So far, I have one finished track (an ambient experimental piece), one half-finished track (a lo-fi downtempo track a la Texel), and two short demos (a chiptune and a solo piano piece). I’ll try to finish the track I have in progress tomorrow, so with any luck by this time next week I’ll have three or four finished tracks, and four or five demos ready to be expanded upon.

Some random things I’ve learned so far:

• Plugin soft-synths are super, super handy when you’re in a hurry — just drop them in a MIDI track, load up a preset, and you’re good to go, without worrying about routing signals or configuring external software or hardware.
• Speaking of soft-synths, the TAL-NoiseMaker native VST synth is my new go-to synth. It’s a standard analog-style synth, but it sounds great and has a straightforward UI and a solid feature set.
• Ardour 3 is still a bit crashy while working with MIDI, but it’s made some nice improvements recently, like being able to double-click to enter or leave note edit mode, and the addition of a drop-down list of synth plugins in the “new track” dialog, so you can start composing more quickly. I could switch back to Qtractor, but even with the crashes I think I’m more productive in Ardour, just because I’m more familiar with it.
• Sound design is fun! It’s hard not to have a good time when I fire up the Blofeld and start twiddling knobs. I should do it more often!
• In fact, I should do this whole music thing more often. I might not come up with something interesting every time I sit in front of the keyboard, but definitely won’t come up with anything if I don’t try.

I talked a little about the production in an earlier post, but I have included some further details after the jump.

This is one of my shortest tracks, and that was a deliberate departure for me — in the past I feel like I’ve often stretched things out to get to a certain length, and let them get a bit too repetitive in the process, so I deliberately avoided that here. In hindsight, there’s a lot going on in there, so perhaps it could’ve done with being a bit longer, but it definitely shouldn’t be boring at the very least!

In terms of sounds, the pads and bass are from the Blofeld, but everything else is software — Calf Monosynth for the sixteenth-note bleeps in the background, Linuxsampler for the Rhodes, and Hydrogen for the drums. The Rhodes is the excellent jRhodes3 soundfont, which I converted to SFZ format using sfZed, and the kick and snare come from the Microhammer Gnomehammer collection of wintry sounds.

The vocal sounds were a bit of fun. I recorded myself singing a note in to my mic, chopped out a small section that had a nice, stable pitch, and then time-stretched that to produce a longer vocal sound that sounds a little strange and ethereal. Each of the little vocal flourishes was created from that by pitch-shifting it to create the different pitches, and then performing a bunch of micro-edits for the glitchy, stuttery bits.

Effects-wise, I’ve already discussed Loomer Cumulus — it’s responsible for the messed-up drum break in the middle, and a few drum hits in the latter half of the track. I’ve also used the Decimator plugin a bit: the Cumulus drums, vocal sounds, and sixteenth-note bleeps all have automated Decimator plugins. The warbly, lo-fi delay on the Rhodes is courtesy of my Behringer VM1 analog delay pedal.

The new track is a bit of a downbeat, ambient-ish thing, with some lo-fi sounds mixed in with some glitchy elements. I definitely put Ardour’s MIDI features to the test: there are MIDI tracks running out to my Blofeld and to Hydrogen, along with LV2 synths (Calf Monosynth and Linuxsampler), along with automation of CC parameters on the Blofeld and automation of plugin paramaters on Calf Monosynth. I’ve done quite a bit of effects automation as well, particularly with the bitcrushing Decimator plugin.

There’s even a VST plugin in there now; I had been beta-testing Loomer Cumulus, using it as a standalone synth, but with Ardour’s new VST support I now have it running within Ardour directly. Cumulus is somewhere between a synth and an effect: it lets you load a sample, and then trigger its playback using granular synthesis with varying paramaters, altering the starting point, pitch, and playback rate, among other things. You can define up to eight sets of those parameters, and then trigger those via MIDI keys. It can turn all sorts of sounds in to eerie textures, but it can just as easily take a drum loop and turn it in to a wonderfully glitchy mess, which is exactly what I used it for.

I’m pretty sure the track is done, but I don’t want to release it just yet. I plan to sit on it for a few days at least, while I read more of my copy of Mixing Secrets for the Small Studio, but I like the idea of putting together at least an EP with a couple of other tracks and releasing them all at once. That might not be practical if it takes me four months to finish each track, though, so I may post the individual tracks here when they’re ready, and then do an official Bandcamp release once they’re all done.

