Saturday, August 15, 2026
Friday, August 14, 2026
Cubase 14: Play Probability & Velocity Variance
By John Walden
The new Velocity Variance and Play Probability lanes let you easily experiment with performance variations to keep your MIDI loops from feeling static.
Cubase’s Play Probability and Velocity Variance tools can bring your MIDI patterns to life.
Unless you’re intentionally making music with a robotic and repetitive feel, it’s always useful to spice up your MIDI parts to make them feel more ‘human’. Abandoning quantise and playing complete parts in live from start to finish are obviously options, but not everyone has those performance chops, and if you happen to be working with short MIDI loops for things like drum, bass or piano parts, you might want to look to alternative strategies in any case. Cubase has plenty of options on this front, and if you have either the Pro or Artist editions of Cubase 14 there are two new options you can exploit: Play Probability and Velocity Variance. Below, I’ll consider a couple of examples, one a humble MIDI drum loop and the other a simple MIDI bass groove, to see whether we can give these MIDI performances a little extra life. I’ve created a few audio examples to accompany what’s written below, and you can find these on the SOS website at https://sosm.ag/cubase-0825.
Let The Lane Take The Strain
The new Play Probability and Velocity Variance lanes are available in both the Drum Editor and Key Editor windows. In both windows, you can toggle the display of these lanes via the pop‑open menu from the ‘+’ button located towards the bottom right. As for the note velocity lane, the Key Editor shows data for all notes in the MIDI clip, while in the Drum Editor, if you select a specific drum element (kick, snare, hi‑hat...) then you can see data just for that element.
In the Probability Lane, the probability that any given note will trigger on playback is determined by the value (height) of a vertical bar for that note, and this can be set between 100 (full height) and 0 percent. Each time the clip is played, these probabilities are applied. So, for any note with a probability less than 100, you’re essentially choosing a degree of randomisation. In the Velocity Variance lane, the values are initially set to zero (centred upon the line within the lane) and you can apply either positive or negative bias away from this on a per‑note basis. On playback, this then applies a degree of randomisation to the note’s velocity up to the maximum positive or negative bias you’ve set.
Used very subtly, both options can just add a degree of seemingly random variation as your MIDI clip loops during playback. However, if pushed too far and applied to every note, the coherence of the part can be lost. So, let’s consider some useful strategies to ensure that our efforts to inject a human feel also remain musical.
Changing Hats
Let’s start with one of the more obvious candidates: adding some character to a hi‑hat pattern. For example, the screenshot shows a two‑bar hi‑hat pattern in the Drum Editor. In the first bar, this is a simple 16th‑note pattern, while in the second bar the hits follow a 32nd‑note pattern. Some small amounts of positive or negative Velocity Variance have been applied throughout (so that the loop doesn’t sound too robotic on playback), but the hits placed on the 16th‑note grid are all set to 100 percent Play Probability. They will, therefore, always play, and thus ensure the pattern benefits from a consistent ‘core performance’.
That’s not the case with the 32nd notes in the second bar, though. These have Velocity Variance applied but they also have low Play Probability values. These start at around 10 at the beginning of the bar and reach about 35 at the end. On playback, therefore, each of these 32nd‑note hi‑hat hits have a relatively low probability of being triggered, but they’re more likely to be triggered towards the very end of the bar. The end result (which you can hear in the audio examples that accompany this workshop on the SOS website) is a variable degree of 32nd‑note syncopation that’s somewhat stronger at the very end of the two‑bar phrase. This can be really effective: a solid core to the part, but with some variations superimposed upon it each time playback cycles through the loop.
There are two general Play Probability strategies being combined in this example that make it more likely you’ll end up with a musically useful result, and they provide a useful starting point for your own experimentation. First, don’t apply Play Probability to the core hits that give the drum part its foundation — instead, use it on ‘extra’ notes. Second, don’t apply it everywhere. Focus on just a part of the overall pattern to give the variations themselves a sense of regularity. In this case, that’s achieved by the two‑bar structure (one bar played straight, the other with variations).
