Welcome to the Sound On Sound Recording and Mixing podcast channel where you’ll find shows packed with Hints & Tips about getting the most out of the recording, mixing and mastering process.
More information and content can be found at www.soundonsound.com/podcasts | Facebook, Twitter and Instagram - @soundonsoundmag | YouTube - https://www.youtube.com/user/soundonsoundvideo
Hello and welcome to the Sound On Sound podcast. In this episode I'll be talking about the history and the evolution of guitar synthesizers. The subject of guitar synths is a bit of an odd one as we can divide the sector into systems that actually work from a conventionally playable guitar and those devices that look like a guitar and use guitar fingering, but they're not actually guitars.
The forerunner to guitar synthesis was probably the Vox V251 organ guitar that dates back to the 1960s. It was based on a modified Vox Phantom guitar, but the electronic circuitry, based on a Vox Continental organ, relied on insulated frets, with six contacts built into each one, so that when a string was held down, it would complete a circuit and an organ note would sound. You only needed your left hand to play it. Guitarists didn't really like the one handed technique, and the build up of sweat and corrosion on the divided frets made it very unreliable, so it wasn't particularly successful.
In 1971, EMS released the Synthi Hi Fly, which was designed by David Cockerell, the man behind many of the current electro harmonics designs. Though marketed as a guitar synth, the Hi Fly was really an early multi effects unit that included ring modulation, fuzz, octave shifting, attack and decay controls, modulation and filtering. They're now very collectible but they're still not guitar synths.
Things went quiet for a few years and then, in 1977, ARP announced the Avatar. Essentially an Odyssey synth hooked into a hex pickup system that could be fixed to pretty much any guitar. I took one home to try, but couldn't get it to trigger or track reliably, so I took it back and tried instead the Roland GR500, which was released at almost the same time. The GR500 came with its own guitar, complete with built in hex pickup. Its synth section combined polyphonic sounds created by processing the guitar string waveforms. There was a sub octave section that could be applied to selected low strings and it had a monophonic analogue synth that followed the last note that you played. It worked well enough to convince me to buy one and I had some great results from it, but you had to play really carefully and the tracking delay was quite obvious. It also needed a general helping of chorus, reverb and delay to make it sound alive. One useful feature of the GR500 is that it included CV and gate outputs, so it could be used to control other analogue synths.
This next guitar synth, the GR300, was launched in 1980. Again, this came with its own guitar, but this time the synth engine was based on the string waveforms, augmented by some clever phase lock loop technology that allowed pitch shifting and pitch gliding. Essentially, the string waveform cycles were used to reset a ramp generator, waveform at the same pitch as the guitar string. The sounds were pretty much limited to pseudo string and brass types, but the GR300 was truly polyphonic and there were no apparent tracking or latency issues which made it a lot of fun to play. It followed the guitar input exactly, including string bends. It was also very simple to operate, so guitarists were not intimidated by the technology. Andy Summers, then in The Police, made good use of one, and yes, I bought one of them too. A simplified version called the GR100 was also available.
Here's an extract from one of my 4 track recordings made in the very early 1980s using the GR300.
And then along came MIDI. With MIDI came some wonderful sample based synths and so the race was on to build a guitar synth that could convert pitch to MIDI to drive the same kind of sound engine you might find in a keyboard. However, extracting pitch from a vibrating guitar string is no easy task, as the waveform is far from pure and also has a very noisy attack. Once the pitch has been extracted, it then has to be converted to MIDI, and that was a big ask for the technology of the time.
Initially, Roland again opted for a system with their own guitar. The GR700 was teamed with the rather oddly shaped GR707, a guitar that I recall christening the Dalek's Handbag, due to the long stabilising bar between the body and the headstock. I reviewed it back in 1984 and it produced a wide range of keyboard sounds and it could also output MIDI for controlling other synths and modules. The tracking could still be a bit hit and miss and there was some processing delay, but things were moving forward. This was to be Roland's last system that relied on a dedicated guitar.
Also in the 1980s, Casio came out with their own integrative MIDI guitar that was both affordable and surprisingly usable. It could be used to control any conventional MIDI sound module, but it also had on board sounds.
