FABFILTER VOLCANO 3
Danny inspired me to write a post on a few Fabfilter Volcano 3 mixing techniques, as this plug-in can be a really colourful filter and effective in mixing as well as creative duties.
HARMONIC SATURATION AND CLIPPING CHARACTERISTICS
The Volcano 3 user manual describes the filter drive algorithms as:
- Classic, the original filter style taken from our award-winning FabFilter One synthesizer
- Smooth, like the cream in your coffee
- Raw, lots of overdrive and exhibits a character of its own. Great for distortion guitar sounds
- Hard, moderately distorting filter, with a nice clean whistle
- Hollow, juicy moderate distortion with fairly much low-end self-oscillation
- Extreme, for more wild sonic ideas
- Gentle, a more smooth and clean general purpose style
- Tube, with a warmer sound and nice overdrive, great for synth sounds
- Metal, with a rough, sharper sound and distortion
- Easy Going, a softer version of the Tube filter
- Clean, linear behavior with no drive or clipping at all
I ran each filter drive algorithm through Plugin Doctor Harmonic Analysis and Oscilloscope modules to review harmonic saturation and clipping behaviours.
Classic: Harmonic Analysis
Classic: OscilloscopeAnalysis
Smooth: Harmonic Analysis
Smooth: Oscilloscope Analysis
Raw: Harmonic Analysis
Raw: Oscilloscope Analysis
Hard: Harmonic Analysis
Hard: Oscilloscope Analysis
Hollow: Harmonic Analysis
Hollow: Oscilloscope Analysis
Extreme Harmonic Analysis
Extreme Oscilloscope Analysis
Gentle: Harmonic Analysis
Gentle: Oscilloscope Analysis
Tube: Harmonic Analysis
Tube: Oscilloscope Analysis
Metal: Harmonic Analysis
Metal: Oscilloscope Analysis
Easy Going: Harmonic Analysis
Easy Going: Oscilloscope Analysis
Clean: Oscilloscope Analysis
All non-linear filter algorithms generate odd and even harmonic saturation, resulting in asymmetrical clipping. Asymmetrical clipping means that the positive and negative portions of the waveform are clipped differently; the more asymmetrical, the more pronounced the even-order harmonics.
The harder the transition to clipping (the sharper the transfer function), the greater the number of higher-ordered harmonics are produced, as can be seen with the Classic and Extreme filter drive algorithms.
Hard clipping behaviour generates higher-order harmonics, which can add edge and bite or a sharp attack quality to percussive material. However, excessive high-ordered harmonic content produces a harsh, unfocused sound that is rarely desirable.
Soft clipping behaviour generates low-ordered harmonics, for example, the second, third, fourth, and fifth, which are considered smooth, warm, fat, and full-sounding, as they are harmonically consonant. The second is an octave above the fundamental, the third an octave and a fifth, the fourth two octaves, and the fifth two octaves and a Major third.
MIXING TECHNIQUES
Now that we have explored the sonic characteristics of Volcano 3, let’s dive into a few mixing techniques.
COLOURFUL EQUALISER
Volcano 3 can be used as a colourful filter box. In a recent production for Toolroom Academy, I used Volcano 3 to cut the low- and high-frequency energy of a pad sound, coupled with a midrange bell boost using the Raw filter type. Compared to Pro Q3, this added a warm, rich, open sound to the pads.
Volcano 3: Hammerstein Module; 100 Hz Bell Filter; Clean
The above illustrates a 100 Hz Bell filter using the Clean filter type. As expected, there is no harmonic saturation.
Volcano 3: Hammerstein Module; 100 Hz Bell Filter; Tube
The above illustrates a 100 Hz bell filter using the Tube filter type with Drive at 0 dB. Because the Tube algorithm is non-linear, it generates harmonic saturation; in this case, low-ordered odd and even up to the fifth harmonic. The Hammerstein module demonstrates harmonic boosts that are relative to the harmonic i.e. second harmonic 200 Hz, third, 300 Hz etc. Similar harmonic curves are also found for the other non-linear filter types. While this behaviour is also present in non-linear EQs such as Manley Massive Passive and Pultec EQP-1A, for example, these equalisers do not allow for the harmonics to be driven, notably not per band or with different saturation characteristics as can be achieved in Volcano 3.
This could be an interesting technique for adding subtle saturation, weight, or body to low-frequency content like 808s, kicks, and bass in specific frequency regions. For example, a bell boost at the fundamental could add additional harmonic energy, an octave and an octave and a fifth above, which will help these low-frequency instruments cut through on band-limited playback systems like mobile phones and portable Bluetooth speakers.
Additionally, I could see this technique being effective for harsh synth sounds when rolling off high frequencies; for example, if the high-cut filter has effectively removed harshness but made the sound a little closed or dull, pushing the drive could add a little controlled air or openness back into the sound. Of course, this technique would also be effective with low-cut filters.
