What Is Spectral Processing? Get To Know The Technology Behind The Next Generation Of Effects Plugins

Spectral effects are a relatively new field of technology and have exploded in popularity in recent years. Developers around the world are releasing plugins with a new buzzword: spectral processing. iZotope, Sonible, Mastering The Mix, and more have joined forces, but what exactly can they do?

First, note that spectral processing is a technology only possible in the digital domain. Spectral effects use Fast Fourier Transform (FFT) calculations to break down sound into a representation of the frequency spectrum. The spectrum is then broken down into many different bands and these bands are processed individually.

In many ways, spectral processing can be considered "hyper-multiband", using more bands to get the job done. But while a user can program settings for three or four bands in a multiband effects processor, multiband spectral processing must rely on rules that determine how each band processes sound. These rules can be programmed by the developer or can be controlled by user-defined parameters.

Spectral processing technology can be applied in many ways. Perhaps the most obvious is the RX iZotope, which can perform several functions using spectral processing technology. For example, RX can split a mix into individual instruments, remove background noise, and even perform spectral restoration on poor-sounding audio.

Elsewhere, plug-ins like Oeksound's Soothe2 have taken manufacturers' plug-in folders by storm, offering softening of loudness and resonance wherever they occur in the frequency spectrum. This type of "moving target" processing is ideal for spectral applications, acting as a dynamic balancer with many active bands waiting to enter.

The 6 most advanced plugins that surprised us in 2022

It can also be used to enhance more traditional sound effects such as compressors. Sonible's Smart:comp2, for example, uses spectral compression for sharper compression results. smart:comp2 is actually a multi-band compressor with hundreds of bands across the frequency range. This means the strong bass in the audio signal will not compress the high frequencies, which can help create full mixes and is ideal for transparency.

The story continues

Not only are third-party developers starting to produce exciting instruments, but DAW developers are also jumping in. Ableton introduces several new spectral effects processors in Live with Live 11: Spectral Resonator and Spectral Time add extra depth to traditional resonators and delay devices. They're great for weird and otherworldly sounding designs, especially if you're a big fan of slightly unnatural and distorted digital audio.

In this article we will look at some examples of spectral processing in the plugin world. We have seen many powerful and revolutionary applications in recent years, but the potential applications of spectral processing technology in sound effects are very exciting. Here we are, the rest is yet to come.

The first three spectral processors

Equalizer processing is divided into two types: one affects only the transition material in the audio signal, and the other only works on the existing tonal sound. With SplitEQ, there are tons of new ideas for experienced mixing engineers and creative applications and sound designers to try.

2. New sound saturation

In this plugin, clipping is subjected to spectral processing. By cutting multiple strips separately, Saturate can cut or boost sound while still having headroom. The result is clear scissors that can penetrate where it can melt.

3.GS Magic Field DSP

Want to change pitch and frequency with characteristics that match the frequency spectrum? MagicPitch is designed for this and gives you all kinds of boundary-pushing creative effects.

Like Serum, Little AlterBoy and VintageVerb before it, Soothe by Oeksound is one of those plugins that really captures the mood of music producers. Other developers have focused on emphasizing resonance and one of them is Reso from Mastering The Mix.

Reso intelligently analyzes your sound, identifies the resonant frequencies, creates a band EQ for each, and lets you adjust the damping of each resonant frequency to your liking. Or it can automatically suppress any resonant frequency.

Note that on our garage-style practice track, the drums are quite loud in the upper half when the snare hits. This is where Returns comes in handy. Load Reso into the thrombus, play the song and press the Compute Target button.

A row of red + icons will appear, representing the resonance frequency. By pressing + you can move it up/down to change the threshold at which the node is activated and left/right to change its frequency. Move the mouse wheel to adjust Q.

Clicking Target Engage automatically attenuates any detected resonant frequencies with the EQ band setting that Reso deems best. The resonance attenuation is displayed as a red graph at the top of the spectrum analyzer.

Resonant frequencies are often difficult to hear, and it may be difficult for you to hear the difference between refined and raw drums. By clicking on the "delta" icon, we can hear only the omitted resonant frequency.

Spectral Suite Beetroot

After some controversy online, Bitwig's new Spectral Suite is now included as part of the core DAW package for Bitwig Studio 4.4. This spectral processor is a very clever signal splitter that allows you to split the signal and process its different qualities in different ways.

The simplest tool is probably the Transition Partition. In this way, it is possible to separate the signal into transient and non-transient matter, rebalance and process each with a different sequence of effects.

The most "creepy" effect is probably Harmonic Split, which detects and isolates roots, oddities and even harmonics, allowing you to isolate, mute and rebalance any of the three, as well as process them separately. Loud Split is something to behold, splitting the signal to its extreme on the spectral level.

Split audio into harmonics and transitions in Bitwig Studio 4.4

Let's start with temporal division. We can balance the transient/non-transient parts of the device or isolate one of them. Or enter an analysis that leads to the separation of the two elements. In this case, the slider to the right of the waveform gives a heavy transition result.

We apply flanger only to the tonal and supporting elements, without painting the transitions. For example, we can only add a distortion processor to the transitions, making them sound louder without affecting the rest of the signal.

Let's move on to Harmonic Split, which can split a signal into harmonic or fundamental groups, and nothing else. This basic separation allows us to enhance the lowest parts of our signal, or apply effects to only that part, or only the harmonics.

Here's an interesting application: We only added a Pitch Shifter to the root, lowered it by 12 semitones and set the Mix to about 60%, and we just added a sub (or new bottom root) to our instrument .

