Symptoms and troubleshooting steps

The two Arturia products I had installed were Emulator II V and Farfisa V. Both of them install VST3 plugins as well as standalone applications. I first tried loading the plugins in Cubase. In a new Cubase project I was able to select each of the plugins from my list of virtual instruments, however the plugin UIs never appeared and I wasn’t able to play any sounds with them. After saving and reloading the Cubase project, I noticed that instead of listing the plugin names on the instrument tracks, there was a little message like “Couldn’t load plugin!” Next I tried launching the standalone versions of each instrument. Neither instrument loaded, and I was stumped.

Here are all the things I did next:

• I monitored processes in the Windows Task Manager, and I could see that when I attempted to launch either program, a process named ArturiaSoftwareCenterAgent.exe would run for a few seconds and then quit, and then the program I launched would quit as well.
• I opened Arturia Software Center and made sure my V Collection 8 license was activated locally. I even de-activated and re-activated the license and tried the apps again with no success.
• I uninstalled and reinstalled both products. Same results.
• I submitted a support request at Arturia’s site and continued with my day.

When Arturia’s support staff eventually got back to me, their advice was to uninstall and re-install the plugins (which I had already tried, and already mentioned that I tried in my original support ticket). I did it once more for kicks, but of course this didn’t resolve the problem. Realizing that Arturia probably wasn’t going to help me either, I decided to do some further digging.

I ran the Sysinternals Process Monitor to see what kind of activity occurred on my computer when I launched one of the Arturia standalone apps, and I noticed that Emulator II V was accessing a file named log.conf in a directory nested in the hidden ProgramData folder. The exact path of the file was C:\ProgramData\Arturia\Emulator II V\resources\log.conf. “Could this be a logging configuration file?” I wondered.

And that’s exactly what it turned out to be! Arturia turns out to have a pretty sophisticated logging framework for their plugins that can be enabled in the field (rather than requiring special debug binaries). There are numerous ways you can customize the logging on a per-application basis, but if you simply want to see all available logs, you just have to uncomment and edit two lines of the log.conf file.

• Uncomment the OutputFile line by removing the preceding “//” marks, and edit it to the path and filename you would like it to save. IMPORTANT: This needs to point to a directory that your Windows user account is able to write to without admin approval. I chose the root of my user profile. Example: “OutputFile C:\Users\joshm\EmulatorLogs” Note: You don’t need to add a filename extension; the app will use the filename you specified just as a prefix and then add a date and timestamp to it, like “EmulatorLogs_2026_May__7_09–51–48” It’s just a stanard text file so you can open it in any editor.
• Uncomment the DebugAllFlag line (remove the “//”) and set the value to 1.

After you save the file, you can run the app again and you will find a detailed log file at the place you specified. And this is where I discovered what was really going on. I saw two suspicious lines in the resulting log file:

| 50 | T32716 | description | error | [XML] Error on parse file C:\ProgramData\Arturia\Emulator II V\resources\Emulator II V.xml : Failed to parse xml file ‘C:/??/Volume{1d393a8f-7798–4b3b-bfdf-0d80ba60c197}/Arturia/Emulator II V/resources/vc8_common_vst_params-generated.xml’(C:/jenkins/root/workspace/VC/release/emu/wrapperlib/src/ArturiaProcessor/ContentLoader.cpp:755)

and

| 55 | T34516 | database | error | Exception caught in PresetBrowserController::populateDatabase: boost::filesystem::directory_iterator::construct: The filename, directory name, or volume label syntax is incorrect: “C:/??\Volume{1d393a8f-7798–4b3b-bfdf-0d80ba60c197}\Arturia\Presets”

The Arturia code failed to build the correct file paths to locate files the instrument needed to operate. Why was this? Well I immediately had a suspicion.

File system foibles

The computer I am currently using as a DAW PC originally came with a 1TB M.2 NVMe drive for storage. I knew I was going to need a lot more space than that so I acquired an additional 2TB drive to use as my main system drive, and formatted the original 1TB drive to use as extra storage. In all prior computers I built, I would use a RAID 0 storage configuration which allowed me to install a number of physical drives and combine them into a single Windows volume. (So even if I had five different SSD drives in my computer, Windows only saw one giant C:\ drive.) A problem I faced with my new computer, however, is that the B760 chipset that my motherboard uses does not support enabling RAID 0 with NVMe drives.

I researched some of the alternatives to RAID 0 that work on Windows 11, but each of them seemed to have as many disadvantages as they had advantages. I ended up choosing the mount point folder method to make Windows use my 1TB drive as an extension of my system volume. Full instructions are at the page I just linked, but basically, you create an empty folder on the extra storage device, and then you tell Windows to use that directory as if it was part of a different physical drive.

So, in my case, I created a folder called “ExtraSpace” on my 1TB drive and I configured Windows so that it appears as C:\ExtraSpace. The problem (for Arturia) is that you can no longer assume that the C: drive comprises only a single volume.

And here’s the thing. When I first installed installed Arturia Software Center, I chose a directory under my mount point folder, because I didn’t want any sample content being installed on my C: drive. So instead of the default “C:\ProgramData\Arturia” I entered “C:\ExtraSpace\Arturia”. In most cases this shouldn’t be a problem, but I think Arturia might be using some logic or APIs that assume that any directory under the system drive must be on the same logical volume- which is not a safe assumption to make on Windows 11.

So, when the Arturia apps were building the paths to that XML file and the Presets folder, they were trying to access files that existed on a Windows volume they didn’t know about, and their access attempts failed as a result.

The solution

Once I realized what was really happening, the fix was simple:

• I uninstalled all Arturia products, including Arturia Software Center.
• I deleted the “Arturia” junction point shortcut that ASC had created under C:\ProgramData at install time. (It creates this shortcut only when you specify an alternative Resources directory.)
• I reinstalled ASC, and kept the Resources directory at the default location (C:\ProgramData\Arturia).
• I then re-installed both of the plugins I wanted to try out: Emulator II V and Farfisa V. And they both work now, as standalones and as virtual instrument plugins!

I replied back to Arturia with my findings, and asked them to resolve the ticket. I’m just putting this out there in case another mount point user experiences the same problem, since I didn’t find anyone else online with the exact same symptoms or solution.

Symptoms and troubleshooting steps

With both synthesizers I found I had the same problem: They each powered on, and I could get audio out of them by playing their internal sequencers or switching their bottom row of buttons to keyboard mode (so I could use the buttons like a virtual MIDI keyboard), but when I tried to play any MIDI into them from my DAW, I saw no evidence that they were receiving MIDI, and they certainly weren’t playing anything.

Things I tried:

• I plugged the audio and MIDI cables from a known-working synth (a Roland XV-3080) and connected them to the KORG synths. Playing the same sequence with the same settings that worked on the Roland had no effect on the KORGs. No joy.
• I tried using both the DIN MIDI and USB MIDI inputs for each synth. Nada.
• I checked and double-checked that my DAW was sending to the correct MIDI port and channel and that the synths were set to the same channel as my DAW. Bupkis.
• I performed the factory reset sequence on both synths to ensure they didn’t have any weirdo custom settings causing problems. Nope!
• Then I took my PC out of the equation and ran a MIDI cable from my Arturia Keystep 37’s MIDI OUT port directly to the Minilogue XD’s MIDI in- and that worked!
• Next I did the same thing with my Native Instruments Komplete Kontrol S49 (I had to do a little configuration with MIDI-OX in order to set this up) but that worked too!
• Finally I fired up Ableton Live 11 Standard and configured it the same way I had Cubase set up- and that also worked. Was Cubase the problem all along?