Following the GR700 system, Roland relied on third party guitars fitted with Roland GK Series Hicks pickups, or their bass guitar equivalents. The GK pickups could be fitted by the end user and were compatible with most guitars other than the Telecaster where the bridge plate prevented the pickup from being correctly positioned close to the bridge where it needed to be. Fitting the pickup correctly with the recommended string spacing was essential for reliable operation. A small control box attached to the pickup connected to the guitar synth module it was controlling via a 13 pin DIN cable, and that became a standard.
Fender and Roland also co-developed the Roland Ready Strat with a GK Hex pickup built in, yes I have one, and numerous other guitar companies, most noticeably Godin, produced their own instruments with a Roland compatible 13 pin output.
Here are a couple of brief sound examples from my old GR33.
Of course, Roland didn't have it all their own way with guitar synths. There were companies like Axon, Shadow and Fishman all coming up with their own solutions. But it's probably fair to say that Roland have put more effort into developing the guitar synth than anybody else. A succession of GR series guitar synths were produced for use with the GK and GK compatible pickup systems. We saw the GR1, the GR9, the GR30, the GR33 and the GR55, the latter also including some guitar modelling features and effects using COSM modelling.
Now there is some inevitable tracking delay with any pitch to MIDI approach as the system has to read at least half a cycle of the incoming audio to be able to extract the pitch. However, as the technology improved, the tracking delay was brought down to what most players consider to be an acceptable minimum.
Roland's GR Series instruments were all sample based with sound engines that worked like a simplified version of their keyboard counterparts. They gave guitar players a much wider palette of sounds to play with, but there was always a learning curve as guitarists learnt to play in the style of the instrument the synth was emulating. I mean, there were still players who would set up a slow attack string or brass sound, and then complain that the synth couldn't keep up with their shredding. One big difference between the guitar and the keyboard, from a playing perspective, is the way that chords are voiced. A guitar can spread the chord notes over a couple of octaves, whereas with a keyboard, each hand can only cover a fairly restricted range of notes.
While Roland went down the pitch to midi route, other companies came up with their own guitar synth systems, many being guitar shaped controllers that didn't actually behave like guitars. The weirdest was probably the Synthax with its angled neck, equally spaced frets, and separate picking and fretting strings. It was incredibly expensive and not very guitarist friendly. I thought it looked a little like a chainsaw.
Step is another brand that came and went, but at least their picking strings were in line with their fretting strings. Their DGX instrument had separate picking and playing strings as well as a rectangular body covered in buttons.
In the late 1980s, Yamaha came out with their own futuristic midi guitar, the G10, and that was strung with all the same gauge strings. These only had to be set to a comfortable tension as they produced no sound. The G10 used advanced optical and ultrasonic scanning to work out where the player's fingers were, and it certainly looked distinctive with its ultra slim body.
Zeta's Mirror 6 used a clever combination of wired frets to get around the latency, but then it switched to pitch tracking as soon as the string vibrations settled down. Sound On Sound's Dave Lockwood reviewed this back in 1991 and found its performance to be pretty impressive, but at a price that would put off most potential purchases. And talking of expensive, guitarist Pat Matheny was often seen playing what looks like a modified Roland guitar plugged into a Sinclavier system.
Another pitch to MIDI instrument worth looking at is the Xyvix Jamstick Studio, which has a hex pickup and DSP built into a very tidy guitar. There's a mini TRS jack for the midi output, and also a regular jack output for the guitar pickups, and so it can hook up into any midi system. Its onboard battery is charged via a USB C port, which can also carry MIDI. This was reviewed by Dave Lockwood of Sound On Sound in January of 2021 and he was extremely impressed by its tracking accuracy.
Surprisingly, the most successful of the not really a guitar controllers came in the form of the U Rock MIDI Guitar, where the pluck strings are separate metal bars and the fretboard, which has pseudo strings moulded onto it, behaves as a series of switches.
The Starlab Zitar, which works on a similar principle, has six narrow push switches between each fret position. You can't bend notes on these instruments in the same way as you can on a normal guitar. You have to use an electronic vibrato arm, but for playing keyboard style chords and melody lines, they actually work pretty well.
Today we have pedals that are not quite synths, but they change the sound of the guitar in radical ways to sound like one. For example, the Electra Harmonix 9 Series pedals can emulate polyphonic organs, synths, string machines and so on, by combining polyphonic pitch tracking with some rather clever wave reshaping, pitch shifting and filtering, all based on the incoming guitar waveform. These offer the player the simplicity of effects pedals, but they can produce remarkably convincing sounds. The advantage from the player's perspective is that they feel like you're playing through an effect and they'll work with pretty much any standard guitar.
Roland also developed a guitar synth that could be used with a standard guitar and this was the Boss SY300. The SY300 felt to me almost like a step back in time to the Roland GR300, where most of the sounds are created by reshaping the waveform coming from the guitar pickups. But in the case of the SY 00, the hex pickup is not required. The range of sounds is reminiscent of analogue synths and there's no MIDI output, but for those who want a guitar synth that plays with the ease of an effects pedal, it's a great option. It also contains some powerful effects, as well as the option of mixing guitar and synth sounds. The same technology was also shoehorned into the BOSS SY1 compact pedal, but without the effects.
The follow up to the SY300 was the BOSS SY1000. This effectively combines the principles of the SY300, but adds on a pitch tracking synth fed from a GKX pickup. Again, the sounds are mainly analogue, but the addition of an oscillator based synth provides it with a wider range of sounds. It can be used with a standard guitar for patches that don't use the oscillator based synth, but really to get the most out of it, you need a guitar with a GK Series pickup or compatible.
Now, not wishing to throw a spanner in the guitar synth works, but it is possible to create a very wide range of synth style sounds by combining effects pedals or plugins such as delay, envelope filters, envelope shapers, pitch shifting and so on. There's also the Ebow attachment, a type of electronic bow that can be used to create infinite sustain monophonic melody lines on a standard guitar, and this has been around since the 1980s.
Here's the E bow with a few plug in effects added.
Now, if you slow the attack of the guitar using a pedal such as the EHX Attack and Decay, or a plug in such as Eventide's Physion, the sound takes on a very different character, after which you can add octave up shimmer delay to create a very effective substitute for analogue strings. You can also use granular processing, pedals such as Hologram's Microcosm or Red Panda's Particle 2 and in the realm of plugins there's the wonderful Portal from US Company Output, or the EFX Fragments from Arturia. This kind of processor can really transform the sound of a guitar and further blurs the lines between guitars and synthesizers.
While leaving MIDI triggering behind limits the types of sounds that can be produced by guitar synths or stacked effects units, the playability of an approach that uses the string sound as the source waveform has a lot going for it. For example, there's no tracking to worry about, so you don't need to be so clinical with your playing. There's imperceptible latency and every nuance of your guitar playing translates as a change in sound, whether that's playing pinched harmonics or simply bending notes and adding finger vibrato.
These days, when I want to trigger MIDI sounds in the studio using a guitar, JamOrigin's MIDI guitar software actually does a pretty solid job and it works on a regular guitar, so I don't feel the need for a hardware guitar to MIDI system. Any unintentional MIDI notes are usually easily identified by their low velocities and short duration, making them easy to edit out.
As to the future, I think machine learning will be exploited further to make note identification more precise, but I think DAW software developers could also do more for the MIDI guitarist. For example, having dedicated MIDI guitar tracks that instantiate a synth voice for each string, as is conventionally done using a different MIDI channel for each string along with MIDI mode 4. It would also be useful to have a global instrument bend range setting that you can call up when you're using a MIDI guitar track, regardless of its settings when used on a keyboard track. It would also be possible to create an automatic filter for low velocity short duration rogue notes, accidental low velocity pull offs to open strings and other common MIDI guitar flotsam. Given that guitarists significantly outnumber keyboard players, I'm sure that would actually make good business sense, wouldn't it?
Thank you for listening, and be sure to check out the show notes page for this episode, where you'll find further information, along with web links and details of all the other episodes. And just before you go, let me point you to the soundonsound.com/podcasts website page, where you can explore what's playing on our other channels.