We can obviously apply these strategies to other elements of a drum part, and snare ghost notes (additional snare hits played softly and usually syncopated) are a case in point. Again, the core heartbeat of the kick/snare groove might be left intact, but additional snare hits, set initially with a lower MIDI velocity, can be placed within the pattern. These can then have some Velocity Variation applied. However, the key element is the Play Probability, and with suitable values you can control where these additional ghost hits are most likely to appear within the overall pattern. Again, I’ve provided an audio example (with some additional commentary) that demonstrates this in practice.
Accidental Bass
These same principles can be applied to instruments other than drums, and a good candidate is bass grooves. Again, a two‑bar looping pattern makes a good starting point and the screenshot shows the ‘after’ version. All the core notes of the performance have Play probability set to 100 percent. However, four additional ‘accent’ notes, placed in the second half of each bar, have been added to the part (you can hear both the ‘before’ and ‘after’ versions in the audio examples) and these have been given Play Probability values in the 30‑40‑percent range.
Play Probability can also be used to add accent notes to melodic instruments such as bass.
Again, the core notes of the pattern play every time, but these accent notes pop in and out as the pattern is looped to add some performance variety in the second half of each bar. You can obviously adjust when/where these notes might appear (for example, only in the final bar of a four‑bar loop) and how busy the effect might be (more added notes). Of course, exactly the same principles can be applied to other melodic instruments.
Rein In The Randomness
Both Velocity Variance and Play Probability are, essentially, randomisation processes, and although you can steer those processes with the settings you use, it’s kind of the point that some unpredictable details will appear within the performance. You create your core loop, add some randomisation elements, and then copy that loop multiple times to occupy suitable sections of your overall arrangement (a verse or chorus, for example) along the project’s timeline.
Of course, while the results can be great, and very musical, they’ll also be different every time you hit the playback button. There might be occasions when you are fine with this but you might not get what’s desired when you hit Export to generate your final mix. So, is there a way to ‘lock in’ the best results from your randomisation experiments? Yes! The Merge MIDI In Loop function (from the MIDI menu) lets you do just that. You can use different combinations of steps to achieve this, but a sensible approach is as follows...
The Merge MIDI In Loop command lets you lock in the performance variations, as shown here using multiple copies of the two‑bar bass loop from the earlier screenshot.
First, duplicate the required MIDI/virtual instrument track (the ‘sensible’ bit; you can always go back to the original if required) and then solo that track. Second, place the left/right locators around the copies of the MIDI loop where you want to lock in the Velocity Variance and Play Probability settings. Third, from the MIDI menu, select Merge MIDI In Loop, and in the dialogue box that appears make sure to tick the Erase Destination button. When you then click OK, a single new MIDI clip replaces all of the MIDI loop clips.
If you then open that clip in the Drum or Key Editor, you’ll see that your randomisation elements have been applied, and the Velocity Variance and Play Probability values reset. You can then audition sections of this new MIDI clip to find the best performance variations your randomisation efforts have created and simply copy/paste these as required along the timeline to create your final performance, complete with all of its interesting human‑esque variations. Ta‑da! Your super‑cool performance additions and accents will now appear at the same points every time you play through your project.
Thursday, August 13, 2026
Wednesday, August 12, 2026
Cubase 14: Create Driving Bass Sounds
All the settings required for our Retrologue bass synth sound can be accessed from a single page of the GUI.
Explore the world of bass synth sound design with Cubase’s Retrologue and HALion Sonic.
Whether in electronic music or contemporary film, TV or game scores, powerful pulsing bass synth sounds are often used to drive the music or visual action along. If you want such a sound, you might get lucky searching for the perfect preset — but designing your own is not only satisfying, but will also ensure you get a sound that’s a perfect fit. It’s something that you can easily do in Cubase Artist/Pro’s Retrologue 2 synth plug‑in. And while Elements users could buy that synth separately, they already have access to HALion Sonic, which is more than capable enough for the job. So let’s get rolling some DIY driving bass synth sounds...
Bass‑ics
Bass synth patches come in all sorts of sonic shapes and sizes, so before we start we need a target in mind. We’ll aim for something that can provide a pulse‑like rhythmic element, and that perhaps suggests a fast attack and a short release, allowing short staccato notes to be played without a rapid progression of notes getting in the way of each other. I’d also like it to be dynamic. In this case, I’m thinking it should provide a solid (maybe slightly subdued?) low‑end tone for any low‑intensity sections of our project, but, while retaining that low‑end foundation, it should also be able to get more strident (maybe more aggressive?) when greater intensity is required. Finally, I’d like to be able to gradually transition between these two sonic characters in real time, which means we may need to include some velocity or controller‑based parameter modulation.