Special note about the KORG MIDI drivers

Here’s a little sidebar. While doing my web searches I ran across a number of posts and videos that talked about issues with KORG’s own USB MIDI drivers. I am a Windows user, and Windows has historically been kind of notorious for having audio and MIDI support that simply didn’t meet professional standards. Steinberg created ASIO to make up for Windows’ sub-par digital audio support, and many MIDI device manufacturers (like KORG) wrote their own MIDI drivers to provide features such as multi-client support to make up for Windows single-client limitations. But over the years, Windows has evolved. Recent major Windows releases have incrementally improved both the audio and MIDI support, and in a recent Windows 11 update, Microsoft has actually surpassed the technology available in a lot of these third-party MIDI drivers (including KORG’s).

I am running Windows 11, and now both KORG and Microsoft are urging users not to install the KORG USB MIDI drivers on their systems. So, my problems were not related to the KORG drivers, and I was fairly confident that installing them would probably make things worse.

Fixing the Minilogue XD problem

Once I knew that the problem was Cubase-specific that allowed me to narrow my web searches, and I found this post, by a person having the same kind of trouble with a different KORG synth. This was where I learned about the “Project Synchronization Setup” dialog box in Cubase. I’m not entirely sure what it all does (it’s not even documented very well in the Cubase documentation), but it was there that I discovered what I needed to do in order to fix my Minilogue XD.

From the Cubase project window, go to Transport > Project Synchronization Setup > Destinations > MIDI Clock Destinations and select the MIDI port connected to your Minilogue XD (whether you’re using the synth’s USB connection or a DIN MIDI connection). After I clicked OK, I was able to play a sequence from Cubase into my Minilogue XD with no problems.

I don’t understand why it worked. I just know that it did. But even after making this change my KORG Opsix still wasn’t working!

Fixing the Opsix issue

Despite being a bit crestfallen upon discovering that the Project Synchronization change didn’t fix my Opsix, after some more specific web searching I learned about the Opsix’s “Rx Transport” feature.

The factory default settings enable Rx Transport, so I had to disable it in order to play MIDI from Cubase into the Opsix. On the Opsix, press SHIFT + GLOBAL, then select the MIDI page, locate the Rx Transport setting, and set it to Off.

Once I did that, Cubase worked fine with Opsix. Also, the “Project Synchronization Setup” change didn’t affect the Opsix one way or the other. If you own both synths, they will both work with it in place.

The inciting incident

Sometime last year I wanted to check out the Lo-Fi Glow Kontakt library, that was included with my purchase of Komplete Ultimate 13 Ultimate. So I installed Lo-Fi Glow, launched Kontakt 6, and was surprised that I didn’t see the library listed in my available instruments. I later opened up Komplete Kontrol, where I located Lo-Fi Glow- only to have the program inform me that I needed a newer version of Kontakt in order to use the instrument. But how could a library included in Komplete 13 not run in the version of Kontakt that came with that bundle?

Well as it turns out, sometimes when Native Instruments updates their content libraries, the updates end up changing the file formats in a way that they require newer versions of their host applications than they were originally designed for. So while Lo-Fi Glow originally worked with Kontakt 6, it no longer does due to updates published after my version of the Komplete bundle was published. And Native Access doesn’t care what versions of which programs you own or have installed when updating content libraries; it simply gives you the latest available version. The good news is that Native Instruments does offer free “player” versions of Kontakt and some other programs, for loading licensed instruments- so I still have the ability to use all my libraries. I just need to know which ones require one of the player apps instead of my fully-licensed version of Kontakt 6.

So anyway, this got me thinking: How can I know which of my Native Instruments content products require newer versions of their host apps? I was already familiar with the XML files that Native Access uses to manage Native Instruments-compatible products (I had learned the file format when writing my Ultimate Plugin Tool quite some time ago), and I discovered that the XML files indicate the minimum versions of applications required, where applicable.

Once I learned how to divine this info from the XML files I was able to write an open source web-based app that parses the files and shows you the required application versions for every product in the Native Instruments database. I call it the Native DB App.

About the app

The app currently has two forms: A web-based version that uses XML files that I have uploaded to my source repository, and works with any device that can connect to the web, and a local version that uses XML files on your own Windows or Mac OS computer.

You can access the web version here: UltimateOutsider’s Native DB App

I’ll discuss how to get the local version in the next section, but first let me give you a quick walkthrough of the app.

First off, if you’re greeted with a message that says, “This app has gone to sleep due to inactivity,” don’t fret. The web app is hosted on a free community platform and they shut down apps that aren’t in use to save resources. All you have to do is click the “Yes, get this app back up!” button and in a few seconds it will start up.

When the app first loads it will look for Native Access XML files. For the web app, it will look in an online directory I’ve already populated. For the local version, it will search the following directory: “C:\Program Files\Common Files\Native Instruments\Service Center” If you are running Mac OS, or if you have installed Service Center in a different location, the program won’t find any XML files and it will prompt you to enter the path to your Native Access files. (On Mac OS the default path would be “[Your System Drive]/Library/Application Support/Native Instruments/Service Center”) Once you enter a new path to search, click the “Search for XML files” button. If the app locates files in the path you specified it will come up looking something like this:

The default view shows a table of products that includes the name, company, and product type of every Native Instruments-compatible product located in the various files.

On the left sidebar of the app you will find three ways to filter the content under “Display Options:”

• Filter by dependency: You can select an app from this list in order to only show products that depend on the selected app. For example, choosing Reaktor will narrow the list only to Reaktor-based instruments, or expansions that include Reaktor content. Choosing an app from this list adds a column to the table that lists the minimum required version of the selected app for each product. Select None to reset this filter.
• Filter by company: Only shows products released by the selected publisher. Select All to reset this filter.
• Filter by product type: Narrows the list down to certain kinds of products (apps, plugins, content libraries, etc.). Select All to reset this filter.

If you float over any table with your mouse, a little tooltip menu appears with four options.

The tooltip menu options are:

• Show/Hide columns: You can choose which columns to display when viewing the table.
• Download as CSV: Downloads a comma-separated file to your device with the currently displayed contents of the table.
• Search: Pops up a little search box for locating specific items in the table.
• Fullscreen: Expands the table to fill your whole browser window. (Click it again to return to the original size.)

Another detail worth noting is that you can sort the table in any order you want by clicking the column headers. Clicking the same header twice reverses the sort order.

If you are running in a small browser window or on a smaller device like a phone, you might have to expand the Display Options menu in order to filter the results. Look for the “>>” symbol and touch it to open the filter list, and then touch “<<” to collapse the menu after you’ve made your selections.

While I wrote the app specifically so I could look up which of my libraries required certain versions of Kontakt or Reaktor, I’ve made some interesting discoveries while playing with the different filters and features:

• I discovered a number of free Kontakt player-licensed products by simply searching the database for: “free” A lot of them are still available for download.
• I’ve discovered some interesting commercial instruments I never would have found otherwise by sorting the lists by company name and then doing web searches for the products.
• It’s helped me discover content in expansions I already own (for example, sorting by Massive shows all products that include Massive presets).

The nitty gritty

I wrote this app in Python as an exercise in learning the pandas data analysis library and the Streamlit web app framework. I have published the project as open source under the AGPL 3.0 license.

Installing and running the app locally

At the time I am writing this, I don’t recommend attempting to run this app locally unless you are familiar with Git and Python. I am looking into making an installer that will make it easier for non-engineers to run the app on their own computers. I will update this post if I ever release an installer.

