a TD-3-MO. I'm not buying one because it only does one thing, and after I make a couple songs in a genre, I get bored. Sure looks like a fun toy though! That said, if they make one in Atomic Purple I will have no choice but to get one. Sorry Robin!


You must log in to comment.