This was whipped up on Sunday night mostly in Qtractor, with Hydrogen for the drums, and my Blofeld for all the other sounds. I’d normally record everything in to Ardour and mix it there, but I stayed in Qtractor for this one, and it did a fine job; I had no trouble replicating my usual trick of running the drums on to separate tracks so that I can apply individual effects to each, for instance. The result is a bit trite, but it’s fun, it loops pretty smoothly, and I think it suits the game well.

The open kit synths: MeeBlip and Shruthi-1

The MeeBlip and Shruthi-1 both answer the same question — just how much of a synth can you build on top of a low-cost Arduino-style platform? The MeeBlip squeezes an analogue-style two-oscillator monophonic synth on to it’s tiny 8-bit AVR CPU, and while the sounds may be digital, the interface is very old-school. It’s eight knobs and sixteen switches give you direct control over the sound, and it’s a good thing, too, since the MeeBlip lacks patch storage.

The Shruthi-1, with super-cool tranparent case. Image © Mutable Instruments

The more expensive Shruthi-1 has less direct control, but it does have patch storage, and quite a few more sound generation options. Like the MeeBlip, it’s a monosynth that runs virtual oscillators on a bargain CPU, but it augments those digital sounds with an analogue low-pass filter. Not only does that give you some genuine analogue feel, and the ability to filter sound from external sources, but it also leaves more CPU power for the virtual oscillators. That lets it cover more ground: while the MeeBlip has just saw and pulse waves, the Shruthi-1 adds super-saws, sweepable wavetables, vocal formants, and CZ-style phase distortion.

What both synths share is their design philosophy: both are completely open, with the hardware designs and the firmware source code available under open licences. If you have the means, you can source the parts, grab the firmware, and build your own synth from scratch; for the rest of us, both synths are available in kits with all of the bits you need. As more of these synths get in to the hands of developers, it’s going to be very interesting to see what new features find their way in to their firmwares.

The mass-produced dark horse: the Monotron

Korg surprised the hell out of everyone when it announced the Monotron — the last thing anyone expected from Korg was a 100% analogue synth for US$60. The Monotron is incredibly simple: there’s a single sawtooth oscillator, a low-pass filter (based on the filter from the venerable MS-10), an LFO for modulation, and a ribbon controller for input. There are no envelopes, and definitely no MIDI, but it’s still capable of some crazy sounds, and like the Shruthi-1, its filter can be used to process external audio.

The Monotron is capable of far more impressive things in the right hands, though, because its hardware is remarkably hackable. Korg recently released the synth’s circuit diagram, but it hardly needed to — the Monotron’s design is so clean and simple that it had been hacked by the community even before its official release. Still, Korg should be applauded for embracing, rather than discouraging, the modding community that that Monotron has inspired.

Many of the hacks involve adding external control systems — YouTube is awash with videos showing Monotrons under CV/Gate control from PCs using MIDI-to-CV/Gate converters, or under more direct control from custom Arduino-based boards. More intrepid hackers have gone beyond pitch and gate controls, adding external control of the filter, extra oscillators, and distortion circuits, among other things.

Affordable synths for the rest of us

While these new synths are bringing hardware down to software prices, the thought of assembling a synth from parts is, admittedly, a bit intimidating. Thankfully, mainstream instrument-makers are also delivering some remarkably affordable synths, even if they’re not quite as bargain-priced as the DIY options.

The Alesis Micron and Korg microKORG both give you small keyboards and virtual-analogue sound on a budget, the DSI Mopho squeezes an analogue monosynth in to a tiny box, and the DSI Evolver delivers unique hybrid analogue/digital crazyness. Even Moog is getting in to the game with the Slim Phatty — US$849 isn’t exactly cheap, but for an honest-to-goodness, 100% analogue Moog synth, it’s a bargain.

The Slim Phatty brings the classic Moog sound to a new low price

I still think that the Waldorf Blofeld — my synth of choice — is a great option, too, especially at its current second-hand prices. It covers a great range of sounds, from bread-and-butter analogue sounds through to distorted wavetable madness, and it has enough polyphony in multi-mode to run an entire track’s worth of sounds at once.

Linux soft-synths are certainly improving, but if you’re serious about making electronic music under Linux, you owe it to yourself to check out some of these affordable hardware options.

Here, then, is a “performance” of my track tiny droplets — the various MIDI loops used are all pre-sequenced, but I’m triggering them all in realtime using my QWERTY keyboard. In this case, seq24 is driving Hydrogen and my Blofeld, and I’m using Ardour as a live mixer to process and mix the audio from those synths in to a stereo stream.

UPDATE: If you’d prefer to download the video rather than streaming it on YouTube, I’ve uploaded a WebM version of it. WebM is still quite new, but current versions of VLC and MPlayer support it.