But for now, go to the native-db-app repository on GitHub and clone the code to your local drive. The readme file at the repo includes info on the specific versions of Python and the relevant libraries you need in order to run the app, and also lists the commands for launching the application.

It will open in your default web browser. If you don’t have Native Access installed or if you are running on an OS other than Windows, the app will prompt you for a directory location where your Native Access XML files reside. (You can also point it to the contents of src/webdata in the repo, as those are all the same XML files consumed by the web version.)

Random notes

• I will probably make minor updates as time passes. In fact I’ve already added a new “Version” column to the tables to indicate the latest known version of each product, so the screenshots in this post are already slightly outdated!
• Different versions of individual content libraries can have different dependencies (like the Lo-Fi Glow example I gave before). When filtering the list by dependency, the app only looks up dependency info for the most recent version of a given product. So if version 1.0 of Lo-Fi Glow supported Kontakt 6 but version 11.0 requires Kontakt 7.6.0, you will only see a dependency for Kontakt 7.6.0 reported in the table- no entries for previous versions.
• If, like me, you have multiple generations of Native Instruments XML files on your computer, there might be multiple versions of the same product represented in those files. Native DB App only displays info on the most recent version (the one with the highest “version” attribute in its Product element).
• If you can think of anything else interesting we might be able to figure out from this information, let me know- I might be able to add it to the app!

In 1982 Roland released the SH-101, a relatively affordable portable analog monosynth that could run on battery power and even transformed into a keytar with the optional MGS-1 grip set. This keyboard had a similar voice architecture to most of Roland’s JUNO synths, although it lacked polyphony or any onboard effects- and it also lacked MIDI support.

The synth’s portability and comparatively low price made it an early staple in the 80s electronic music scene, but the absence of MIDI and a quite limited feature set make it not a particularly sought-after piece of vintage gear today. But in 2017, Roland released a significantly improved digital recreation, the SH-01A, capturing the character of the original synth while including a large number of sound design and quality of life improvements.

Notes on the hardware

Perhaps because Roland anticipated artists would be strutting around on-camera and on-stage with their SH-101s strapped to their necks, Roland released the original synth in three different colors: Red, Blue and Gray. Roland cheekily also issued three color models of the SH-01A, although the reds are less common, and the blues are quite difficult to find on the used market.

There are so many differences between the SH-01A and the original I’m just going to focus on things that directly affect sound design or studio integration:

• Three additional new voice modes, besides the default monophonic mode: 4-voice polyphonic, 4-voice chord mode, and an extremely limited Unison mode
• MIDI USB and DIN MIDI in and out
• USB audio is also supported (the device functions as an audio interface when connected to a computer)
• 64 memory slots for patch presets (the original synth had no patch memory)
• Six oscillator octave positions instead of the original’s four.
• A 4-element display for navigating the menu system.
• The LFO has two additional waveforms (sawtooth and reverse saw), bringing the total to six.
• An optional “advanced” LFO mode that unlocks LFO speeds that enter the audible frequency range.
• An optional version of the noise generator that produces more typical white noise versus the darker original noise generator.
• While the SH-01A retains the original’s CV and Gate outputs for controlling external analog synths, Roland dropped the CV/gate inputs for the Boutique release. (But with USB and DIN MIDI you still have plenty of connectivity options.)
• Like all other Boutique models, the SH-01A has a 1/8" stereo audio output jack, but it just duplicates the mono signal on both channels.

Sound design on the SH-01A

The SH-101 was a budget synth, and its voice architecture (re-created on the SH-01A) reflects that:

• A single oscillator, with four sound sources whose levels are adjusted in the “Source Mixer” section: A modulatable pulse wave, a sawtooth wave, a multi-mode sub oscillator, and a noise generator.
• A single LFO, that can either be free-running or note-triggered, depending on the “GATE” switch. The LFO has 6 waveform shapes, and when switched into the optional “advanced” mode, the LFO can even serve as an additional sound source.
• A single envelope, for filter and amp. There is an Env amount control in the filter section, and the amp section can either be set to follow the envelope or work in gate mode (simple note-on, sustain, note-off behavior).
• There is a portamento control that offers always-on or legato modes, with adjustable travel time. (I used the legato mode for the “Slide Synth” part in this song.)
• The LFO is this synth’s primary modulation source, but you can also use the bender and mod strips for some tweaking.
• While the SH-01A thankfully includes a polyphonic mode (I wouldn’t have been able to produce this track as-is without it), there is no on-board chorus like the JUNOs had, so raw patches mostly sound pretty thin.
• The Unison mode is largely disappointing. It consumes all 4 voices to produce a thicker, detuned sound however there is no way to control the number of voices or the detuning amount. Also, since the synth still has effectively monophonic outputs, there is no stereo width to the unison effect. I used it on one of the patches in this track.

While the limitations can at times be frustrating, they can also push you to be thoughtful in your sound design. I always found myself looking for ways to make the patches more interesting by modulating things like the pulse width, filter cutoff, or pitch with the LFO, and making different sonic blends with the various oscillator sound sources.

Free patches!

I’ve decided to start offering free patches that I design for these tracks as a fun little bonus. I designed 15 patches for this project (14 of them are in the song; the first one was the INIT patch I created as a starting point).

If you have an SH-01A you can try them out yourself by downloading them here. The file Roland-SH-01A-Patches.zip contains a readme that describes the individual patches. The instructions for backing up and restoring (installing) patches on the SH-01A are described in the “Data Backup/Restore” section of the SH-01A Owners Manual.

Producing the track

I chose to cover AERwave’s track Burnout because it appears to entirely use subtractive synthesis, offering a limited sound palette that was more realistic to achive, given the SH-01A’s limitations. At the end of the day I got a good song out of this synth, but I did have to make some sound design compromises and I needed to add a lot of external effects (chorus, modulation, reverb, saturation) to make most of my patches sound like I wanted in the mix. The bass sounds didn’t require much help sounding good.

As a Roland completist I’m glad I had some time to work with this little synth, but it’s certainly not the first device I would choose when deciding to record a new song. I prefer something with more oscillators, envelopes, and LFOs at a minimum. But even with the SH-01A’s limitations I’d much rather work with this little synth than I would the original SH-101.

I produced the song in Cubase Pro 15. I recorded all audio through a Radial JD6 DI into a Focusrite Scarlett 18i20 4th-gen audio interface. In addition to Cubase’s built-in plugins I used the following third-party tools to produce this track:

• Baby Audio Super VHS (saturation)
• Blue Cat Audio PatchWork (parallel processing)
• Boz Digital Pan Knob 2 (panning automation)
• CableGuys ShaperBox 3 (modulation)
• FabFilter Pro-L 2 (track limiting)
• FabFilter Pro-Q 4 (EQ)
• Goodhertz Lohi 3 (filter)
• Kazrog KClip3 (clipping)
• LoudMax LoudMax (master limiting)
• Native Instruments Guitar Rig 6 (modulation effects)
• Native Instruments Solid Bus Comp (master compression)
• Nicky Romero Kickstart 2 (ducking)
• Sound Toys Decapitator (saturation)
• Sound Toys PhaseMistress (modulation)
• TBProAudio dpMeter5 (gain staging)
• TAL Software TAL-Chorus-LX (JUNO-like chorus)
• UVI Tape Chorus (tape emulation)
• Valhalla DSP ValhallaRoom (reverb)
• Valhalla DSP ValhallaVintageVerb (reverb)
• Waves Bass Rider Stereo (volume management)
• Waves H-Delay Stereo (delay)
• WavesFactory Trackspacer 2.5 (dynamic EQ)

Watch it on YouTube

Hear it on SoundCloud

It took me several days of trial and error, but I eventually figured out the entire process. Since there doesn’t appear to be any one place on the internet that documents the whole procedure (including several of the common technical problems that can occur), I wanted to document the method that ended up working for me, in hopes that other Blofeld Desktop and Keyboard owners have an easier time. Please note: I am a Windows user, and I do not know which of the tools I mention below are available for Mac OS. I am simply documenting my own journey.

