Bach Goes Different Synths and Signal Chains (not another analogue vs. digital debate)

The Praeludium Nr. VIII aus dem Wohltemperierten Klavier Teil 1 (BMW 853) is a stunning piece even by Bach’s standards. That’s why it was the first Bach composition I learned to play on the piano.

Decades later, I MIDI-tracked and recorded a rather peculiar version of it.

More than another decade later, I took the same project, used a different signal chain, and recorded two mixes.

This can be seen in the context of preparing for my More Than Moore project. That also the second exemplary recording is a cover tune of a long-deceased composer may or may not be considered a surprise.

Analogue Engineering

This was another trial run for the More Than Moore project, which means it has an analogue signal chain, albeit this time only to the output of the mixing stage.

Synth-wise, both lead voices are double tracked: the lively part where I arpeggiated the slow melody from the original is played by a 0-Coast and a Matrix-1000, while the accents which happen about every other bar are a Pulse 2 and a DFAM (thus demonstrating that the DFAM can also do melodic stuff).Synths.jpg

The DFAM gets its control signals from the note, velocity and gate outputs of a Doepfer A-190-2 MIDI interface. To also make the 0-Coast velocity-sensitive, I sent the velocity signal (which it generated from its second Gate/CV MIDI interface pair) to a Dreadbox VCA Eurorack module, to which the 0-Coast’s output was sent.

The chord parts, which consists of just the chords and an arpeggiated version of them were done with the Tetr4 (using the Wagnerian String preset), and an original patch on the new Minilogue.

Bass duties were again covered by the Minibrute. This one was also sent through a VCA, which got its control from the remaining control channel of the A-190-2, to which the velocity info was fed via a Logical plugin in Cubase.

Effects-wise, there’s a slow delay on one lead part (which is three VD400 in series fed back into the first), of which the output was doubled through a T-Resonator, its cutoff frequency modulated by a LFO. There’s also two old Ibanez stereo chorus stompboxes (one on the Tetr4, one highpassed one on the bass), and a slapback echo through an Ibanez phaser on the fast lead part.

StompboxThe composition has two passages which today might be called “breakdown sections”. This was an issue insofar as the Matrix-1000 patch had some incredibly long release – which would’ve ruined said section. Realtime patch editing was out of the question (as it only affects release times of notes triggered after the edit). A second option would have been to control an expander or compressor with an artifical side chain signal. And finally riding faders as the third option. As last time, the reverb was the only faked portion of the signal chain: Zoom’s 1201 with a plate reverb setting.

I opted for the third option, mainly because scoring mixing commands is part of the concept I consider mandatory for the album to come. It was a simple implementation here, but at least on this level of complexity the concept could be proven.

A look at the score with some mix markings in it.

To stick with the analogue signal chain, I opted for a mastering chain consisting of plugins that acted “as if”. Next to SPL’s stunning PassEQ and elysia’s alpha for compression duties, this meant Airwindows’ ToTape (for fake tape) and TapeDither (for a dither which aims to sound like tape as well).


Analogue Synths, Digital Mix

The approach here was simple: record the synths from before into Cubase, and mix it there. I used a similar approach for the effects treatment, using OhmBoyz Delay as the choice for delay, Steinberg’s Studio Chorus for the chorus effects, OhmBoyz Mobilohm phaser and Steinberg’s REVelation algorithmic reverb.

Other processing was limited to Cubase’s channel strip, and dynamic effects were used both on the bass channel (compression) and on the Matrix-1000 (expansion).

Interestingly, I felt the need to do more mix automation than I had done fader-riding in the analogue mix, all of which related to the breakdown section.

For mastering, the choice was WaveLab’s MasterRig with only the standard EQ, compressor and limiter, and MBit+ for dither.

First Version: Software Synths

Although completed last, this was a prequel, and my memories on why I made the creative choices the way I did are somewhat hazy.

For synths, I had chosen two each by Arturia, NI and Steinberg, albeit in the software version only one synth per part was used. Also differently to the hardware/analogue version I used presets exclusively, also owing to the fact that back at the time the instrumentation was just considered a sketch. For statistics, there were three digital vs. three virtual analogue models, and three hardware emulations vs. three original synths.

The fast-paced lead is Padshop, while the slow one is Massive (also through OhmBoyz). There’s two Arturias for the chordal parts, namely the CS-80 for the chords versus SEM V for the arpeggiated version. Prologue (with a preset called “80s Stab” or so) provided the bass.


When revisiting my mix from waybackwhen and comparing to what I had done with the analogue synths, it was interesting to discover that in this version I had used no mix automation, while there was considerably more EQ and dynamic processing in the channel strips.

The dynamic processing in the individual channels would have a very pronounced effect in doing the master: here, I only used SPLs FullRanger (maybe the coolest graphic EQ available), there wasn’t even a limiter, or rather there was, only it didn’t get to do anything. As before, the mastering chain was completed by MBit+ dither.

A Comparison

The title already says it, but I’ll say it again: this chapter is not meant to be a hardware vs. software, or analogue vs. digital, or vintage vs new competition. There’ll still be an opinion which version I like best, but that has nothing to do with one technological approach being better or worse.

There’s one important difference in the approaches: to revisit that age-old project in the DAW, reloading it took mere seconds. It took a few minutes to load the last 32bit version of Cubase so I could load the FM7 patch, export that, load the 64bit version again and import that into FM8. But all that is a small effort compared to connecting various synths, hardware effects, a large mixing console, and setting them up as before (which is only possible if you documented that properly).

One thing I found interesting is that, even though I did the analogue mix before the digital one, it appeared a faster, more fluid way to get my mix done in the analogue domain. Why this is – different audio properties, console vs. DAW with faderbox, or something else entirely – I don’t know.

Apart from that, the advantages or disadvantages to me were more in individual components than in technologies. The OhmBoyz Delay is wonderful and so much more fun than a combination of four stompboxes. Padshop had a preset that hit it for me, and I felt to convincingly get what I was after using my analogue hardware synths. The same is true for the short lead notes, where Pulse 2 plus DFAM instantly gave me what I had tried to achieve with Massive, and failed.

It’s three different versions all right. And while it was clear that the new versions might be vastly different to the original, virtual version, it came as a surprise to me that the two new versions ended up so differently. Make no mistake: they wouldn’t have needed to – I could just have spent more time recreating one mix with the other setup. But this is where each setup had led me. And that on its own is an interesting statement.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s