The RE-303 and more Syntorial praise
2025-01-25
In this post I will talk about the RE-303, a remake of the 1981 TB-303 synthesizer, and I will reflect on what I'm learning from the Syntorial curriculum now that I am further along in it.
The RE-303
The RE-303 is a DIY recreation of the 1981 Roland TB-303 synthesizer.
The TB-303
The Roland TB-303 or "303" for short is the defining sound of acid house but like many other iconic Roland products it was a commercial flop when it was introduced and it only became popular once it was appropriated by underground musicians. The 303 was meant to be a bass synthesizer ("TB" stands for "Transistor Bass") or rather the bass equivalent of a drum machine: an electronic replacement for a real bass player. It was marketed as a sibling product to the TR-606 drum machine. You can synchronize the two and if you look closely you see that they are built into nearly the same enclosure. The idea was that you could program the drum and bass parts of your song into the respective machines and then play along to them on your main instrument.
Source: Synthtopia. A photo from a Roland advertisement showing jazz pianist Oscar Peterson seated behind a piano, with the TR-606/TB-303 rhythm section propped up on the music stand.
The 303 has a peculiar workflow which probably turned off the target audience. It's telling that the Heaven 17 song Let Me Go, one of the few pop songs from the time that used the 303 the way it was intended, uses a single note pattern that gets transposed. They don't use the pitch function of the sequencer (which we'll get back to below).
In acid house, people supposedly enter random patterns into the machine until they hit upon something that sounds good because deliberately programming the machine is too cumbersome. On the one hand the 303 sequencer is reviled but on the other hand some people suggest it helped define the sound of acid house.
The RE-303 is a "replica"
As acid house became more popular second-hand TB-303's became less and less affordable. As I write this original TB-303's are offered for about €3000 on online marketplaces. Companies have been producing clones of the TB-303 since the 1990s, and likewise there have been DIY recreation projects. The DinSync RE-303 is one of the more recent DIY remakes.
What sets the RE-303 apart is the level of authenticity that they strive for. At the heart of the project is a re-creation of the original circuit boards, down to the locations of all the parts and the circuit board traces of the originals. Clones often modernize some part of the original design because some things don't make sense anymore or some parts are hard to get. The 303 for example is battery powered and it has this infuriating user interface. Few clone developers can resist the temptation to improve the user interface.
The RE-303 uses exactly the same power supply as the original so you could power it from batteries if you wanted. And to my amusement, the developers managed to extract the firmware from the original 303. This is non-trivial because it uses a mask ROM, meaning that the firmware is etched into the CPU. The RE-303 comes with a replacement CPU called the "Pixie" RE-CPU which is a modern microcontroller running an emulator for the original firmware. The emulator is 120kB and the original firmware is 2kB.
What drew me to want to build this particular DIY clone was the fact that it is not only evidently a faithful clone of the analog synthesizer part of the 303, but that it also runs the original firmware. (Earlier versions of the RE-303 used CPU's with modern "improved" firmware, but those have been replaced by the RE-CPU emulator and the original binary blob from 1981.)
I need to point out that if you are interested in the 303 sound or its user experience there are many other, non-DIY options available on the market, some costing much less than the RE-303. I chose the RE-303 for myself because I like building DIY synths and because I thought its approach to historical accuracy was endearing.
My RE-303 before I assembled it into its case. The green daughterboard is the Pixie RE-CPU. You can see some IC's are missing in the lower left corner. The original 303 has its battery-backed memory IC's and an address decoder there. The RE-CPU emulates the battery-backed memory using flash storage so we don't need the memory circuitry of the original. This also means that unlike original, the RE-303 does not need batteries to retain sequencer data when powered off.
Living the 303 experience
The first thing I did once my RE-303 was up and running was to go through the original owner's manual. This manual is unusual because it reads more like the syllabus of a course on how to use the 303 than a reference manual. It's nice that if you work through the syllabus you get some practice on how to use the machine, and that concepts are introduced gradually. But it's unhelpful if you want to look up how to do something specific.
The language in the manual is odd at times. It speaks about "memorizing" the bass line which does not mean that you the human operator need to remember it, but rather it means performing the data entry from sheet music into the memory of the 303 sequencer.
Throughout the manual the exercises start from sheet music. You then need to deconstruct the bass melodies into a series of musical pitches and a rhythm, and enter ("memorize") the two into the sequencer separately.