Blofeld Wavetable Specifications

While the Blofeld ships with 68 permanent factory wavetables, there are 38 empty slots for user-provided wavetables. You access these slots by selecting wavetable 80–118 using the Shape control on either Oscillator 1 or 2. By default, these slots have the name, “User Wavetable.”

When creating your own wavetables, it’s important that they conform to the following specifications:

• Samples per wave (“Frame Size”): 128
• Waves (“Cycles”) per wavetable: 64
• Sample rate: 44.1kHz (44100 Hz)
• Bit depth: 8 bits or 16 bits (higher bit depth will result in higher quality wavetables)
• Audio channels: 1 (mono)

Configuring the Blofeld and collecting initial values

Before you create your own wavetables, you need to make sure your Blofeld is connected and configured properly. You will also need to gather a little bit of information in order to successfully create and transfer the wavetable data.

Use a DIN MIDI connection if you’re able

I’ve seen comments from a number of folks online who said they weren’t able to successfully transfer SysEx to their Blofelds when using the USB MIDI connection. The 5-pin MIDI In port appears to be more reliable for this purpose. I have successfully performed this action using both a MOTU MIDI Express 128 MIDI interface, and the MIDI OUT on my Focusrite Scarlett 18i20 audio interface.

Confirm Blofeld settings

Press SHIFT + Global on the Blofeld and use the selection encoder to navigate to Global MIDI screen 1/4. Set MIDI Channel to omni and write down the listed Device ID. The factory default for this is 0 (00h), but you need to know the current value when creating your SysEx files.

Confirm the desired wavetable slot

Press the Play button on the Blofeld to exit the global settings, and then select Oscillator 1 or Oscillator 2 (with the Utility button) and use the Shape encoder to locate the first available “User Wavetable” slot. If you device doesn’t have any user wavetables installed yet, your first available slot will be Wavetable 80. Write down the slot number.

Creating the wavetables

While there are several free tools for this sort of thing, I read good reviews of the OSS Wavetable Creator, so I purchased a copy of that.

In Wavetable Creator, I selected the correct Frame Size, Cycles, Sample Rate, and Bit Depth for a Blofeld wavetable, and also selected my audio interface so I could audition the wavetable as I designed it. In the example below, I have a wavetable that mixes various pulse and saw waveforms with a clipping effect applied.

When I clicked the File button, the tool created a new file called “PulSaw_128_64_44100_16bit.wav” in my Documents\Ocean Swift Wavetable Creator\Wavetables directory (the program’s default storage location).

Converting the WAV file to SysEx format

The Blofeld can only receive wavetables in SysEx format, not the WAV files exported by Ocean Swift Wavetable Creator. For this I used the extremely odd (but extremely powerful) free tool, Audio-Term by Mathias Gurk.

From what I can tell, the original download site for Audio-Term is no longer online, but I found a legitimate downloadable version at Wusik.

After downloading the file WIN_AudioTerm.zip, I extracted the contents into a directory named WIN_AudioTerm. Inside that directory was an executable named WIN_AudioTerm.exe. I launched it and accepted the default settings, resulting in a directory structure like this:

I copied my PulSaw_128_64_44100_16bit.wav file from Wavetable Creator into the “data\01 Samples (WAV)” directory created by the WIN_AudioTerm installer. The contents of that directory then looked like this:

Next, I launched Audio-Term.exe from the WIN_AudioTerm directory and clicked the KTERM<> button at the bottom-left of the Audio-Term window. On the next page I pressed the DISK<> button at the bottom of the screen, and I was greeted with the following page that automatically displayed the contents of that “data\01 Samples (WAV)” directory from before. Note that Audio-Term uses old-fashioned “8.3” filenames for files. The one in the example named “PULSAW~1.WAV” is the one I copied there.

Note: In Audio-Term, if you ever find yourself looking at the wrong directory contents, press the F_TYPE<> button at the bottom of the screen and then press the button corresponding to the type of files you wish to browse on the File Operations screen that appears. For example, to view my WAV samples, I would click the “->01” button below.

On the SAMPLES <WAV> screen I selected my sample and clicked LOAD<>.

When Audio-Term asked if I wanted to treat the file as a scannable wavetable, I clicked YES. On the following screen I clicked APPLY, and was greeted with this 3D view of my wavetable. (Epic!)

In order to export this as a SysEx file, I clicked DISK<> at the bottom of the screen, followed by F_TYPE<> on the next screen. From the list of file types on the next screen that appeared, I clicked the ->05 button, for “KTERM — WAVETABLES <BLOFELD, SYX>” This shows the contents of the “WIN_AudioTerm\data\05 KTERM — Wavetables (Blofeld, SYX)” directory, where the SysEx (SYX) file will be saved:

Click SAVE<> on this page to reach a screen where you will enter the filename of the soon-to-be-exported SysEx file. The filename can only be 8 characters long. I like to include the number of the target wavetable in the filename. In this example, I’ve entered the name “PULSAW82” to indicate that it will go into Wavetable slot 82 on my Blofeld:

After you press ENTER on the filename screen, you’ll be prompted for your Blofeld’s device ID (described in the “Configuring the Blofeld and collecting initial values” section earlier in this post). If you need to change the device ID to something other than the default of 0, press the “<-|” button (or use your BACKSPACE key) to clear the field and then use your keyboard to enter the new device ID.

After you press ENTER you’ll now be prompted to specify the wavetable slot number (also described earlier in this post). In my example, I already had custom wavetables in slots 80 and 81, so I entered 82. As with the previous screen, you can use BACKSPACE or the <-| button to clear out the field if you need to enter a new value.

Once you press ENTER you should now be greeted with a screen showing the name of your newly-created SysEx (SYX) file. (It will have arrows pointing to it on this screen.)

At this point you can quit Audio-Term.

Uploading the SysEx file to the Blofeld

It is this part of the process that took me the longest to figure out. I wasn’t able to get it working using the USB MIDI connection (I only had success with the Blofeld’s DIN MIDI IN port). Also, one tool that I had seen numerous people recommend (MIDI-OX) failed every single attempt I made with a “No Buffers Available” error. I was never able to get past that error despite following suggestions from several forum posts I found regarding the problem.

I ended up using the free tool, Bome Send SX. Here’s what I did:

After starting up Send SX, I opened the MIDI Out menu and selected the MIDI port that connects to my Blofeld. (In my case that’s Port 3 on my MOTU MIDI Express 128.)

Next, I went to Options > Settings and selected “Very Slow” for the Sending MIDI speed.

Next I went to File > Open and navigated to the “WIN_AudioTerm\data\05 KTERM — Wavetables (Blofeld, SYX)” directory where my SysEx file was saved, and I selected my newly created file, PULSAW82.SYX.

Next, I clicked Send (F4) to upload the file. Now when this works properly, Send SX will say “Sending line…” at the bottom of the window and the red MIDI indicator on your Blofeld will flash. When the upload gets near completion, your Blofeld will say, “receiving SysEx” on the LCD display for a couple of seconds. (This took me a while to figure out… at first I thought Send SX wasn’t working, because I didn’t see the “receiving SysEx” message on the Blofeld while the file was uploading- and since the message only appears for a couple of seconds, I kept missing it. But this is how the process goes when executed correctly.)

