Korg All Access: Greg Phillinganes demos the Korg Krome Music Workstation

Podcast Dialogue Editing In Cubase

 By John Walden

The Channel Strip's audio-sweetening tools are as good for dialogue as they are for music.

Reduce the time you spend on podcasting post-production, with these Cubase voice-editing tips.

Some podcasters are happy simply to record their dialogue in a single take, top and tail it with their standard channel intro and outro, and publish, warts and all. But most — while still usually recording an episode in a single take — will perform in a way that allows obvious gaffes, over-long pauses, ums and errs, to be edited out later. Knowing that mistakes can be edited can mean a more relaxed attitude when recording and, for example, allow you to take a deep breath and start again if you fluff your lines. But equally, anticipation of an extended period of detailed editing could deter you from creating a podcast in the first place! So it's important to strike the right balance.

Thankfully, a little once-only work to organise things in Cubase can streamline many repetitive editing tasks — and have you transforming your raw takes into polished 'performances' in no time.

Podcasting Presets

Let's consider a typical situation where the dialogue has been recorded as a single take to a Cubase audio track. Everything you need to do to lick this performance into shape falls into two broad categories: you'll no doubt want to 'sweeten' and 'clean' the audio a touch, using EQ and compression; and you'll need to edit the recording to create a better-paced, more polished-sounding performance.

There are plenty of suitable third-party compressor and EQ plug‑ins, but in the vast majority of cases the Cubase Channel Strip has more than enough firepower. If you record your content using a consistent equipment/location setup, it's worth spending time up front figuring out some EQ, de-essing and compression settings, and creating a Track Preset based on them, for rapid repeat use. The Channel Strip's DeEsser can help tame any sibilance, and I also tend to use the low-cut filter in the Pre section to reduce the impact of any plosive sounds that crept past my pop shield. A combination of a fairly high frequency (perhaps 150Hz) and a steepish filter slope (24dB per octave) will generally take care of the worst offenders without robbing the voice of too much body. You can, of course, adjust the cutoff frequency to taste, or perhaps combine a steep high-pass filter cut with a gentle low-shelving boost that restores any lost body. (And if this isn't cutting the mustard, consider addressing the problem while recording before you look for more post-prod solutions!). Finally, if you've recorded different takes with different mics, Cubase Pro users might consider using the bundled Voxengo Curve EQ's 'match EQ' facility to make the takes more consistent before further processing is applied.

Note that while Cubase's plug‑in suite is great for music production, spoken-word can require more attention to extraneous noises and Cubase doesn't include a dedicated noise-reduction plug‑in. Again, try to address this in your recording environment, but if Cubase's EQ and (multiband) gate, compressor and expander options aren't up to the job, you'll probably want to explore third-party de-noisers (eg. iZotope RX, the freeware Cockos ReaFIR or, indeed, Steinberg's own Wavelab).

If you create this macro and assign a key command to it, each edit requires nothing more than you making a range selection and pressing a key.

Speedy Macros

With some time-saving Channel Strip settings in the bag, it's time to move on to the detailed audio editing. In general, this involves two kinds of task: removing unwanted sections of audio to tighten up the overall flow of the dialogue (including 'ums' and 'errs'); and reducing the level of breaths and other such natural noises (removing them completely often sounds unnatural and thus just as distracting to listeners as when they're too loud!). Both tasks can involve lots of tedious, repetitive, manual editing — but Cubase's powerful system of key commands and macros can turn 'tedious and repetitive' into 'speedy and efficient'!

We've looked at the Cubase Macro system many times (see SOS August 2018: https://sosm.ag/cubase-0818) but the uninitiated need to know that a macro is simply a sequence of key commands, and that the Key Commands dialogue box (accessed via the File menu) is where you view, create and edit macros; just open the dialogue and click the Show Macros button.

Macros can be created and edited in the lower pane of the Key Commands window, and you can assign a macro its own key command in the Macro section of the upper pane. The command sequence for the Cut Time And Fade macro is shown in the lower pane.

Steinberg have made some useful additions to this panel since I last covered it (there are additional buttons on the left) so creating and editing a macro is now much easier. As before, though, once you've created a new empty macro and named it, you simply select key commands in the upper panel and 'add' them to the macro. Key commands can be moved up/down or removed from the macro. And, once you've created your macro, if you go to the Macros section of the key command list, you can also assign a dedicated key command to run the macro. You can also run macros within macros in this way... so a lot is possible.

Cut It Out

If you prefer your podcasts free of ums, errs and outright 'bad' takes where you had to repeat a section, the most time-consuming editing task will be cutting out all these vocal fails. You need not only to audition your performance, but to zoom in and select the offending sections, cut the unwanted audio, and move what remains to the right of the cut section so it neatly fills the gap you just created, butting up nicely against audio to the left of the cut section, apply suitable crossfades, and then repeat the process multiple times until all your cuts are made.

A little once-only work to organise things in Cubase can streamline many repetitive editing tasks — and have you transforming your raw takes into polished 'performances' in no time.

If this sounds like a recipe for RSI, fear not, as Cubase can prevent the pain! Switch to the Range Select tool and you can highlight the section of audio you wish to remove. Then, execute the Edit>Range>Cut Time command (the default shortcut is Cmd-Shift-X on the Mac, Ctrl-Shift-X on Windows), which will instantly delete the selected audio and move all audio to the right of the cut on the current track accordingly. Select, key command, done, repeat.

If 'auto fades' is switched on (via the Project menu's Auto Fades settings dialogue), every time you cut audio in this fashion a fade-out/fade-in will be applied on playback, reducing the possibilities of any click and pops at your edit points. Note that auto-fades aren't actually visible on the clips themselves, though. If, like me, you prefer to see any fades but don't want to have to insert them all manually, the Cut Time And Fade macro shown in the screenshot can do the Cut Time process and add the required short fades to the events either side of the cut.

This macro includes a number of steps. The first three commands set the left and right locators around your selection, move the cursor to the left locator, and execute the Cut Time command. The remainder of the commands move the cursor either side of the edit point, select each audio event that lies under the cursor in turn, and then apply suitable short fades. Finally, the last command ensures that the Range selection tool is still selected, so that, after the macro has run, you can move swiftly to your next edit point.

If you create this macro and assign a key command to it, each edit requires nothing more than you making a range selection and pressing a key. This is perhaps the most common type of audio edit any podcaster will make and, via this macro, the process is about as efficient as it can get — it's a massive time-saver!

A breath noise that has been processed via the Breath And Fade macro. The breath itself has been reduced in volume and a fade-in/out applied to the audio event containing the breath. Fades have also been added to the audio events on either side of the selection.

Breathe Easy

The Breath And Fade macro comprises a long series of key commands, but the whole sequence can be executed with a single key.Breath noises can often be distracting. But as I said above, deleting them can sound unnatural, and simply reducing their volume so they're less prominent can often lead to a better result. The same can go for some other mouth noises.

Again, a macro can be created so each edit can be executed with only a simple range selection operation and a keystroke. I included a macro for reducing the level of breath noises in the SOS August 2018 column I linked to above, but I've expanded on that process in the Breath And Fade macro shown here. Many commands used are similar to the Cut Time And Fade macro, but adding the fades requires more steps, simply because three audio events are involved rather than two.

Fades aside, the key part of the process splits the audio range you've selected containing the breath into a separate audio event. It then applies six instances of the Decrement Event Volume command, each instance adjusting the event volume handle by about 1dB. You can, of course, change the number of instances for less or greater volume reduction to suit your needs.

Published June 2020

Korg All Access: Aaron Draper Preps for Dr. Dre's Coachella Set with his Wavedrum

Cubase: Delta Monitoring Explained

By John Walden

With some simple signal routing, Cubase can tell you exactly what a plug‑in is doing to your audio.

A limiter's primary function is to prevent the loudest signal peaks from exceeding a specified maximum level, but they're often also used as a 'maximiser' to increase loudness — as you raise the input signal, everything below the limiter's peak detection threshold gets louder, while the peaks are reduced to the threshold level. But the latter tactic can mean the limiter both acting more frequently and applying more attenuation, and the more you ask your limiter to do, the more unwanted sonic artifacts it will leave behind. Eventually, it will become audibly unpleasant.

The Brickwall Limiter inserted on a drum bus Group Track, with the track's Pre Gain control used to adjust the input gain.You can't simply crank up the gain and hope for the best, then; you must train yourself to hear exactly how the limiter is changing your audio. When you overdo it significantly, it's very easy to hear any damage being done. But, especially when you first start experimenting with limiters, it can be harder to judge where the sweet spot is.

To make this a little easier, you can use a technique that's often referred to as 'delta monitoring'. Essentially, this requires you to subtract the processed signal from the unprocessed one, so that you can listen to the remainder, which is what your limiter is removing. Some third-party plug‑ins helpfully include delta monitoring facilities (Tokyo Dawn Labs' Kotelnikov and Limiter 6, for example), but a little creative audio routing allows you to achieve it with almost any plug‑in in Cubase (or, indeed, any other DAW). I've used Cubase's simple Brickwall Limiter for the examples that follow.

Hitting Bricks

You can start by familiarising yourself with the Brickwall Limiter by using it on a drum bus track, as per the first screenshot — I've used the Gain control in the MixConsole's Pre section to set the signal level coming into the Brickwall Limiter. This plug‑in offers switchable dual-mono or stereo operation (via the Link button; best left engaged unless you have specific reasons to change it) and an auto-release option that generally works well. The most important control, though, is the Threshold slider, which sets the maximum level a signal can reach before limiting is applied. Any peak that exceeds this level is quickly brought down to the threshold.

To reduce the possibility of exceeding 0dBFS, you should set the Threshold to allow a small margin of error — the screenshot example shows 2dB of headroom left on a drum bus limiter. But if in doubt, the DIC (Detect Intersample Clipping) feature adds an extra level of safety; it uses a lookahead oversampling process that, at the cost of an additional 1ms of latency, ensures the signal won't exceed the Threshold, even at a point between two samples.

Take It Away

Delta monitoring: the final track and routing configuration required to isolate the audio differences between the processed and unprocessed versions of the drum bus signal.Now let's look at how we can set up delta monitoring to better hear what this limiter is doing — the final configuration is set out in the second screen, but a number of steps are required to get there. I'll assume we're starting with a limiter on the drum bus, as described above.

First, create two FX Tracks and insert an active (ie. not bypassed) instance of the Brickwall Limiter on each. Initially, configure these two instances identically to the one on your drum bus. Then disable the drum bus track's output (in the Routing section), and create two sends, each at unity gain (0dB), each going to a different one of your new FX Tracks. The audio from the drum bus now flows to the master output only via these two FX Tracks, so you can bypass the Brickwall Limiter on the drum bus track. Don't engage playback just yet, though, or the signal will be twice as loud as the original!

Now, in the Pre panel of the second FX Track, switch the 'phase' (polarity) from 0 to 180 degrees. Engage playback now, and you should hear silence — the two FX Tracks are playing identical audio out of phase and therefore perfectly cancel. Finally, bypass the Brickwall Limiter on the second FX Track. Now, identical parts of the audio will still cancel, but where the limiter has acted you'll hear only what it has removed.

Split The Difference

That was easy enough, but what does it tell you? Well, for a quick example (because it results in an obvious difference), experiment with the Threshold setting and listen to how the characteristics of the delta signal change. In practice, though, the Threshold is likely to be a 'set and forget' parameter. It's more interesting to change the input level (in this case, via the Gain control within the drum bus track's Pre section).

The delta monitoring signal will include any gain differences introduced by the processor.

As you'd expect, as you increase the gain, you force the Brickwall Limiter to work harder, and the differences between the processed and unprocessed audio become increasingly obvious. These differences manifest themselves in three ways. First, you can hear which elements within your audio are being altered most by the limiting (whether on a drum bus or a full mix, this will probably be the kick and snare drums). Second, you may be able to hear some audio nasties, which may encourage you to think about just how far you should push your limiter. Third, the delta monitoring signal will include any gain differences that are introduced by the processor itself.

In my example, the gain change is applied equally to both FX Tracks because it's performed on the drum bus Group Track that feeds them — so you won't hear any gain differences. But if the plug‑in being monitored is one with an input gain control (Cubase's Limiter, for example), and you adjust this in the FX Track instance of the plug‑in, then the delta signal will include this gain difference, and it will be present whether or not the limiter's gain reduction circuit is active. (It's also a great way to spot plug‑in presets that try to trick your ears with a sneaky 'louder-sounds-better' boost!) There are pros and cons to both approaches: it can be handy to hear just the limiting, or a combination of the limiting and any gain change.

Listen & Learn

Of course, while I've used limiting for this example, the same approach can work with almost any plug‑in. For example, try it with a compressor, or tape saturation processor if you struggle to hear precisely what it is they're doing. As a means of educating your ears, it's a really helpful technique. Hearing the differences isolated in this way can be really interesting, and when you then go back to monitoring the processed signal in the normal way, you'll hopefully have a better idea of just what artifacts to listen out for as you adjust the plug‑in's controls. 

Published July 2020

Korg All Access: Lance Tolbert demos the Korg Krome Music Workstation

How To Use Cubase's Spectral Comparison EQ

By John Walden

Cubase Pro 10.5's Spectral Comparison EQ can be helpful when adjusting the frequency relationship between two sources, such as kick drum and bass guitar, whose frequency ranges overlap.

Cubase Pro's Spectral Comparison EQ makes it easy to figure out why things sound too different or too similar!

In Cubase Pro 10.5, Steinberg added a very useful Spectral Comparison EQ option to the Channel Settings window. This makes it easy to compare the frequency distribution of two audio signals, in real time, providing a very useful visual guide to complement what your ears might already be telling you. There are plenty of scenarios in which such comparison can be useful, and I'll explore two here:

• managing the relationship between a kick drum and a bass guitar.
• comparing the overall balance of your mix to that of a reference track in the same musical style.

EQ For Two

Let's start with the kick and bass example. In the first screenshot (above) I've selected my bass guitar track and, in the Equaliser tab of its Channel Setting window, engaged the Activate Channel Comparison button. In the second (orange) drop-down slot, I've selected my kick drum track. As the project plays, I now see the EQ spectrum for my kick drum (orange) through that of the bass guitar (blue). This makes it easy to spot the two parts' peaks and troughs and identify frequencies where one part might be masking the other. You can then devise EQ settings to give each part enough space to play its intended role in the mix.

But this tool isn't just a visual aid; the Spectral Comparison feature also makes it easier to do the second part of this job. By default, the displayed EQ controls will be active for the currently selected channel (in my case, the bass guitar in blue). But clicking on the orange drop-down tab toggles the EQ controls to the kick drum (orange) track — so I can quickly flick back and forth to make the required EQ adjustments. As both tabs also feature solo buttons, you can isolate either signal (or both) from the rest of the mix.

In this kick/bass example, I eventually took a few decibels out of the bass at about 60Hz and a few out of the kick at around 110Hz. In conjunction with a little side-chain compression (to duck the bass guitar by one or two dB during kick-drum hits) this improved the clarity of the low end of the mix, without adding lots of additional energy that might devour the remaining mix headroom.

Master Strokes

In SOS July 2019's Cubase 'Reference Tracks' workshop, Matt Houghton outlined various strategies for comparing your own mix with reference tracks when preparing your final mix or doing DIY mastering. The new Spectral Comparison EQ feature can play a useful role in this process, helping you visualise what EQ moves might be required to nudge the tonal balance of your own mix towards that of your target. It can also be combined with Cubase Pro's metering options, so you can compare loudness between your mix and the reference.

It's not just a visual aid... the Spectral Comparison feature also makes it easier to adjust the EQ of the two channels.

To compare the tonal balance or loudness of your mix to those of a reference, you need to be able to switch quickly between hearing your own mix, with any master-bus processing applied, and the reference mix, without that also passing through your master-bus signal chain. As Matt described, there are various ways to set this up in Cubase, but if you want to exploit the Spectral Comparison EQ alongside any overall level metering, the following approach will allow you to switch between the mix-in-progress and the reference track very efficiently. It requires a little configuration, and there are six steps.

First, configure a Group Channel to act as the 'master bus' for your own mix (we'll call it 'Master Bus'). Other tracks from your mix that are normally routed to Cubase's Stereo Output should be routed to this channel, including group channels such as drum buses that act as submix buses. Any mastering-style processing you want to apply to your mix can be inserted on this new Master Bus group track, which can then be set to output to the Stereo Output.

Second, you need to add your reference track to your project and insert it on a standard stereo audio track. This track should also be routed to the Stereo Output track, so it doesn't pass through any 'mastering' processing you apply on the Master Bus channel. In the Project window, drag this 'Reference Mix' track so it appears directly below the Master Bus one. Your Master Bus and Reference Mix tracks should now also be sitting side-by-side in the MixConsole.

Third, we need to create two simple macros, as per the screenshot. I've named these Track Left Solo and Track Right Solo; when executed, all they do is toggle the mute status of the currently selected channel in the MixConsole, navigate to select the track immediately to the left (or right), and then toggle the mute status of that track. Once created, assign these macros to convenient key commands.

Don't be scared: the two macros described in the text are easy to create in the Key Commands window.

Fourth, with the MixConsole in the foreground, select the Master Bus track. Open the Channel Settings window for that track and, in the Equaliser tab, click the Activate Channel Comparison button. Select the Reference Mix track in the orange drop-down menu (the Master Bus track will already be selected under the blue drop-down). Open the Channel Settings Function Menu from the down-arrow icon located top right and deselect the option called "Follow e buttons or selection changes". Normally, the contents of the Channel Settings window change to reflect the selected channel in the MixConsole but deselecting this means that even if we change the track selection in the MixConsole, the Channel Settings will continue to display the Master Bus track settings, so we can keep our comparative EQ display for the Master Bus and Reference Mix tracks in view.

Fifth, open the Right Zone on the MixConsole and select the Meter option (upper tab), and the Loudness option (lower tab), to display the Loudness Meter. A detailed discussion of the options here is a topic for another day but, for now, just focus on the visual feedback of the meter if you want to compare the loudness of your mix to the reference.

Finally, mute the Reference Mix track and select the Master Bus track. When you initiate playback, you'll hear your mix; but if you then execute the Track Right Solo macro, the Master Bus track will be muted and the Reference Mix track selected and unmuted. You've instantly toggled playback from your mix on the Master Bus track, to the reference mix. Apply the Track Left Solo macro to flip back to your own mix.

Together, Cubase's Spectral Comparison EQ facility and its Loudness Meter can provide very useful insights when comparing your mix with a reference track.

As you do this, the Spectral Comparison EQ display allows you to compare the tonal balance of the mix and reference in real time. And, if you spot frequency ranges where your mix might have relatively more or less energy than your reference, you can easily experiment with EQ adjustments.

In terms of loudness comparisons, watching where the Loudness Meter settles as you switch between the two tracks is the best place to start. But if you want to use the numerical readouts, make sure you reset the meter (the rightmost button above the numerical panel) as you switch between the mix and the reference.

It's also worth noting that, while the overall shape/slope of the two EQ distributions can indicate the similarity (or otherwise) of the tonal balance between your mix and the reference track, the vertical axis of the EQ spectrum is also quantifying differences in signal level. So, for example, if the overall frequency curves show a similar distribution of low, mids and highs, but your mix just has lower amounts of all frequencies (as per the example in the second screenshot, where the blue curve mirrors the orange curve but sits consistently beneath it), you might want to add some overall gain to your mix — or attenuate the reference — before making more detailed EQ tweaks to refine the frequency-distribution match.

Making Space

Of course, sorting the kick and bass relationship and informing mastering decisions aren't the only applications for real-time frequency comparison. For example, you might compare your lead vocal with a bus containing the instrumental elements that make up the rest of your mix. You can then experiment with EQing to create space in the instrument bed for the most important vocal frequencies, to help improve the clarity of the vocal and intelligibility of the lyrics. Spectral Comparison EQ lets you go compare! 

Published August 2020

Korg All Access: Backstage with keyboardist Mike McKnight