New audio possibilities with Bitwig Loud Split

Saying goodbye is a revelation. With spectral processing, it doesn't just "lock" the signal on volume; You isolate a different volume at each point of the frequency spectrum. Let's play with horizontal lines to understand our attitude…

…and then we just add reverb to the higher parts so they sound louder but don't blur the project. Since it only affects frequencies above and beyond our threshold, the effect is fun to play with.

What about the fading effect on the intermediate part of a signal? Or take a slower beat and add modulation to soften it? We can even apply saturation and gain to improve compression or process vinyl recording artifacts.

Spectral compression is the term Sonible uses to introduce the second version of smart:comp. Building on the recent popularity of spectral effects, smart:comp 2 takes this spectral processing technology and applies it to the realm of compression.

So what does it look like in the case of spectral compression? Well, you can think of smart:comp 2 as a multiband compressor with 2000 bands of spectrum. This means that smart:comp 2 can compress certain areas of the frequency spectrum with high precision, producing a balanced yet transparent compressed sound.

This is particularly useful for items such as drums, submixes and master channels, where frequency information appears across the spectrum at various points along the track. With a traditional single-band compressor, it can be difficult to effectively compress a song because it will focus on different areas of the frequency spectrum at different times.

This spectral compression is enabled by default when working with smart:comp2, but can be disabled for more traditional compression approaches. Sonible also lets you use a spectral algorithm to control the pitch of your voice, making it darker or lighter.

To explore the spectral compression possibilities of smart:comp 2, we will use it on the main bus of this dance line. Spectral compression is suitable for use on the main bus because the full track contains information for the entire frequency spectrum.

We start with the famous Sonible artificial intelligence. Select Mix | Email profile from the drop-down menu at the top of the user interface, click "Register" and click "Play". smart:comp 2 listens to the song and applies the compression settings it thinks are best.

The Spectral Com (pressure) knob can be set from 0 to 150, where 0 indicates single-band compression and 150 uses all possible bands. By default, smart:comp 2 is set to 100. Increase it to 150 to hear what maximum spectral compression sounds like.

smart:comp 2 delivers crystal clear compressed sound thanks to spectral compression. It's hard to hear without a comparison point, so here's what compression sounds like when the Spectral Compression switch is set to 0.

The differences are subtle, but single-band compressed sound is more limited in tonality. Spectral compression technology can also colorize audio using a color scale. We set the brightness to maximum to highlight these keys and strings.

A strong feature is that you can adjust the frequency range over which spectral compression is applied using two knobs. By default they are set to 20Hz and 20kHz, but by dragging them you can change the compression spectral range.

The simple MSspectralDelay mode contains 20 powerful spectral delay algorithms. Choose from a light, medium, or brutal algorithm category and play around with different preset parameters to apply all kinds of delay effects, from subtle repeats to full-voice transformations.

Go to the edit page to access the advanced MSspectralDelay control. From here you can adjust the usual delay options, including dry/wet amount, feedback level, delay mode and delay time. Let's set up a live table tennis delay with 8 synchronized tones with a medium mix and high feedback value.

Below the delay control is a spectral graph that allows you to scale parameters based on the frequency content of the original sound. In this case, we reduced the bass response to remove fading. Double-click the graph to create a new node.

The MSspectralDelay analysis and measurement functions allow you to clearly see the audio content of your audio source, giving you a visual indication of where there is too much frequency content. You can turn the analyzer and sonogram on and off by clicking on them.

You can shape the overall frequency content of a delayed signal using the EQ tab, which is useful for cleaning up the bottom end of a delayed signal or smoothing out high frequencies when the top end is being delayed with high feedback values. This is where the high point comes out.

The Spectral Transformation tab allows you to change the response frequency to another. Activate the Transform function in the spectral window, then increase the Transform value. Costs can rise quickly, but this can be controlled by reducing the number of conversions.

Then average the delay response using the conversion rate graph. The x-axis represents the input level and the y-axis represents the output level. Experiment with this graph to create spectral compression effects or input to control frequency response delay conversion.

You can also shift the delay time based on the frequency content. Click the Panorama tab and click Enable. Dragging the black nodes vertically controls the depth of the pan, and dragging the white nodes changes the wavelength. Drag the black button border horizontally to soften the effect.

The full set of modulation options includes a number of unique modulation sources. In the utility window, click on the modulation site. Then select the modulation target from the list or use the learn function. There is an envelope follower that reduces the amount of feedback.

Relaxed spectral plugins, including Smooth Operator and Reso (mentioned earlier), can make all the difference when you're looking for a smooth, professional mix. By smoothing out the frequency spectrum and smoothing out resonances, you get a balanced track, group or entire mix that sounds as experienced sound engineers and mixing engineers have observed.

But then again, there's something about driving this processor too hard that gives you a dull, lifeless sound that rings and sounds like you've never been near a room full of air. Of course, it is the "imperfections" that make music believable.

So if you're looking for resonance dampers, smoother trackers, tone balancers, or some of the other newer processors like the ones we covered in this article, remember to use them with care and a light hand.

The smooth operator is one of the most popular AI de-resonance removal tools in recent times. The Smooth Operator is $100 cheaper than Soothe 2 and much more affordable, but how does it work? Let's take a look at the action on this track.

The main arpeggiator of the synth is a bit loud at higher frequencies due to the use of a bitcrusher. Fixed by loading the Smooth Operator into the channel. You will notice that the Smooth Operator suppresses the higher frequency sounds.

However, Smooth Operator currently focuses on the mids and lows. For higher frequencies we can control the threshold and tonal balance of suppression using the manipulator on the center line of the UI.

We have now increased the overall smooth operator threshold to reduce layer averaging. We use high-frequency simultaneous keystrokes to improve typing here. This gives us a less brittle arpeggiated sound.

How technology turns people into products