PLUGIN WEEK 2026: Can you can remember exactly where you were when you heard your favourite album (or a life-changing track) for the very first time? We are fortunate enough to have had many moments like that. But there was one particular album where this writer can remember calmly sitting down in front of a (now old-fashioned) hi-fi system, and having his mind thoroughly blown from the very first track onwards…
I think we can all agree that Radiohead are quite ‘the’ experimental band. Their exploration of the various extremes of music have taken them to the very fringes of what’s usually acceptable in the studio - including the use of some exceedingly cutting-edge production techniques. They’ve also incorporated a fair share of vintage instrumentation, such as the Ondes Martenot, as well as strings and well… you name it, Thom, Jonny, Ed and co have probably done it at some point.
Having experienced exceptional success with their more chart-friendly albums The Bends and OK Computer, the creativity dial was set to maximum for the electronic-leaning follow-up, Kid A in 2000.
The crystalline simplicity of the opening track, Everything in its Right Place (which we’ve analysed at length in an exhaustive song story here), was the first indication that we were in for something special…
The use of Prophet-5 synth as an accompaniment to Thom Yorke’s audio-scrubbed vocals (a relatively new technique which was made possible by Pro Tools-based production) set the tone for an album that would introduce millions to sounds and innovative attitudes previously the remit of a select, Warp Records-signed few.
By the time the third track on the album made it to the front of the queue - the dark, circular bass groove and frenzied, blaring horns of The National Anthem - our socks were thoroughly blown off.
One thing that we particularly remember being amazed by at the time was the exceptionally cool effect being applied to Thom's vocals.
“I think, [Thom] is someone who's very into the fine voice that he has, and at the same time he's always looking for new ways to try and express different personae,” said bassist Colin Greenwood back in 2001 in an interview with the BBC. “A way of doing that is by treating the voice”
Although we can’t be fully certain of the chain used to achieve the particular effect, we know for a fact that, at this time, Radiohead were adopters of a largely experimental package which can certainly achieve the same result.
In fact, this particular piece of software formed part of my own regular arsenal of effects prior to Kid A coming out, so I could conceivably stake the claim that I used it before Radiohead!
Just so that we’re completely clear about the glass-like effect that we’re talking about, we have to wait until around 01: 36 on the timestamp of the song, and the first vocal entry - after the main groove and overlaid Ondes Martenot have set the scene.
It is identifiable by the resonant sound that we hear after Thom delivers a line - it seems to leave a lasting impression in the audio, almost the aural equivalent of a bright light being seared into your retina.
It imprints once the vocal line hits a particular frequency. It's an effect that gives a metallic quality to the vocal, and can be achieved using a comb filter.
Casting my our minds back to the late 90s, and computer-based effects had really started to hit their stride, but pre-VST effects and plugins, an organisation called INA GRM (Groupe de Recherches Musicales) had long been established as a leading centre for research into the use of electronics in music, particularly experimental music.
First founded in 1951 by Pierre Schaeffer, it was known as a centre for the development of musique concrète; the groundbreaking technique of creating music from found sounds using analog tape (an early form of sampling, but using tape and razor blades) which gave inspiration to organisations like the BBC Radiophonic Workshop.
In 1990, INA GRM released GRM Tools; an entirely unique standalone suite of software and processing effects, running in a DAW-less capacity on an Apple Mac. It’s where we first heard the GRM Comb and its highly recognisable stylings.
“The first GRM Tools came out in the early 1990s but they didn’t just come out of nowhere,” INA GRM’s director François Bonnet told Ableton. “Actually, some of them were trying to go back to some early machines like the Phonogène and the Morphophone [early innovative instruments]. But mostly they were continuing the first real-time treatment system developed at GRM : SYTER (SYsteme TEmps Reel / Realtime System). The Syter algorithms were really inspiring to start a collection of sound treatments as plugins. So as plugins, they were really quickly designed as commercial software, but they come from a long period of prototypes."
Due to its standalone nature, it meant that many avant garde and experimental composers used GRM Tools with an accompanying recording device, such as a Digital Audio Tape machine (DAT to its friends), recording the experimental output generated by the GRM Tools suite.
With the traction of VST plugins, and the adaptation to other formats such as AAX and AU, GRM moved with the times, offering elements from the original Tools suite in individual plugin form.