An example from the 303 manual. The first line shows a 4 bar bass melody. The second line shows a table representation of the first 2 bars. The first row of the table is the pitch and the second row represents the rhythm.
This way you can make "patterns" which are usually bars but sometimes just fragments of bars. You can then chain and transpose the patterns together to form a track. In Heaven 17's Let Me Go all you hear is a 1-bar pattern where all notes have the same pitch; the pattern gets transposed according to the chord changes of the song. I don't know if Heaven 17 used a single note melody because they were frustrated with the way the sequencer works but having used it myself now I wouldn't be surprised.
In defence of the 303, as long as you follow its approach, namely to first figure out the bass line outside the 303, then write it down, then do the data entry, then it works fine. It is a simple but pretty flexible design. The problem for me is that if I'm writing a bass part on another instrument then I might as well sequence and use that other instrument in my composition. Using the 303 the way it was intended requires too much planning and deliberation for me.
The other thing for me to try now is to use it in the "acid" way and create more or less random patterns. This is a technique in itself and I haven't gotten very far yet. Not every random pattern sounds good so it's not really all that random. I'll keep trying though.
I can't say if the RE-303 sounds like the original because I don't have the original. It does sound nice and even though I already have instruments that can make similar sounds, the RE-303 clearly has its own distinct voice.
Syntorial
I've previously written about how happy I am to have discovered the Syntorial self-study synthesizer course, even if I'm 11 years late to the party. I am now 24 chapters into the 33-chapter curriculum and I thought it would be nice to reflect on how I like it thus far. The short version is that I still think it's amazing.
A quick recap: Syntorial is a desktop app that lets you do ear training for synthesizer sounds. It has a built-in synthesizer so all you need is a pair of headphones and a MIDI keyboard. Besides short video demonstrations, its main educational technique is a kind of challenge where you are presented with a "hidden" sound that you can play and hear but not see the parameters of. You must then recreate the hidden sound by tweaking the parameters of the built-in synth.
The course is designed for beginners but I cannot judge how well it works if you start from scratch because I am not a literal beginner. What I do know is that I am constantly learning new things even though I already know a lot. The course presents three kinds of information: (1) what do the knobs do, (2) how can you hear what knobs are being used, and (3) how do people commonly use these knobs. When it comes to (1) I am learning almost nothing new. This is why I say I'm not a beginner. But I am learning so much in areas (2) and (3) that I don't mind having things in area (1) explained to me again. It's a deep and thorough course about the basics.
When I learn something from a book I tend to read the parts I find interesting and skip the exercises. You still learn something but you end up with a gap between what you know about and what you know how to do. Because Syntorial is an interactive computer program and not a book, you are forced to do the exercises before you move on to the next concept. And unlike exercises in a book, the exercises in Syntorial are graded and corrected for you. It does a good job of lowering the friction of keeping up with the course work by doing the exercises.
Something I particularly like is that it's possible to solve the challenges 100% correctly. I think this is because the built-in Syntorial synth is digital and offers perfectly repeatable results. Another major factor is that all the knobs are stepped. It's easier to find the right value on a knob that goes from 0.00 to 1.00 in 6 steps then if it takes 100 steps like it would at full resolution. This way you can focus on whether the hidden sound and your solution really sound identical. If they don't then you can keep going until they do. If I did not know that attaining a perfect score is possible then I would give up sooner and learn less in the end. There is something magical about being almost there and then after a last tweak things fall into place and the sounds are identical. It doesn't happen on every challenge, sometimes the final tweak is too subtle, but when it does happen the "falling into place" is very satisfying.
The "On Your Own" challenges are not working for me
The only criticism I have at this point is around the last challenge of each chapter which is called "On Your Own". The idea of "On Your Own" challenges is to practice the concepts you've learned in the chapter outside of the sandbox of the Syntorial app. At first I was skipping these because the app does not check if you actually did this and skipping exercises is what I do. I later decided I should do these anyway because I want to maximize what I learn from the course. But over time I have been finding the On Your Own exercises more and more frustrating.
I think this is because my own synths are all less flexible and feature-complete than the Syntorial synth. I constantly need find a different synthesizer to do the challenge on and often I can't because the synth I need is in use in the song I'm currently working on. Perhaps the On Your Own challenges work better if you work with software synthesizers.
Conclusion
The 303 is a cute but frustrating instrument. I still wholeheartedly recommend Syntorial if you want to get better at synthesizer sound design.