Additionally, the bands can be set up in either serial or parallel or a mixture of both, with parallel filters arguably having a smoother sound. Inter-channel pan offsets for each band can create highly mono-compatible pseudo-stereo width, and each band can have different filter types and drive amounts; there is lots of creative flexibility here.
If you are unsure what the Hammerstein module is showing; it is the behaviour of harmonics across the frequency spectrum. For example, a linear processor would only depict the fundamental, if a non-linear processor generated a second and third ordered harmonics these lines would also appear, if these harmonics were linear across the frequency spectrum they would be straight lines. However, say the second harmonic was more emphasised at 2 kHz the line would illustrate this.
HAAS EFFECT FOR WIDTH
Yes, we all love the Haas Effect, but why use Volcano 3 for this application?
First, we need to understand how humans localise sound.
Inter-aural differences in the level and arrival time of sounds are primary cues for localising sound in the horizontal plane (Azimuth).
There are four effects that occur when the inter-channel level or arrival time of a coherent sound source is modified:
If the relative level and arrival time of two simultaneous coherent sound sources satisfy certain conditions, a summing virtual source appears at a position where no actual source exists. If the conditions exceed that for summing localisation but remain within specific boundaries, a fused spatial event originates from the position of the leading sound source; directional information from the reflection is suppressed. This phenomenon is known as the precedence effect (Haas Effect). If the delay is increased, and the arrival time of the reflection exceeds the echo threshold, the fused spatial event separates into two localisable sound sources where the reflection is perceived as an echo of the first. If the delayed time is increased further, the reflection becomes independent of the primary sound, and two primary auditory events are perceived.
For this spatial width technique, we are interested in Summing Localisation and the Precedence Effect (Haas Effect).
Summing localisation is utilised in the stereophonic system. When the level and time of the left and right channels are identical, the signal is perceived as emanating from the centre of the two loudspeakers, i.e., the phantom centre.
If one of the signals has an intensity difference or delay, the auditory event will migrate from the midpoint to the loudspeaker that radiates the stronger or first-arriving signal.
An inter-channel intensity difference of approximately 15 dB is sufficient to position the source at either loudspeaker and an inter-channel time difference between 630 μs and 1 ms is sufficient to position the source at the loudspeaker reproducing the leading signal.
When the relative arrival time difference between two coherent sounds exceeds the upper limit of summing localisation (1 ms), a different spatial auditory event occurs called the precedence effect.
In this case, the auditory system perceives a sound as coming from the position of the leading sound when the arrival time difference falls within a specific boundary defined by a lower and an upper limit. With arrival time differences between 1 ms and 30 ms, the two sound sources are perceived as a fused spatial auditory event, which is localised at the lead loudspeaker with no apparent contribution from the lag loudspeaker.
Haas Effect
Haas Effect: Phase Response
5 Millisecond Comb Filter: Frequency Response
In the above example, I have delayed the right channel by 5 ms, which creates comb-filtering nulls (180-degree phase shift) that begin at 100 Hz; this is because the 100 Hz waveform period is 10 ms (360-degrees), and therefore its half cycle (180-degrees) is 5 mis. Additional nulls are at harmonic intervals of 300 Hz, 500 Hz, 700 Hz, and so on, with the centre frequency of peaks at 200 Hz, 400 Hz, 600 Hz, and so on.
It is worth taking note of the wavelength period time (or more accurately its half cycle time) of particular frequencies that null in mono; this can be really effective for creating optimum mono and stereo versions. For example, you could Haas Effect a synth sound so the comb-filtering nulls are in areas of the frequency range that would clash with the vocal on the mono sum. This can be tuned in mono, or dial in the frequency wavelength period where masking occurs in mono (I prefer the former method).
Haas Effect: Stereo Image
Haas Effect: Polar Sample
Haas Effect: Polar Level
The effect has created a wide and expansive sound from the mono source; the source still feels relatively centred because there is no inter-channel difference in level.
Haas Effect with Level and Frequency Attenuation
Haas Effect with Level and Frequency Attenuation: Stereo Image
Haas Effect with Level and Frequency Attenuation: Polar Sample
Haas Effect with Level and Frequency Attenuation: Polar Level
In the above example, the right channel level has been attenuated by 6 dB, and a high cut filter at 10 kHz has been applied; consequently, the source has radiated towards the left channel.
It is worth noting that comb-filtering has been significantly reduced, and the source is much more mono-compatible. As described earlier, this could be further tuned to attain the optimum sound in both mono and stereo by adjusting delay time, level, and filter frequency. Additionally, changing the filter drive type and the Drive amount of the right channel can also change the characteristics of the spatial effect.
The Haas Effect is usable on far too many sources to mention. Shorter delay times are recommended for percussive sounds and longer for more sustained times; typically times range from 1 ms to 40 ms depending on the source.
