Paul T. Schreiber, the man behind Synthesis Technology is a man who you might want to talk to first when you become a eurorack manufacturer. Or when you start DIY-ing a difficult project. Or when you just want to have an insight on instrument-design. He is a great storyteller as well, and has created a lot of truly innovative tools which brought synthesis forward, so we were happy to start a conversation with him.
I’d like to know what made you start working on one of your latest modules, the Quand Temporal Shifter. I’d like to also know what sorts of electronic music you are thinking about – if any, when you are designing your products. Or is that just „pure” music theory?
Way back in 1979 I was thinking about using some ’fancy’ sample & hold parts from a company called PMI (I think they were bought by Analog Devices a few years later). Anyway, these chips were used by my employer (Data General) but were really expensive back then, like $12ea . However, they were very very good and I managed to get 3 samples from them. I was designing some digital VCOs and a clone of the Eventide Harmonizer when I got bored at work, and these chips were for the harmonizer’s de-glitcher in the DACs. I had seen pictures of the Serge ASR, and so I breadboarded up a 3-stage one using these samples. It worked great, but I thought to myself: the droop is still an issue, why not do it with A/Ds and D/As instead and make a ’perfect’ one? Well, the harmonizer clone had a ’really good’ 12-bit A/D that my boss got as a sample, but it was $48ea. The D/As were about $19ea. Processors were relatively cheap, we had drawers of 6502 chips. But to build a full digital ASR back then was about a month’s pay.
I also use musicians that I know and trust that will not just auto-approve anything I do, in fact I *want them* to say „this flat out sucks” if it does, or tell me „the range on this control is not useful”. The last thing I want to do is release products and then go through what seems to the users an endless cycle of firmware updates.
I try not to ’over-think’ the musical context for my designs, above a basic level (this is a LFO, etc). I’m 60yrs old, grew up in the Wendy Carlos/ELP period but now mainly listen to ambient music (Robert Rich is a big influence and contributor). I’m not very ’musical’ myself, although I studied pipe organ a bit but decided it was more fun to build them than to play them (in fact, my design partner’s father is a famous pipe organ builder). I think places like Facebook forums is good idea to get feedback, although you have to apply a lot of filtering sometimes. I do a lot of beta testing, 3-4 months is not unusual, to have up to 6 different people with very wide musical styles comment ton a modules before I ship it.
Lately, some prices got better, and some competitive aspects got worse. The latest flood of eurorack modules has made „hoping for the best” harder as the market has changed, hasn’t it?
I don’t think so, there is always this portion of any product (not just a Euro module) that has this chance to really take off. When I was a kid, this guy put a small rock in a cute box and called it a Pet Rock. He sold MILLIONS of them. You just never know it any product becomes a “hit” or not. Same is true releasing music. But we still keep doing it 🙂
A few years ago on the Muffwiggler forum some people were talking about the lack of ASR’s (i.e. analog shift registers – the editor) (the Serge analog version, with the droop issue) and whining there were not any good ones. So I decided to show them exactly what a ’good one’ would do! The E102 uses true 14-bit accurate A/D and D/As along with true 14-bit level shifting, which is not easy or cheap. But thankfully it’s also not a month’s salary to build! And since the logic needed to do the basic function only required 10% of the total program memory of the microcontroller, we decided to add the voltage-controlled delay between the taps, something never done defore.
This is a hallmark of a SynthTech design: we try to add something never before seen in either HW or plug-ins. This leads into the 2nd part of the question: how do we ’decide’ what to build? And then ’how to build it’?
The fair answer is about 80% market research and 20% „hoping for the best”. It is hard to predict what is a big hit, although in my little corner of the market, I do not try to get close to the number of modules shipped by say Doepfer or Mutable. Rather, I try to approach it as „What is missing that I know how to add value to it?”. Value can mean several things: better audio quality, a new feature (like The E340 Cloud Generator) or a feature that we greatly improve on (the morphing algorithm in the E350).
What made you release a MIDI CV interface as well for example? I mean almost every manufacturer has one now nowadays.
No one to date has one with both USB Host and Device. Doing a Device is easy, but that requires connecting to a computer. The Host part, now that is very difficult in terms of both the HW needed and the USB stack, which is probably 4 *times* the complexity/size of a Device stack. The E620 was put on hold because of my time, but we will ship by the end of the year. I still do not know of anything like the E620 on the market: you are talking about years of development to do it properly. USB is non-trival!
How do you see the latest developments / generations of eurorack modules with plenty of functionality, digital elements like the Disting or other DSP tools?
It’s always nice to see people play catch-up with me (grin). I think what has happened is that the casual hobbiest manufacturer (the majority) have reached an ’end of the Google’ problem: they can’t find any more schematics or service manuals to copy from. It’s all been done to death. So going forward, the innovation will be coming from the trained EE’s that know how to code in DSP and design higher end analog because this stuff just isn’t laying around on the web. You have to know it first. Sure, people have already started buying little development boards and bolting them to a front panel,
I’m surprised no one has stuck on an iPad by now. But this is never cost effective or power-effective. The power consumption of some of these modules is staggering, 100s of milliamps from the +12V rail.
For SynthTech, we plan to offer DSP and 100% analog modules. It’s just REALLY hard to do decent VCFs in DSP. Over last few years, many really fast ARM processors with large memories have been released. Microchip still has the best user libraries to speed code development. We are not “DSP snobs”, we choose the best part for the module in use.
My biggest concern, one that I’m known for on the forums, is the lack of power supply knowledge in Euro. Something as simple as marking where the red stripe goes on a pc board somewhow escapes many of these so called „module designers”. I’m not trying to demean anyone: I’m trying to get everyone on the same page for the customers. Euro manufacturers need to rasise the bar: I try to raise it on every module I design. In fact my new low-cost VCO design I’m doing, called the E330 Multi-Mode VCO, has „The red sripe is -12V and goes here” on the silkscreen. My father used to print catalogs, and he had a big sign in his office that said „INK IS FREE”. The point is: why not put this on the board? It costs YOU nothing, but think of the help it is to your customers. I hear suff like „Well, there is no room on my board” to which I say „Then MAKE ROOM”. I view it as being willfully lazy and sloppy. Why is Euro so inconsistent? What other comsumer electronic family is this way? Put yourself in the customer’s place.
Also, the issue of not using shrouded, keyed headers. I know Doepfer has lead the charge, but that was initially a cost issue and really, you can source Asian shrouded connectors that are maybe $0.23 more than unshrouded. I think this is a small price to pay.
A lot of users ask questions about standards (and realise that there are no such stuff). Also, another issue is build quality. You tend to take the finest components when you build a module. Do you think that people – either users or makers – will be able to see as bad quality pots, switches, jacks may go wrong?
It’s not a matter of ‘being wrong’. There is always a tradeoff in any design for cost versus quality of parts used. In many cases there is a point of diminishing return: added more money doesn’t change the customer’s experience.
Sometimes, the “quality” is in the design itself, not just picking a more expensive part.
There is an art to being really low-cost. I did this for many years at Tandy/Radio Shack and to the extreme at BlackBerry, were saving $0.05 per phone got you promoted.
But there is also a nice “feeling” knowing you are free to pick and chose whatever part you want, and there is so many now to pick from. It’s the path SynthTech has chosen, not everyone can relate.
I am also curious about this VCO design. Anything you could share with the public at the moment?
It is a spinoff from the coding work we did for the E950 Circuit Bent VCO, without the speech. We are always trying out new VCO designs in both HW and SW, I guess that is just my personal thing I like to work on. This new VCO will be shown at Knobcon on Sept 10th in Chicago. I will have details then. The test pc board arrives on Tuesday [which is today 😉 – the editor], I think I will refrain saying more until it actually works.
We were really happy to talk to Paul! See our post on the Quad Temporal Shifter
The Expert Sleepers Disting MK3
multi-purpose module has just been out recently. It is a fantastic modulation and CV source and now offers even more functionality and customisability than ever. We are happy to see that there is growing interest in multi-faceted digital-analog modules like this.
A new 3rd party set of 16 algorithms has just been out for themodule. In the following videos you can see what ‘This Thing’ from Squinky Lab sounds. The pack primarily focuses on sample & hold / track & hold, clock and FM functions! Read more and discuss it in here! Update: The algorithms have only been tested on the MKI so far, however the programmers state that these should be working on MKII’s and III’s as well. Note: You can upload any given number of algorithms to the module’s MK3 version, so you do not need to overwrite your favourite ones if you are interested in trying this out!
Algorithm 1A – Dual Sample & Hold
Algorithm 1B – Sample & Hold Dual Clock
Algorithm 1C – Sample & Hold and Track & Hold
Algorithm 1D – Dual Clock Track & Hold
Algorithm 2A – Probablistic Sample & Hold
Algorithm 2B – Probablistic Track & Hold
Algorithm 2C – Track & Hold ala a-152
Algorithm 2D – Track & Hold ala a-152 with extended functionality (two ThisThings playing together in video!)
Algorithm 3A – 2 stage digital shift register
Algorithm 3B – Dual semitone slope detector is featured in most of the other videos sending triggers to Optomix, etc…
Algorithm 3C – Rise and Fall Detector
Algorithm 3D – Dual Function Quantizer Helper
Algorithm 4A – Precision FM Helper
Algorithm 4B – Octave adder, subtracter
Algorithm 4C – Semitone adder, subtracter
Algorithm 4D – Tuning and Voltage Threshold reference