To test out the upload, enter Edit mode on your Blofeld and select either Oscillator 1 or Oscillator 2 with the Utility button and then use the Shape encoder to locate the wavetable slot you selected. Instead of “User Wavetable” it should now say, the name of the file you saved in Audio-Term.

Random notes and observations

• Before learning how to use Audio-Term I tried to use the free Waldorf Blofeld Wavetable Creator by Kotró László Lehel, but it would not accept any of the WAV files I generated with OSS Wavetable Creator.
• Before I figured out that Send SX was actually working, I looked for a way to convert SysEx/SYX files to MIDI/MID files, because I knew that I could play SysEx MIDI files into my Blofeld from my DAW (Cubase). (This is how I have installed firmware updates in the past.) So I discovered the donationware tool called Sysex to MIDI (SMF) (also known as “SyxToMidi”) by developer Royce Craven. I was able to use his tool to convert my Audio-Term SYX files into MID files that I could then play in Cubase (at 60 BPM) while routing the output to my Blofeld. Similarly to when doing Send SX transfers, the Blofeld only said “receiving SysEx” near the end of the transfer, which is what caused me to go back and try Send SX again.
• There were a few blog posts I wanted to call out for having useful information, starting with “Homegrown Pleasures” by Michael Rother at Broken Gadget.
• This post by Mark Mosher is where I discovered I didn’t need License SL to upload my own wavetables.
• And this post at Kimurataro had some important details regarding the format of the Blofeld’s wavetables.
• I got the idea for this article while writing another article about a track I produced on the Blofeld Desktop. Check it out if you’d like to hear a track fully produced using the Blofeld for all tonal parts.

History and context

In 1975 a musical mad genius named Wolfgang Palm founded the firm Palm Products GmbH (better known as PPG) to manufacture synthesizers of his own design. While the company produced some memorable products (most notably the PPG Wave), they ultimately ran into money troubles and had to shutter in 1987.

The following year, a former distributor of PPG products formed a new company named Waldorf to produce more synthesizers based on Wolfgang Palm’s inventions. This company produced the wavetable-based Microwave I (which you might recall from my discussion of the Waldorf Protein), the virtual analog (VA) Waldorf Q, and several other synths, however this company, too, ran out of money and closed its doors in 2004.

Two years later, a new team of business-minded enthusiasts formed Waldorf Music GmbH, with the continued charter of building synthesizers based on Wolfgang Palm’s chips. In 2007, the new company released its first hardware synthesizer to the market: the Waldorf Blofeld Desktop, the subject of today’s article. This small yet sturdy little synthesizer proved so successful that it is still in production today- almost two decades later! (It’s the synth equivalent of the Alesis SR-16!)

Features and specifications

The Blofeld paid tribute to earlier Palm-inspired synthesizers by including the wavetable synthesis (and even the actual wavetables) of the earlier Waldorf and PPG synths, as well as the VA synthesis of the Waldorf Q in a sleek and solid little package, with a feature set that would set the standard for things to come.

The Blofeld can operate in “Play Sound Mode” where you edit and play a single patch or in “Multi Mode” where it operates as a 16-part multitimbral device. The synth has a maximum of 25 voices, which are allocated automatically in multi mode.

The voice architecture is sophisticated:

• 3 oscillators, per voice. OSC1 and 2 can play either VA waveforms (Sawtooth, Pulse, Triangle, etc) or wavetables, and OSC3 is VA-only. OSC3 can optionally perform sync modulation on OSC2.
• 2 multimode filters (HP, LP, BP, comb, etc.) which can be arranged in serial or parallel fashion.
• 1 noise generator with multiple parameters.
• 4 envelopes (filter, amp, and 2 free) that can operate in multiple modes (ADSR, ADS1DS2R, One Shot, etc.)
• 3 LFOs with multiple wave shapes and parameters
• 16-slot mod matrix with all the typical modulation options, plus “Control W, X, Y, Z” modulators which can be mapped to MIDI CCs, and 4 “modifier units” which can modulate by performing mathematical operations on parameters.
• 2 effects slots for a variety of on-board DSP effects (note that the filters each have an optional drive/distortion effect as well, so you can potentially have 4 different effects on a single patch).
• There is a unisono mode that allows up to 6 detuned voices (which cuts into polyphony) for powerful supersaw patches.

The Blofeld also includes 1024 memory slots for patches, already populated with 1024 presets, made slightly easier to navigate with a built-in category filter. There is also an onboard arpeggiator (which I admit I have never tinkered with, so I don’t know its capabilities).

For an additional cost, Blofeld Desktop users can unlock 60MB of onboard flash memory for storing their own samples on the device. There are also keyboard and plugin versions of the Blofeld, which do not require an additional purchase for loading samples.

Both the keyboard and desktop versions of the Blofeld have been released with white and black finishes. And the plugin version can be used as a software programmer for the hardware Blofelds.

All hardware models feature USB ports for MIDI, as well as a 5-pin DIN MIDI in port. (The keyboard models add a MIDI out port and a sustain pedal input.)

Build and physical interface

Both the desktop and keyboard Blofelds are substantial devices. They feature full-metal enclosures, stainless steel encoders, and satisfyingly clicky buttons and power switches.

There are four types of knobs on the device:

• A selector dial for navigating through presets and menu screens.
• A fixed-run System Volume knob (all other knobs are 360-degree encoders).
• Two knobs below the display which correspond to the parameters displayed at the bottom of the screen at any time.
• Four knobs along the bottom right of the interface, which change values according to the matrix painted above them and the currently-lit LED on the left side of the matrix.

There are also five buttons, three of which select which set of parameters the four matrix knobs are to control.

Programming

Designing your own sounds on the hardware is both more involved than the sleek interface might have you think and less difficult than many online comments about the Blofeld might have you believe. The most fundamental sound design controls are clearly laid out in the LED/knob matrix, while more esoteric or optional features must be accessed via the selector knob and modified with the knobs below the LCD display.

I programmed 15 new patches for this project (and probably created about 15 more just during my learning process), all entirely from hardware. Once I knew where to find all the menu options, I found the programming experience pretty intuitive and seamless.

The Blofeld is a very good example of your primary limitation being your own imagination. They let you modulate just about everything, and with the (optional on Desktop) sample-loading capabilities, you can literally make any sound with this thing.

Bring your own wavetables

Something I don’t think I’ve seen discussed very much when the Blofeld’s mentioned is the fact that you can upload your own wavetables to the device, for use in your own patches. (Note: You do not need the sample license to do this on Blofeld Desktop; all forms of the Blofeld have this capability automatically.) There are 68 factory wavetables from prior Waldorf and PPG synths, but there are 38 empty slots (numbered 80–118) available, and they remain in non-volatile storage.

I very much wanted to exploit this feature while I was producing this track, but I ran into a couple of technical hurdles in the process that I could not surpass before I was finished with the song. I’m happy to report that I have since figured out how to do it, and I have already published a second article on how to do this, as it really adds a fun and cool dimension to this synth. You could, for example, capture waveforms of other classic synths or build your own wavetables for cool sweeping modulations not possible with the factory library.

About the track

I chose to cover Baton’s Starlight for this project because the original song has a more diverse palette of sounds than usual for the genre. Most of the synthwave tracks I like almost exclusively use subtractive synthesis, and I wanted something that would showcase sounds that the Blofeld is capable of beyond basic subtractive. (The song also totally slaps.)

