Ask a viewer why a video felt amateurish and they will usually point at the picture. Ask them to watch two versions of the same lesson – one with clean audio, one without – and the picture stops mattering. Sound is where perceived quality actually lives.
There are four things we do to every audio track before we consider a project finished, and none of them are a one-click filter.
Why audio matters more than picture
Online course students are not watching a film. They are listening to an explanation, often for an hour or more, sometimes with the video in a background tab. Slightly soft focus or a plain background barely registers. A hum, an echo or a voice that jumps in volume is noticed within seconds and never stops being noticed. Poor audio also costs you in reviews: on course platforms, sound quality is one of the most common complaints in low ratings.
1. Noise reduction and 2. EQ
First, noise reduction: room tone, laptop fans, air conditioning and street noise removed without leaving that underwater artefact that heavy processing creates. Second, EQ: cutting the low rumble and taming the harsh frequencies that make an hour of listening tiring.
Both are done by ear, clip by clip. A preset tuned for one room makes another sound thin or metallic, which is why automated “enhance” buttons rarely hold up across a full course.
3. Levelling to the platform standard
Third, levelling. A lesson where the volume changes between clips forces the viewer to keep reaching for the dial, and that is the moment attention breaks. We match loudness across the whole project to the standard the platform expects.
In practice that means consistent integrated loudness, measured in LUFS, across every lesson – usually around -14 LUFS for YouTube and most course platforms, and around -16 LUFS for podcasts – with peaks kept safely below clipping. Students moving from lesson four to lesson five should never notice the join.
4. Timing and pacing
Fourth, timing. Silence between sentences is not dead space to be deleted – it is pacing. We trim the pauses that drag and keep the ones that let a point land.
Removing every pause produces a breathless, robotic delivery that is harder to learn from. The aim is the rhythm of a confident teacher: brisk through the familiar parts, unhurried where something new is being explained.
Getting it right at the recording stage
Editing can fix a lot, but it is always easier to start with a clean recording. A few habits make a large difference:
- Keep the microphone close – a hand span from your mouth for a desk mic, clipped mid-chest for a lavalier.
- Record in the softest room you have. Curtains, a rug and a bookshelf absorb more echo than you would expect.
- Turn off fans, air conditioning and notifications before you press record.
- Record ten seconds of silence at the start. It gives the editor a clean noise profile to work from.
- Switch off automatic gain on your recorder or software so the level stays steady.
- Record at 48 kHz, the standard for video, and keep your peaks well below the red.
Music and sound effects
Music should support the voice, never compete with it. In instructional video we keep background music well below speech and duck it further whenever someone talks, so it adds energy in the gaps without making words harder to follow. Sound effects work the same way: a subtle whoosh on a title or a click on a screen action helps, but a cue on every cut quickly becomes tiring. Every track we use is properly licensed for the platform you publish on.
Different formats, different priorities
A long course lesson needs comfort above all: warm, even audio a student can listen to for an hour. A webinar recording usually needs rescuing – several speakers on different microphones and internet connections, matched as closely as possible. Short vertical clips need clarity at low volume on a phone speaker, so we keep the voice present and forward. A podcast needs consistent levels between host and guest and a slightly lower loudness target. The four steps stay the same; how far we push each one depends on where the audio will be heard.
What can and cannot be rescued
Background hiss, hum, fan noise and uneven levels can almost always be cleaned up convincingly. Moderate room echo can be reduced a great deal. Two problems are much harder: clipping, where the recording was so loud the waveform is flattened, and a microphone placed so far away that the room is louder than the voice. If you are unsure, send us a short sample before you record the whole course and we will tell you honestly what to change.
A quick checklist before you publish
- Listen to the full lesson on headphones and on laptop speakers.
- Check that volume is consistent between lessons, not just within them.
- Listen for clicks, breaths and mouth noise at sentence starts.
- Confirm music never competes with the voice.
If you would rather hand the whole thing over, send us your brief and budget. You get a written scope, a fixed price and a delivery date within one business day. See the full list of video editing services we offer.