Retrologue 2 makes all of this mercifully straightforward. Indeed, all the settings I needed to create this sound can be seen in the first screenshot, and I was able to configure them without leaving the main Synth page of the GUI. So, let’s now break down just how I arrived at this configuration. And if you want to hear how the sound evolves through the stages I describe below, check out the audio examples that we’ve made available on the SOS website at https://sosm.ag/cubase-0925.
Solid Base
When you open a first instance or Retrologue 2, the ‘init’ preset starts you off with a single oscillator enabled and the filter section’s cutoff wide open. The first element of our target sound — that solid low‑end foundation — is easy to configure from here. In this case, I retained the sawtooth waveform for osc 1 (it’s more harmonically complex than a sine wave, but either could work), but turned the Octave control down one notch. I adjusted the amplifier ADSR settings to ensure a fast attack, a longer decay stage, and minimal sustain or release. Within the filter section, the most important thing was to adjust the cutoff to around 50Hz. But I also added a small amount of resonance and tube distortion, and made some adjustments to the filter envelope.
This combination of settings removed the mid/high‑end fizz from the sound, but I then opted to switch on the sub‑oscillator and adjusted its Mix control, just to blend in a little extra low end. I used the triangle waveform for this, as I liked the fairly smooth character this added to the low end.
Character Enhancer
That simple osc1 and sub combination provides the foundation for our ‘low intensity’ bass sound but, at this stage, it will likely be a bit lacking in character. It certainly isn’t going to get us into the ‘high intensity’ territory that we also want to achieve.
To start tackling that, I turned my attention to the second oscillator, osc 2. Again, I used a sawtooth waveform but, in this case, I pitched it an octave above osc 1, and applied a slight detuning (I used a combination of the Coarse and Fine knobs for this). However, I also set the oscillator type to Multi and used three voices, setting the Det (detuning between these voices) at a subtle 10 cents. If you open the filter fully and audition this oscillator on its own (as in the audio example), what it lacks in low end it makes up for with some gritty mid‑frequency drive, while the use of multiple, detuned voices adds some stereo width. Hold that thought for a minute...
That grit becomes much more subtle when the filter cutoff is brought back down to 50Hz. Then, I simply blended osc 2 into the mix with osc 1 and the sub‑oscillator. Without adjusting any other settings, it immediately adds an extra ‘analogue synth’ character: it’s warmer and with a nice touch of saturation, but without (yet) being super aggressive. For any ‘low intensity’ bass synth target, this should do very nicely.
Added Intensity
As noted above, osc 2 provides plenty of additional aggressive character that we can tap into to achieve our ‘high intensity’ bass tone. There are a number of routes we could explore, including changing the blend of osc 1 and 2. However, I opted for a little hands‑on control configured in Retrologue’s Matrix. Here, I set the mod wheel to modulate both the filter cutoff and distortion controls. Therefore, as you move the mod wheel up, the filter opens, adding in more of that osc 2 grit noted earlier, and more of the filter’s distortion.
You can tweak the two Depth sliders to taste but, as demonstrated in the audio example, we can already move from our solid ‘low intensity’ bass sound to a much more aggressive ‘high intensity’ sound, simply by adjusting the mod wheel position. Oh, and as an optional touch, you can also enable the noise oscillator for an additional sense of ‘angry’ that becomes more noticeable the further the filter is opened.
HALion Sonic provides all the options you need, but you have to dip into the oscillator panel, the amplifier envelope and the filter page to access them all.
Elemental Alternatives
Cubase Elements users can achieve something very similar to what I’ve described above in HALion Sonic (HS). However, unlike with Retrologue, which puts all the options you need in a single GUI page, in HS you do have to go digging a little. A useful first step is to initialise your first Program slot (right‑click on the slot and choose Init Program), to ensure you have a ‘blank’ slot to work in. HS Programs can consist of up to four sound layers, and the controls for each are accessed using the L1‑L4 buttons. With a layer selected, you get access to HS’s various synth engine parameters, spread across the Voice, Pitch, Osc, Filter and Amp panels. For our purposes, we need only one layer because (as shown in the composite screenshot) each one offers multiple oscillator options, much like Retrologue.
HALion Sonic’s Quick Control options allow you to configure the mod wheel to modulate between the low‑ and high‑intensity elements of our bass sound.To construct the basics of our sound, we must dip into three areas: the oscillator panel; the amplifier envelope (which can be edited below the osc panel); and the filter panel. The screenshot shows a composite of these different views, and across these three different panels you can configure something broadly similar to the Retrologue example, with two sawtooth oscillators (set to different octaves, slightly detuned, and with one using multiple voices), a sub‑oscillator and, optionally, a noise generator and a filter with tube distortion, with the cutoff initially set to roll off much of the mid/high‑frequency content.
Finally, we need to map the filter cutoff and distortion controls to the mod wheel. As shown in the final screen, this can be done by right‑clicking on each control in turn, and selecting the Assign Quick Control option, which includes the mod wheel option. Multiple parameters can be assigned to a single hardware control. Instead of a modulation depth option, from this same pop‑up menu you can also set the minimum and maximum values allowed for the target parameter. It’s also worth experimenting with the Set Type options; I found the Absolute option the most straightforward to use in this case.
Compared with Retrologue, configuring the combination of oscillators, filter settings and hands‑on modulation in HALion Sonic is a little more fiddly, since it requires you to visit multiple parts of the GUI. But it works! So, from a low‑intensity bass pulse to a high‑intensity, more aggressive bass‑driven rhythm, all versions of Cubase have a perfectly good set of tools for your DIY bass synth sound design adventures.
Monday, August 10, 2026
Cubase 14: Keeping Ambience Effects Under Control
Automated effects ducking can give your vocals a little extra clarity within the mix.
Ducking your reverbs and delays can bring greater clarity to your vocals.
Ambience effects, such as reverb and delay, are an essential part of vocal production in many genres. But slapping on these effects indiscriminately also risks compromising vocal clarity, and even the whole mix: long effect tails can mask or otherwise clash with different sounds, while if you keep the tails too short, then they might not create the sense of space and dimension you’re looking for. In this month’s workshop I’ll explore some ways to unpick this conundrum.
Options
In a DAW like Cubase, one way to tackle the issue is to set your effects up on FX Tracks rather than as inserts, and create an automation envelope for the send levels from the vocal track. Use a lower send level in the busier sections of a vocal to keep things sounding cleaner, but allowing more of the effect to be heard between well‑spaced words or phrases. Alternatively, you can use automation of the level of the FX Track — similar, but this time, we’re controlling the ‘return’ from the reverb or delay, not just what’s sent into it. Either tactic can work well and they both give you very precise control, but it can also be a fiddly and incredibly time‑consuming approach.
The classic alternative is to use ‘ducking’, whereby a gate or compressor reacts to the vocal to pull the effect’s output level down automatically. Some newer reverb and delay plug‑ins — including the reverb and delay modules in Cubase Pro/Artist’s VocalChain plug‑in — have this option built in. Another, perhaps less obvious alternative for Cubase Pro users is to use the new Envelope Follower Modulator: you can set this up to control the send level to the FX Track. If you want to try this, be sure to dial in a negative Modulation Depth, so that the send level is lowered when the vocal signal increases!
I want to be inclusive here, though, so I’m going to take you through an approach that any user can employ, from Cubase Elements upwards, and with any reverb or delay plug‑in. We’ll start by setting up a conventional ducked effect, before refining it to give us finer control over the result.
Line Up Your Ducks
The main screenshot shows the key elements of of our basic ducking configuration. You can hear this in action in the audio examples on the SOS website at: https://sosm.ag/cubase-1025. My first step was to create conventional sends from the vocal track to two FX Tracks. Respectively, these have instances of RoomWorks (for reverb) and MonoDelay (for delay) inserted. You can set the usual send level controls on the vocal track to configure an initial static wet/dry balance.
Next, insert a compressor after RoomWorks, and another after MonoDelay. I’ve used instances of Cubase’s vanilla Compressor plug‑in here but any compressor with a side‑chain input option will do the trick; we’re not really looking for ‘flavour’ here, but simply level control. Note that both Compressor instances have their side‑chain input activated, and that the vocal track has been selected as the side‑chain source. This means that although the compressors are acting on the effects, their gain reduction will only be triggered by the vocal level; in essence, the louder the vocal gets above the compressor threshold, the more the effect will be pulled down.
The final stage involves adjusting the various compressor settings (threshold, ratio, attack and release), the side‑chain send levels and the FX Tracks’ own faders. It’s pretty easy to influence when ducking will occur (threshold), the amount of ducking (ratio) and, broadly, the speed with which any ducking is triggered and then recovers (attack and release). A fast attack will pull the reverb out of the way quickly so that your vocal words can be heard clearly. You have more room for creativity on the release. But in both cases, you don’t want the transition to sound too obvious, so some experimentation may be required to find the best combination for the project in question. The side‑chain send level and the FX Track faders both influence the overall wet/dry balance, and so let you finesse how the effects are tucked out of the way of the vocal as it varies in level throughout the performance to (hopefully!) achieve a suitable end result.
Freq Out
The classic approach described above might well give you all the control you need. But it might not! If you pay attention to third‑party reverb and delay plug‑ins you’ll notice that some newer designs, such as iZotope’s Aurora reverb, do more than simply drop the reverb level — they adopt the kind of spectral balancing algorithms you find in plug‑ins like Soundtheory’s Gullfoss, Oeksound’s Soothe, Baby Audio’s Smooth Operator and iZotope’s Clarity. Aurora, for instance, ducks the reverb return more at those frequencies where the vocal is loudest, and adjusts this response in real‑time as the vocal’s frequency balance changes throughout the performance. The idea is that the reverb gets out of the way only where it needs to, but the overall sense of the effect can still be heard most of the time. It’s a neat trick!
While the algorithms behind those plug‑ins vary, the concept is rather like using many bands of dynamic EQ — and those bands in turn are similar in concept to frequency‑specific compressors, each focusing its efforts on a very narrow band. All of which raises a question: can we achieve something broadly similar if we use a dynamic EQ plug‑in instead of a compressor to duck the effect? Yes, we can! Well, sort of... The second screenshot shows an instance of Cubase Pro’s Frequency 2 plug‑in configured for such a role, and it could be substituted for the instance of Compressor in either/both of the FX Tracks in our earlier approach.
Pro users can experiment with some frequency‑based ducking using Frequency 2.
I’ve set bands 1 and 8 to basic low‑cut and high‑cut filters, respectively, and these are used to simply shave off the low and high ends of the reverb return. I configured the other six bands to operate in their dynamic mode, and centred each on a different frequency. To keep things simple to follow here, I used identical settings for each band’s Q, gain, ratio, attack and release (shown for band 4 in the screenshot). I then adjusted the threshold of each band to control the amount of ‘ducking’ generated when the vocal contained energy within that frequency range. At any instant the different bands will apply different amounts of gain reduction — in the screenshot, band 4 (at about 1kHz) is providing a 10dB cut, while band 5 (around 2kHz) is only pulling the sound down by about 2dB.
It’s not quite as sophisticated a solution as that found in Aurora, but it’s interesting to try and lets you focus your automated reverb (or delay) ducking (or ‘unmasking’ if you prefer that terminology) on the frequency range within which your vocal is sitting at any given instant. Of course, Frequency 2 is only available to Cubase Pro users, but you can use the same approach with any dynamic EQ plug‑in that supports external side‑chaining. If you don’t currently own such a plug‑in, you could always download a suitable free option, and top of my personal list would be Tokyo Dawn Labs’ NOVA. This free download supports all the major plug‑in formats and provides a four‑band dynamic EQ with side‑chain support and additional low‑/high‑pass filters. The final screenshot shows a configuration that might be used to match what I set up for Frequency 2, and it responds equally well to a side‑chain input in Cubase Elements, Artist and Pro.
For Artist and Elements users, TDR NOVA is a very credible free alternative to Frequency 2.
Duck Duck Go!
You won’t always need to duck your effects tails, and on some more spacious arrangements it might even be counter‑productive. But if you’re looking to coax maximum clarity out of a busier mix, ducking can be really beneficial. It’s also possible that you’ll occasionally want so much control that it’s worth making the effort to automate the send and return levels very precisely. But the sort of frequency‑specific automated ducking I’ve just described might well do the trick very nicely. Whichever route you choose, Cubase Elements, Artist and Pro have multiple options.