VoiceAssist: Automatic Or Manual?

One thing many engineers and producers love about VoiceAssist is that it automates jobs they hate. The work that happens before you get to the part of the mix you love. Removing breaths, cleaning noise, and setting levels without relying on a compressor to do a job it was never meant to do. Jobs that used to take an afternoon now take a couple of minutes.

For example, imagine editing a one-hour podcast episode. Traditionally, you would have to manually scan the whole recording, cut out every distracting breath and background noise, and carefully ride the levels to make sure every speaker is even. With VoiceAssist, you can load your recording, let the software handle all the repetitive cleanup automatically, and spend your energy fine-tuning only the moments that matter. It dramatically speeds up the process, freeing you to focus on creative choices rather than tedious editing.

However, if VoiceAssist works so well on its own, when do you let it do its thing, and when do you take control? The honest answer is that VoiceAssist is built so you rarely have to choose. Every section runs an automatic analysis as soon as it 'sees' your audio and lets you intervene on a single word, breath, or resonance without affecting anything else. The skill is knowing when the automatic pass is enough and when a few seconds of manual intervention pays off.

VoiceAssist analyses the material as soon as it sees your audio, making decisions across all active sections.

Start By Trusting The Analysis

When you load VoiceAssist on a clip, it analyses the material immediately and makes decisions across all active sections. For much of the voice work we see, like podcast edits, audiobook chapters, and cleanly recorded dialogue, that first automatic pass is enough. In more challenging situations, such as extremely noisy environments or recordings with overlapping voices, the automatic results might not fully address every issue, and some manual tweaks may be necessary for the best outcome.

The Dynamics section, which is the heart of the plugin, smooths volume inconsistencies and writes editable automation that follows the performance. Breath detection finds breaths and lowers them. The gate detects where the voice starts and stops using vocal detection rather than a fixed threshold, so it does not clip the start of a word like an old-fashioned gate. Left alone on good material, this is usually all you need. The temptation to fiddle with something that already sounds right is worth resisting.

The first rule is simple, even if it is not exciting: listen before making any changes. Most of the time, the automatic results are solid, and the manual controls are there for the rare moments when you need them.

Where Manual Control Can Help

Manual adjustments are usually needed for creative reasons, not technical ones. For example, you might want to keep a breath because it adds emotion, or bring a word forward because of its meaning, not just its volume. VoiceAssist lets you do this with region editing, which works the same way in the breath, sibilance, and gate sections.

Each detected event appears as its own editable region. Adjust one breath without touching the rest.

Each detected event appears as its own editable region on the timeline. You can adjust its volume, change its boundaries, remove it, or add a new one by hovering at the bottom of the waveform. If the analysis lowers a breath too much, you can simply adjust that one region without affecting the rest.

The Dynamics section works similarly under the automation. You can click and drag any part of the curve to raise or lower a phrase. The manual points out something important: changing other settings will not erase the manual adjustments you have already made. This is helpful because you can let the automatic levelling handle most of the work, then fine-tune a few words by hand, knowing your edits will stay even if you change the overall settings later.

The Settings That Change How Much VoiceAssist Does

There is a middle ground between full automation and manual editing: the controls that shape how the analysis works before you adjust individual regions. This is where many important choices happen.

In the Breath section, you can pick a fixed reduction that lowers every breath by the same amount, or set a threshold so only louder breaths are reduced. This lets you decide whether to treat all breaths the same or only adjust the ones that stand out. The Dynamics section offers a target loudness mode, so you can let it match the overall loudness automatically or set your own LUFS value if you have a specific requirement. There is also a smoothing control, which works like a dry/wet balance for how strongly the automation is applied.

The Breath section lets you pick between a fixed reduction or a threshold-based approach — both shape how the automatic analysis behaves.

The Lowcut feature works in a similar way. Its pitch-based mode tracks the lowest pitch of the voice and adjusts the filter to match it, which is helpful when a performance spans different vocal ranges, since a fixed high-pass filter would make the low notes sound thin. If you know exactly where you want the cut, you can use the custom mode to set it yourself. You can also turn it off completely if you do not need it. There is no single right choice; these options let you decide how much control you want to give the plugin.

Both Work Just Fine

In practice, the engineers who get the most from VoiceAssist do not choose just one approach. They usually start with the full automatic analysis, review the results, and then decide whether anything needs a personal touch.

To help you get the most out of VoiceAssist, here is a simple workflow many engineers follow:

1. Load your audio clip and run the automatic analysis. VoiceAssist applies its processing across all active sections as soon as your audio is loaded.
2. Listen to the result without making any manual changes. For clean voice-overs or well-recorded material, you might find the automatic pass gives you exactly what you need.
3. Review the audio for spots that need a personal touch. In more challenging dialogue or complex recordings, you might use the Clean section first to reduce noise and reverb, then let the gate and levelling work automatically.
4. Fine-tune as needed. Spend a few minutes manually adjusting problem areas, like dialling back a breath for emotional effect or smoothing out a sibilant sound. VoiceAssist makes it easy to tweak specific regions without affecting the rest of the track.

This step-by-step approach ensures you benefit from full automation while still having all the control you want for creative or technical tweaks. VoiceAssist handles the repetitive tasks, while you focus on the parts that need your judgment.

Finding the right balance is what VoiceAssist is all about. Automation saves you from spending hours adjusting clip gain for every syllable. Manual control exists because some choices will always need your input, and no analysis can tell which breath gives meaning to a performance.

And the great thing is that we've designed VoiceAssist to make using both modes as easy as possible, so you can get on with the stuff you like doing most… mixing.