I used the Blofeld’s Chorus and Drive effects a lot, with occasional use of the unisono feature. While the Blofeld’s Reverb isn’t the worst I’ve heard, I prefer something with more parameters, so I hardly used the on-board Reverb at all. The patches I designed use a combination of VA waveforms and factory wavetables. I did wavetable-position modulation wherever I could, as that is part of this synth’s sonic signature. I tried to include some oscillator sync action in there, too, but I wasn’t able to come up with an application for that feature that worked in this track.

Every patch that ended up in the final production is exactly what I wanted it to be. (Which is not always the case for me in these Single-Synth tracks; sometimes the given synth just cannot produce the specific sounds I hear in my head, due to its synthesis style or lack of parameters.) The Blofeld is a powerful synth, executed well.

I produced the track in Cubase Pro 15. I also used the following 3rd-party plugins for effects and automation:
- Baby Audio Super VHS (saturation and drift)
- Blue Cat Audio PatchWork (parallel processing)
- Boz Digital Mongoose 2 (stereo width management)
- FabFilter Pro-L 2 (track limiting)
- FabFilter Pro-Q 4 (EQ)
- Goodhertz Lohi 3 (filter)
- Kazrog KClip3 (clipping)
- LoudMax LoudMax (master limiting)
- Native Instruments Solid Bus Comp (master compression)
- Nicky Romero Kickstart 2 (ducking)
- Sound Toys Decapitator (saturation)
- Sound Toys PhaseMistress (modulation)
- TBProAudio dpMeter5 (gain staging)
- Valhalla DSP ValhallaRoom (reverb)
- Valhalla DSP ValhallaVintageVerb (reverb)
- Waves Bass Rider Stereo (volume management)
- Waves H-Delay Stereo (delay)
- WavesFactory Trackspacer 2.5 (dynamic EQ)
- Yamaha OmniVocal Beta (vocal chops)

See it on Youtube

Hear it on SoundCloud

The multitimbral power duo

In 1988, Roland released a pair of 1U multitimbral rack units that share many visual and technical similarities, but work quite differently in practice.

D-110 Multi Timbral Sound Module

The D-110 is a fully-programmable synthesizer with 8 separate instrumental parts and a dedicated rhythm section. It combines short PCM samples and traditional synthesizer waveforms to create complex sounds with up to four oscillators (“partials”). Roland released a separate hardware programmer to help users design their own patches.

U-110 PCM Sound Module

The U-110 has 6 multitimbral parts for playing the 99 built-in presets (“tones”). This was Roland’s first pure sample-based rompler; users cannot create their own sounds or significantly modify the pre-existing tones. The selection of sounds available here is quite limited compared to the D-110’s factory presets and deep programmability, and the sounds are primarily limited to traditional orchestral or band instruments. The few sounds available, however, are more realistic than most of the equivalent D-110 tones, since they are mostly based on longer and higher-quality samples than what the D-110 offers. The U-110 also includes a multi-sampled drum kit roughly presented in the General MIDI drum layout, although this kit only features a subset of the drum sounds available on the D-110.

The idea behind the pairing is simple: Use the U-110 for your bread-and-butter instrument sounds where fidelity is key, and the D-110 for everything else.

Both units also provide multiple audio outputs for live multitimbral playback and offer a variety of ways to distribute sounds to them. I have rarely had a personal use for so many outputs, though, since multitimbrality always comes at the cost of polyphony.

How it works: U-110 Internals

Under the hood, the U-110 is actually quite similar to the D-110. Both machines combine PCM samples and traditional synthesizer waveforms to create their tones. The primary difference is that the U-110 prevents users from directly manipulating those underlying building blocks to make their own sounds. The individual PCMs available on the U-110 are also mostly longer and higher quality than the D-110’s, however they are of far lower quality than the samples used in modern romplers or virtual instruments, simply due to the technical and storage limitations of consumer electronics at the time of production.

The sound architecture of the U-110 comprises three layers (patch, part, and tone). This diagram presents the high-level view:

Patches

In U-110 terms, a “patch” stores a complete configuration for the device. It indicates how the instrument distributes polyphonic voices across the different parts, which physical audio outputs to use, and the configuration of the device’s two onboard effects (chorus and tremolo).

One important aspect of a patch is the “output mode.” There are 50 of these output modes, which indicate how many individual parts are available in the patch, how many of the U-110’s 31 voices are allocated to each part, whether the parts use the unit’s on-board effects, and depending on the mode, whether to output the part in stereo or monophonic sound. I find it odd that the U-110 allocates voices statically via the output modes when Roland knew how to allocate voices dynamically as needed- after all, the D-110 can do it!

There are 64 memory slots for patches. There are no dedicated user patch slots, so you must overwrite factory memory slots when creating your own.

Parts

A single patch can define up to 6 “parts” which determine which sound to use, along with part-specific settings such as tuning, part volume, and LFO/vibrato.

In most output modes, each part in a given patch is monophonic, however there are a number of special output modes that combine parts 1 and 2 to either create a single stereo part, or a single monophonic part with a large number of available voices.

Tones

There are 99 presets, or “tones” to choose from. The tones fall into 5 different categories, and the types of tones used for a given patch determine the unit’s overall polyphony (number of simultaneous voices/notes), out of a maximum of 31 voices.

• Single: This is a single sound (PCM or simple waveform) that uses one voice.
• Velocity Switch (“V-SW”): These tones play one of two mutually exclusive sounds based on note velocity. For example, you might have a V-SW piano tone that plays a “hard touch” sound for higher velocities and “soft touch” sound for lower ones. These tones only use one voice at a time.
• Dual: These tones play two sounds mixed together, using two voices.
• Detune: These tones play two instances of the same sound together, slightly detuned from eachother to make a fatter sound. These use two voices.
• Velocity Mix (“V-MIX”): These mix two different sound depending on the current note value rather than velocity. These use two voices.

Some further details about the U-110’s sample content:

• The U-110 contains 2 megabytes of 12-bit sample content (with alleged sample rates of 15kHz to 30kHz), most of it traditional instrument samples, with the rest being synth and harmonic static waveforms.
• The U-110 has four slots on the front panel for inserting “SN-U110” PCM expansion cards, which each include up to 512 kilobytes of additional sounds, meaning you can potentially double your U-110’s available sample capacity by inserting 4 expansion cards at once.
• I have read that many of the samples the U-110 uses were sourced from the libraries of Roland’s S-50 line of samplers (S-50, S-550, and S-330). While you can’t record your own samples with the U-110, you also don’t have to wait for samples to load off floppy disks like you do with those other instruments.
• Many of the tones use the same underlying samples, but combine or mix them in different ways. For example, there are 10 different acoustic piano tones, but there aren’t 10 different sets of acoustic piano PCMs. Many of them are reused to different effect.

Onboard effects, modulation, and sound design

There are two built-in DSP effects, and they are only available on the L/R stereo outputs, and only in specific output modes. The Chorus effect sounds good at lower “depth” levels- I used it quite a bit for the track I produced with the U-110. I could not find a musical use for the Tremolo effect, though. In my experience, most tones that could benefit from tremolo (like the flute) already had that effect built into the sound.

On the part level you can create a vibrato effect by using the LFO controls to modulate the part’s pitch. You can also make minor adjustments, such as the attack and release times, velocity sensitivity, and tuning.

The U-110 in the studio

While the form factor and limited physical controls make using the U-110 an extensive menu-diving experience, the menus are logically organized and I eventually became quite adept at navigating them when editing and creating my own patches for this track. (It still would have been nice to have an “alpha dial” or basic 360-degree encoder, especially for entering patch names.)

In the track I produced with my U-110 (a cover of the Lofi Hip-hop track, “Seeking Change” by Yasumu), I recorded each part separately. Depending on the type of instrument I was recording, I used one of two output modes:

• Output mode 21: This combines the resources of parts 1 and 2 into a stereo pair for a richer sounding tone. It also assigns all 31 voices to this single part. This mode also enables the onboard DSP effects.
• Output mode 22: Like #21, this mode assigns all potential voices to one combined part, but as a monophonic tone, rather than a stereo one. DSP effects are disabled in this mode.

Since I like to mix all my drums separately (and the way the U-110’s drum section is implemented prevents assigning individual drums to the separate audio outputs), I recorded all of the individual drum sounds from the U-110 with SampleRobot and built my own virtual drum machine with them in Cubase. I did end up using two external samples for this track: A reverse cymbal and a tambourine, since none of the U-110’s built-in drums were adequate for those purposes.

While the U-110 sounded good for its time, the factory presets just aren’t up to modern standards, quality-wise. Rather than embrace the dated cheesiness of the U-110 (like I did with the track I produced on the Roland D-110), I chose the Lofi genre to showcase this synth since the deep filtering and saturation of the musical style obscure some of the instrument’s sonic weaknesses. (If you’d like to hear the U-110’s sounds in their raw, unvarnished form check out this good video from the SynthMania guy.)

To add further ambiance, I mixed in some vinyl pops and scratches along with an environmental recording I made on a rainy day many years ago.

Produced in Cubase Pro 15 with the following third-party plugins for effects and automation:
- Waves Bass Rider Stereo (volume management)
- Blue Cat Audio PatchWork (parallel processing)
- TBProAudio dpMeter5 (gain staging)
- Goodhertz Lohi 3 (filter)
- Waves H-Delay Stereo (delay)
- Kazrog KClip3 (clipping)
- Nicky Romero Kickstart 2 (ducking)
- LoudMax LoudMax (master limiting)
- Boz Digital Mongoose 2 (stereo width management)
- FabFilter Pro-L 2 (track limiting)
- FabFilter Pro-Q 4 (EQ)
- XLN Audio RC-20 Retro Color (vintage processing)
- Native Instruments Solid Bus Comp (master compression)
- Baby Audio Super VHS (saturation)
- WavesFactory Trackspacer 2.5 (dynamic EQ)
- Kazrog True Iron (saturation)
- Valhalla DSP ValhallaRoom (reverb)
- Valhalla DSP ValhallaVintageVerb (reverb)

Watch it on YouTube

Hear it on SoundCloud

About the MKS-80

Originally released in 1984, the MKS-80 Super Jupiter is a bit of an oddity. To my knowledge it is the only rack synth Roland produced that was fully analog. (Its oscillators have a habit of falling out of tune, necessitating the anachronistically-named “Auto Tune” button to bring them back in line.) Also, unlike many of Roland’s rack-based synths, the MKS-80 was not intended to be the rack version of an existing keyboard-based synth- it has a unique combination of features, circuitry, and parameters not fully shared with any of Roland’s other products.

There are two distinct “flavors” of MKS-80: Early models are known as “Revision 4” (“Rev. 4” for short), and they contain the same oscillator and filter chips as the Roland Jupiter-6 (which is why some folks mistake the MKS-80 for a rack-mounted version of that synth). But models produced later in the unit’s production run are “Revision 5,” which use different chips for the oscillators, filters, and amplifier (VCA) than the Rev. 4 models.

The MKS-80 is 8-voice polyphonic, with 2 oscillators per voice. Like the much later Roland D-50, the MKS-80 can operate as a 2-part multitimbral unit, or you can combine the two parts into upper and lower “layers” that play two tones together (for half the polyphony).

While it is technically possible to program the MKS-80 from the front panel, this is an extremely onerous process. Roland released a hardware programmer for this synth, the MPG-80. These devices are quite rare, and can sometimes sell for more than the synthesizer itself on the used market. Luckily, several third party manufacturers have released less expensive alternatives in the years since.

Production notes

While I had done a few recordings in prior years where I used only a single hardware synth for most or all of the sounds, this was my first serious attempt at producing a track in the synthwave style, so I chose to cover a song that I thought wouldn’t be too challenging for a beginner. I felt Xtract’s song “Audiotool Day 2016” was a good candidate for its minimal arrangement and lightweight production, largely propelled by its anthemic lead melody and chord progression. I also felt it would be a good showcase for the MKS-80’s lush pads and deep basses.

I used a KiwiTechnics Patch Editor to program new patches on my Rev. 4 MKS-80. I have programmed a number of synths with this device before, and the MKS-80 was the most complicated one, mostly because of its 2-layer support. (There are twice as many controls to navigate through as a result.)

Most synthesizers in my collection are equipped with some manner of unbalanced (“T/S”) style audio outputs, which I generally run through a DI box and then into my audio interface. The MKS-80 additionally offers a pair of balanced XLR outputs, which I ran directly into my audio interface. I had my MKS-80 configured in “mix” mode where I combined the upper and lower layers to create 4-oscillator voices.

When I first worked on this track, I had just gotten Cubase Pro 14, which included a new Drum Machine module which featured drum synthesis support. I created the drum sounds for the track with Cubase’s drum synth features just as a way to learn the new tools. I have replaced all the drum synth sounds with samples in my remixed version.

I also designed a couple of risers in Steinberg’s Retrologue plugin because at the time it didn’t occur to me that I could use the MKS-80’s white noise waveform for that (which I think is only available on VCO 2). I didn’t re-record any audio for this remix, but I always use my hardware synths to make risers in my current Single-Synth tracks whenever possible.

I remixed the track in Cubase Pro 15 and used the following 3rd-party plugins for effects and automation:

• BABY Audio: Super VHS
• Boz Digital: Mongoose 2
• FabFilter: Pro-L 2 and Pro-Q 4
• Goodhertz: Ghz Lohi 3
• Kazrog: KClip 3
• LoudMax: LoudMax
• Native Instruments: Solid Bus Comp
• Nicky Romero: Kickstart 2
• Plugin Alliance: Black Box Analog Design HG-2
• Softube: Saturation Knob
• TBProAudio: dpMeter5
• UVI: Tape Flanger
• Valhalla DSP: ValhallaVintageVerb
• Waves: Bass Rider, Doubler2, H-Delay, and MetaFilter
• WavesFactory: Trackspacer 2.5

Watch it on YouTube

Hear it on SoundCloud

A little history

The D-50 shared the majority of its “Linear Arithmetic” architecture with the other “D-series” instruments Roland released between 1987 and 1989, including the consumer-focused MT-32. The D-50 has higher-quality DACs than the others, as well as some useful features such as pulse width modulation as well as more and better on-board effects. (The later D-70 is more similar to the JV and XV instruments that would come in subsequent years.)

Released thirty years after the originals in 2017, the D-05 is perhaps the most complete and accurate reproduction of a classic synth in the entire Boutique line, partly because the D-50 was Roland’s first all-digital synthesizer, meaning Roland didn’t have to model any emulations of analog circuitry. The D-05 also doesn’t suffer any of the limitations that several of the other Boutique synths do, but includes some of the perks, such as a sequencer and arpeggiator. Also, in addition to including all of the original factory D-50 patches, the D-05 includes all content from the four original D-50 expansion cards, plus an additional bank of all-new sounds.