On a brief side note, I have to give a shout-out to my good friend AutoStatic for describing his new video capture process using Xephyr and FFmpeg — I used it here, and the results look great. The audio was captured with JACK TimeMachine, and in another first for me, I edited it all together using the brilliant Kdenlive.

This is another Seq24/Hydrogen/Ardour recording, with Blofeld synths, though I also created my own drum sounds (mostly on the Blofeld again) for this one. I also used PHASEX as the synth for the lead arpeggio — it’s a simple patch, but I really liked how it sounded, so it stayed in the final version.

EDIT: Turns out that the download links were broken! I’ve fixed them now, so if you had trouble downloading, please try again now.

That’s not to say that I started with a totally hardware-centric workflow, though. My work has always revolved around Ardour, not just for recording, but also for effects, and I’ve used Hydrogen for drums on just about everything I’ve recorded so far. However, on my first track (atlantis), the instrumentation was all Blofeld, and it was mostly played live, with just a few bits of piecemeal sequencing.

Just another day at the office: Qtractor, Ardour and Hydrogen running in unison

By comparison, my latest track (frozen summer) was made entirely in the box, though it’s perhaps not a fair comparison point since it was 100% sample-based and I don’t have a hardware sampler. A better comparison is my cover of Enjoy the Silence, where everything but the (Hydrogen) drums was sequenced from start to finish in Qtractor, using software synths as well as software effects.

Some of that was FluidSynth playing Soundfonts for the less “synthy” sounds (guitar, horns, etc.), but I also made extensive use of WhySynth, an analog-style synth. I even worked with WhySynth the way I’d work with my Blofeld, crafting my own patches instead of using presets for most sounds. Why did I ditch my hardware and embrace softsynths?

The box is convenient

The answer, perhaps unsurprisingly, was convenience. With a separate softsynth on each track, it’s very easy to create custom patches on each one, and then add in custom effects chains that process the results on each channel, too. Using multi-mode on the Blofeld I can run multiple patches at once, and edit them individually, but the results come out a single stereo output in to my PC’s single stereo input, so I can’t add realtime effects to individual instruments unless I use the Blofeld’s (very limited) internal effects. Recall is also an issue — with softsynths, when you load your session, it’s exactly as you left it, but with a hardware synth, you usually have to set it back up yourself.

The other obviously convenience factor is portability. I spend a bit of time on buses, and working with softsynths means that I can do everything on the go. My laptop has a 2.4Ghz Core 2 Duo, and it’s easily handled everything I’ve thrown at it so far.

WhySynth doesn't inspire, but it's a good bread-and-butter synth

That’s not quite the end of the story though, because I did end up using the Blofeld on that track. After sequencing it all on my laptop, I moved it to my desktop to add some polish, and while all the FluidSynth sounds stayed in the final product, I replaced some of the WhySynth sounds. It’s a perfectly serviceable analog-style synth but, try as I might, I couldn’t get enough “oomph” in the bass part, or the right filter squelch in the “bieuuw” effect sound that comes in around the second chorus. With the Blofeld, I was able to nail both sounds very quickly.

There’s also a tactile element that softsynths are typically unable to capture. A good synth is an instrument in itself, with an interface that beckons the user to create new sounds and interact with them in realtime. A softsynth might be capable of making the same sounds, but I’m yet to find one that inspires me like a good hardware synth can.

Finding a balance

Ultimately, a lot of the choices about when to use hardware or software come down to compromise. Using softsynths plugins inside Qtractor is very convenient, but the available synths have limited possibilities. Using hardware gives me better sounds that I can program more quickly, but it ties me to my home studio, and limits my effects options while sequencing.

There’s even a middle-ground between the two extremes under Linux — JACK softsynths like Specimen and PHASEX. I can run these on my laptop, and depending on how they’re set up I run separate realtime effects racks for each under Ardour, but they require setup each time you start them, so you lose the convenience of total recall that you get with softsynth plugins.

I think the answer to the question of when to use hardware or software for a job is “it depends” — each track is going to have its own ideal line in the sand, and that’ll vary from person to person, too. Sometimes I’m going to start with a killer sound or riff on the Blofeld and build a track around that, so it’s going to make sense to use hardware all the way through. Other times, I’ll be able to get away with softsynth plugins, even if they’re just guide sounds that I end up replacing later, or it could just be that a softsynth is the best tool I have for a particular job.

Perhaps the best answer is to use whatever gets me the best results with minimal fuss and maximum enjoyment. I don’t think softsynths will ever replace my hardware, but that adding them to my toolkit has definitely made me more productive.

mp3 / vorbis / flac: 4 minutes 26 seconds

mp3 | vorbis | 4:10