We know that producer Nigel Godrich and Thom Yorke were keen users, and so it seems likely that it was the GRM Comb Filter plugin that was responsible for the effect we hear on The National Anthem.
The effect was described more broadly as a ‘ring modulator’ by guitarist Ed O’Brien in an interview with MC2. “We are delay and reverb snobs. I think the interesting thing is, we don't go for nice sounds. For example, we processed the vocals on The National Anthem with a ring modulator. We can get up our own asses a bit, but what we do recognise is, when it comes to sounds, I think we've got quite good taste. We're good at judging what is and what is not a good sound. For us, part of recording is finding out different methodologies and techniques to create new sounds - which is what you should be doing in the recording studio."
It's therefore necessary to have either this plugin (or something very similar) in order to recreate the same effect.
With perfect timing, INA GRM has just re-engineered all of the classic Tools from the original suite into a new standalone and plugin format, which it calls Atelier. Moreover, it is also now available for Mac and PC. Nice!
‘What is comb filtering anyway?’ I hear you ask…
The literal definition is as a form of distortion, which occurs when a single audio signal is combined with a delayed facsimile of itself.
However, in the effect domain, it is quite unlike conventional high or low pass filters. A comb filter offers a number of spikes, a little bit like heavily narrowed ‘Q’ points on an EQ curve.
Also much like an EQ, these spikes are exaggerated and excited by the presence of a matching audio frequency, hence it picks out particular notes and frequencies - which can be defined by the user - to create resonant tones and chordal effects.
In order to achieve a similar effect, you are going to need two things; firstly, some form of audio signal to excite the filter - which we are going to adopt as a vocal part - and some form of comb filter plugin.
You can download a demo version of GRM Atelier if you just want to dip your toe in the comb filtering pool, although the originators are now accompanied by a plethora of other brands that have picked up the comb filtering mantle.
Examples such as Polyverse Supermodal or Kilohearts Comb Filter (which is also free) are all ripe for exploration and can be used to follow this exercise.
Start with a simple key note and build the first frequency plot from here
Once you have your vocal part on a track within your DAW, load up your chosen comb filter plugin.
We need to think about which notes we want to excite from our vocal signal. As many plugins use frequency as their parameter of choice, it's a good idea to nudge the frequency plot up and down, until you find a satisfactory note to excite.
To give you a rough idea of value, the note A above middle C on a piano keyboard, has a frequency of 440 Hz. The note A, one octave higher, has a frequency of 880 Hz - so the rest is pure maths!
Some comb filter plugins also offer a pitch value, so you can simply use a note name instead of a frequency value to construct your desired chord.
In our example, we are using GRM Tools Atelier, but many parameters will be duplicated within other plugins. The Decay parameter adjusts the volume of the resonance, and will dictate the length of time it takes for the resonance to decay to nothing. Be careful, because higher values here can be extreme in volume.
The Damp control limits the number of overtones, so this should be adjusted, particularly if you only want a single note to be exaggerated.
Once you've added your first frequency point, it's time to think about adding a second, third or more, but once again, you need to consider which notes you want to bloom.
If your vocal line or song is in a major key, consider exaggerating the 3rd of the chord (in much the same way that Radiohead did). If you are in a minor key, try exaggerating the 5th of the chord, which will also work well in a major key too.
In our example, our song is in C Major, and we have chosen to place the note C on the top, with a G and D below, creating a Sus2 chord in an inverted state.
A note though, the notes in the chord won’t be constant, only appearing as the vocal line stays across the assigned notes in the plugin, exciting the frequency.
Add additional frequency points to build up your chord
As you add each frequency point, increasing its amplitude will bring out each note.
You may choose to temper and increase/decrease each frequency point, meaning that you can balance the chord to suit your purpose. In all cases though, it is important to listen, as a spirited vocal performance can have an extreme effect on an outputted signal.
Always work with a degree of caution - each time you add a frequency point, and increase its resonance, it will stick out like the proverbial sore thumb.
It's therefore important not to overcook frequency points, which is a sensible mantra for the effect itself.
There is nothing so old and obvious as an overused effect, so it might be worth holding this effect back for a point in the song where it really counts. After all, it took Radiohead over a minute and a half on The National Anthem before Thom Yorke’s comb-filtered vocal made an appearance.
When you get it right - as Radiohead absolutely did - it’s an effect that leaves a lasting impression on the ears of listeners.