The D-05 in practice

The D-05 has all of the D-50’s original PCMs (samples), supports the name max number of voices, and includes the original set of factory presets, plus the content from all 4 official expansion cards, and a new preset bank made just for the D-05. It also supports the full MIDI specification of the originals, meaning it works with hardware and software patch programmers made for those synths.

I strongly encourage using an external programmer of some kind, because the D-50 voice structure is one of the most complex that Roland ever designed. For example, a single Juno-60 patch comprises about 33 parameters, while a D-50 patch can consist of between 110 and 315 parameters, depending on the number of “partials” it includes. I designed the patches for this song using the Dtronics DT-1000 programmer.

Like the original PG-1000, the DT-1000 is capable of loading patch data from the D-05 and listing the current values of the named parameters on display. For this to work, you must route a MIDI cable from the DT-1000’s MIDI OUT to the D-05’s MIDI IN, and another cable from the D-05’s MIDI OUT to the programmer’s PARAMETER IN.

Dtronics also released a compact programmer for these synths called the DT-01. It uses knob-style encoders instead of faders, but lacks a display and sends MIDI CC data out only; it does not load patches into local memory like the other programmers can.

The D-50/D-05 voice architecture is multi-tiered: A single patch can comprise up to two “tones” and each tone can contain up to two “partials.” Partials, which are basically individual oscillators, have one of two types: Synthesizer or PCM. Synthesizer partials use typical subtractive synthesis, and offer two waveform types (Square or Sawtooth). PCM partials use sample-based synthesis to play back sounds that fall into a variety of categories.

There are 100 PCM sounds to choose from, and different kinds serve different purposes. There are short “attack” sounds that represent the transients of an instrument- for example, the pluck of a guitar string or the mallet strike of a vibraphone. There are also single waveforms that capture the harmonic qualities of instruments, such as organs, pianos, or something evoking a human voice. There are “spectrum” PCMs which can sound drastically different from one note to the next, perhaps useful for cinematic sound design. And there are a number of loops that combine a number of the other PCMs in rhythmic patterns that I have never found a musical application for. (Oh, there’s also a single “Noise” PCM, which is a standard white noise sound.)

A disappointing limitation of the PCM partials is that they do not pass through the filter; only synthesizer partials can do that. So, for example, when I designed the riser for my D-05 demo track, I had to rely on external filters to sculpt the sound.

Roland went hog-wild designing the D-50, fully exploring the benefits of a purely digital architecture. The synth has robust key-follow support, “bias” ranges that add the ability to modulate over a specific portion of the keyboard, and the ability to select direct or inverse relationships between LFOs and envelopes.

Speaking of envelopes- while many synths (even today) use 4-parameter ADSR envelopes, the D-50 has 12-parameter envelopes with 5 distinct stages for both the filter and amp sections.

While the D-50 has a better reputation than the other D-series instruments due to the richer sounds you can get out of it (thanks to the chorus effect and better digital-to-analog converters), a lot of the classic sounds it’s known for are very “of its time.” Patches with the attack and looped PCMs can sound particularly cheesy to modern ears. Out of the eight original patches I designed for this track, only one of them uses an attack PCM. The others either use synthesizer partials or the harmonic waveform PCMs.

The D-05 in summary

Some final thoughts on the D-05:

• This is the most complete and successful Boutique instrument that Roland produced, in terms of faithfulness to the original product.
• It has a powerful subtractive synth engine, but you need a hardware or software programmer to really unlock this thing’s potential.
• I wish that you could pass PCM partials through the filter, and I wish the unit gave you more control over panning. (The only panning capability I know of are the fixed-position output modes that send the lower and upper tones to separate channels.)
• I have seen several people complain online that you can’t change the startup patch to something other than the dreadful “Neo Horizon” that loads up by default- but you can! The D-05 defaults to whatever patch is saved in the first sequence in the device’s memory bank: Patch 1–1 (also called “Neo Horizon”). Just edit the “Pattern Patch” parameter of the sequence and save it to memory. I changed mine to my custom “Init” patch that just loads up a single sawtooth partial.

About the track

I produced the track in Cubase Pro 15. Wherever possible I tried to use the D-05’s own effects (primarily Chorus and Reverb). I also used the following 3rd-party plugins for effects and automation:

• Baby Audio: Super VHS
• Boz Digital Labs: Mongoose 2
• FabFilter: Pro-Q 4, Pro-L 2
• FX23: PsyScope Pro
• Goodhertz, Inc: Lohi 3
• Kazrog: KClip 3
• Loudmax: LoudMax (free plugin)
• Native Instruments: Solid Bus Comp
• Nicky Romero: Kickstart 2
• Valhalla DSP: ValhallaRoom, ValhallaVintageVerb, ValhallaDelay
• Vengeance Sound: VPS Trance Gate
• Waves: H-Delay, Bass Rider
• Wavesfactory: TrackSpacer 2.5

Watch on YouTube

Listen on SoundCloud

The Roland MKS-7 Super Quartet is a 4-part multitimbral synth released in 1985. Three of the parts are based on the Juno 106 sound engine, and the fourth part is for drums (it has a selection of samples from the TR-707 drum machine).

I recorded all instrument tracks from the unit’s “Chord” part, which is 4-note polyphonic. (The “Melody” part is 2-note polyphonic, and the “Bass” part is monophonic.). Each part has a dedicated mono output, but I recorded from the stereo “mixer output.”

The MKS-7 does not have any controls for editing its sounds, but you can use MIDI SysEx to program the device in real time. I used a KiwiTechnics Patch Editor in Juno 106 mode to do all my patch programming.

There are 16 audio tracks in this project. Some of the patches are slightly tweaked presets, and the rest are brand new. Because the MKS-7 does not have any user preset storage, I was not able to save any of the patches after recording.

When I first worked on this track I thought my MKS-7 had a defect because I wasn’t able to get the noise generator working for the riser. (I used Steinberg Retrologue for that instead.) It turns out that the MKS-7 does indeed have a noise generator, but it’s only available on the “Melody” part, not Bass or Chord.

I have also since learned that you can put the MKS-7 into a special “Whole Mode” that disables the Melody part and grants the Chord part full 6-voice polyphony! (See page 19 of the Owner’s Manual.)

The drum sounds in this track are samples from my MKS-7, with some tweaks: I pitched the Ride sample up a bit and shortened its envelope, and pitched the Kick down a couple semitones. I also used a TR-727 Cabasa sample because the MKS-7 doesn’t have any shaker sounds.

Despite being fairly limited, the MKS-7 sounds great and was easy to program with the help of the Patch Editor. (I just wish it was possible to save user patches on the device.)

I used the following 3rd-party plugins for effects and automation:

• BABY Audio: Super VHS
• Blue Cat: Gain 3
• Boz Digital: Mongoose 2
• FabFilter: Saturn 2 and Pro-Q 4
• Goodhertz, Inc.: LoHi 3
• Kazrog: KClip 3
• LoudMax: LoudMax
• Native Instruments: Transient Master
• Nicky Romero: Kickstart 2
• Phil Speiser: THE_SPICE
• Slate Digital: Fresh Air
• TAL Software: TAL-Chorus-LX, TAL-Filter 2
• TBProAudio: dpMeter
• Valhalla DSP: ValhallaRoom, ValhallaVintageVerb and ValhallaPlate
• Waves: Curves AQ, Doubler, H-Delay and MetaFlanger
• Wavesfactory: Trackspacer 2.5

See it on YouTube

Hear it on SoundCloud