The Haas Effect is often better suited than, say, Ozone Imager or StageOne when you want an expansive or out-of-the-speakers sound. It is difficult for these plug-ins to achieve that sound, as they are either a volume control for the side signal or a delayed and inverted mono signal (comb-filtering in the side channel instead of the mid) to create stereo content, which cannot achieve that super wide sound of the Haas Effect.
Check out my Izotope Ozone Imager and Leapwing StageOne article here
L/R INTER-CHANNEL FILTER DIFFERENCE
Inter-channel Filter Difference
Inter-channel Filter Difference: Stereo Image
Inter-channel Filter Difference: Polar Sample
Inter-channel Filter Difference: Polar Level
Slight differences between the left and right channel filters’ centre/corner frequencies or slopes can create stereo width that is highly mono-compatible.
I have found this technique to be particularly effective for contrasting sounds; for example guitar. or synth tracks panned left and right; contrasting filters can produce an expansive sound without changing instrument positioning. For example, in the above example, the corner frequency for the low cut filter is at 100 Hz (75 Hz left and 150 Hz right) the opposite could be applied to the guitar or synth track on the other channel (left 150 Hz and Right 75 Hz). Of course, this technique does not work for highly coherent sound sources.
It is also a good technique when you wish to move particular frequency ranges out of the way of centre instruments, such as vocals, to reduce masking.
M/S FILTERING FOR WIDTH
M/S Filtering
Dry Stereo Sound: Stereo Image
Dry Stereo Sound: Polar Sample
Dry Stereo Sound: Polar Level
M/S Filtering:Stereo Image*
M/S Filtering: Polar Sample*
M/S Filtering: Polar Level*
M/S filtering is very good for changing the frequency balance between the mid and side signals of a stereo source; it can add subtle stereo width that is mono-compatible. This technique will not work on a mono source and will only have limited effectiveness on a stereo source with limited side signal; you cannot boost what is not there!
An interesting aspect of using Volcano 3 instead of, say, Pro Q3 for this technique is pushing the drive to add harmonics, which creates additional frequency content in the mid or side channel. While the saturation is not dependent on the filter’s centre or corner frequency, it is heavily dependent on filter settings.
PARALLEL BANDPASS FOR DENSITY
Using a driven parallel band-pass filter can add density and fullness to a particular frequency range. This technique is flexible, as you can experiment with different filter drive types, adjust level, pan, slope, and even delay.
Parallel Band Pass Filter
Parallel Band Pass Filter: Frequency Analysis
Parallel Band Pass Filter: Hammerstein Analysis
Parallel Band Pass Filter: Harmonic Analysis 100 Hz
Parallel Band Pass Filter: Harmonic Analysis 1 kHz
An interesting aspect of the above example is the harmonic contour. The first band is a bell filter at unity with the Clean filter type and generates no harmonics; consequently, there are no harmonics with the 100 Hz test tone. However, the 1 kHz test tone depicts harmonic saturation from the second filter, which is a band-pass filter in parallel with the Extreme filter type.
This technique is really good for adding density to sustained sounds like pads, synths, and vocals. It can also be used to add weight or body to kick and snare drums.
The filter drive algorithm and drive amount is important for achieving the desired sound.
Remember to turn band 1 up by 6 dB for unity gain; set band 1 to Clean mode if you want the original signal to be clean.
ALL-PASS FILTER WIDTH
This technique is good for adding a sense of depth and width to a sound without it being clearly localisable. It is relatively mono-compatible, depending on settings, of course.
All-Pass Filter
All-Pass Filter: Phase Response
All-Pass Filter: Polar Sample
All-Pass Filter: Polar Level
This technique is particularly effective for background sounds or background layers of stacked sounds. Frequency ranges can be pushed out of the way of primary sounds, adding a sense of depth and width without changing the positioning on the sound source.
TRANSIENT FOLLOWER FOR ATTACK
Transient Follower to Filter Cutoff
When working with stems or sample loops that need a little more attack or shape to the sound, you can use Volcano 3 with an envelope follower modulating the filter cutoff.
Set the filter cutoff to where you want modulation to start, choose filter type, drive, etc.; then add an envelope follower and assign it to the filter cutoff. Increase the modulation amount and time constants of the envelope follower until the desired sound is achieved.
An envelope follower traces the waveform, and its output is based on the dynamic contour of the input. However, an envelope generator could also be useful for this application where you wanted “static” envelope shaping on all hits.
The envelope follower or envelope generator could also be assigned to the Volume parameter for level-based envelope shaping.
I have found this technique particularly effective when reshaping a stack of synths, and for fitting percussive loops into a mix.
I hope you enjoy this post; please comment if you have any Volcano 3 techniques or any suggestions, I always love to chat about this stuff slight_smile: