Podcast and spoken word mixes need a different ear

Mixing for podcasts and spoken word is not simply a smaller version of music production. The finished piece has to make every word understandable, keep the speaker’s personality intact and remain consistent from one episode to the next. A listener may be commuting through Sydney, walking along Melbourne’s Bourke Street or cooking dinner while a narrator talks, so there is little tolerance for a buried voice, distracting noise or an abrupt level jump.

That said, the goal is not sterile perfection. Spoken delivery benefits from natural breaths, small changes of pace and the texture of a real voice. The engineer’s task is to remove what distracts while protecting what makes the recording credible. A polished dialogue mix can sound warm and cinematic without turning a conversational performance into a heavily compressed broadcast.

A different job from a music mix

A music mix often builds atmosphere through stereo width, instrument balance, movement and detailed automation. Spoken word depends on a narrower set of priorities: intelligibility, tonal consistency, controlled dynamics and an unobtrusive bed for music or sound effects. Two speakers recorded in different rooms should still feel like they belong to the same programme, while a dramatic scene can demand more space without sacrificing the words.

Australian hosts often move easily between dry humour, colloquial phrasing and precise explanation. The mix should preserve that conversational character rather than smoothing it into generic broadcast copy. The workflow therefore changes as well. For music, the mixer may open the session by shaping drums, bass, guitars and synthesisers; for dialogue, clean edits, gain staging and voice matching usually come first.

An experienced music engineer can transfer the technical discipline, but they still need to hear speech as speech rather than force it into the habits of a record mix. That distinction matters when an episode contains interviews, narration, advertisements and an original score.

Build the dialogue around intelligibility

Begin with gain staging. The signal should reach the converter with enough headroom, while the recorded voice should sit comfortably above the noise floor without requiring aggressive digital repair. A boom-mounted microphone or a well-chosen dynamic mic may suit a shared studio; a lavalier can work for an unscripted interview, but handling noise and chest resonance need a light touch. Use a high-pass filter conservatively, often around 70 to 100 Hz, then check whether consonants remain full.

Next comes corrective processing. A de-esser can tame harsh “s” sounds, but Australian vowels and consonants should not be thinned simply to meet a technical target. Compression can even out syllables without removing the rise and fall of the speaker’s voice. Start with a modest ratio, listen for unnatural pumping around pauses and reduce gain reduction if every sentence sounds equally forced. Short phrases, jokes and vulnerable moments often need more dynamic shape than a commercial read.

If music or stock audio appears under the speech, it needs a separate mix bus. High-pass the bed, carve a controlled area around the voice and automate the change with tasteful ducking rather than abrupt volume jumps. This is also where Australian copyright comes into the picture: spoken-word productions using commercial music may need a podcast or synchronisation licence from APRA AMCOS, while advertising or commercial radio use can bring ACAN and other recording rights into scope. Confirm licences before publication, not after a complaint arrives.

Edit for speech, rhythm and personality

Edit the conversation before fine-tuning the mix. Remove clipped words, long gaps and obvious mistakes, but keep breaths and small hesitations when they reveal character. Mark breaths that must remain audible and fades that need smoother transitions. Record several minutes of room tone so edits can be joined without sounding as though someone vanished between sentences. In a noisy flat, a café near Brisbane’s Fortitude Valley or a home office in Perth, the background can change between segments; a consistent noise floor and carefully chosen transitions can make that problem manageable.

Watch body language through the waveform. A speaker leaning away from the microphone should not be made to sound farther away unless the story needs that effect. Normalise obvious level differences, remove clicks and mouth noises, and use crossfades where they will not blur consonants. If two hosts overlap, either edit for clarity or preserve the exchange with separate, intelligible levels. The result should reflect the pace of the conversation rather than the grid of a pop record.

A spoken piece can borrow the atmosphere of a live session without imitating its balance. Looking at live session workflows can help a producer consider how musicians, room reflections and foreground sound are captured together. For narration, the voice still comes first. A modest room sound can provide depth, while a strong music cue should support the message and never ask the listener to work harder to hear it.

Set levels without flattening the voice

Loudness targets are useful, but they are not a substitute for listening. Many podcast platforms normalise around -16 LUFS for stereo speech, while some mono spoken-word specifications use roughly -19 LUFS; video platforms and advertisers may request another figure. Check the destination before export and measure the final programme, not an isolated voice track. Leave suitable true-peak headroom, commonly around -1 dBTP for lossy delivery, so loud consonants do not distort after encoding.

Use the loudness range as another diagnostic. A very wide range can make quiet passages disappear on a phone speaker or in a noisy car. A tiny range may indicate that compression has erased natural expression. Compare the integrated level with the spoken sections and adjust the overall gain only after the balance is right. A limiter at the end can prevent accidents, yet pushing the threshold too low can square off transients and make every syllable feel identical.

Stereo width also requires judgement. Interview dialogue is usually best mixed mono or made mono-compatible, which means checking for phase cancellation on headphones and mono speakers. Atmosphere or a sound bed can remain wide, but the voice should remain centred and clear. This matters for international listeners as well: in Australia, a producer may edit a guest from Perth while working in Brisbane on AEST, or coordinate with Sydney staff when AEDT starts. A shared clock, explicit versions and a written handover prevent time-zone confusion from becoming a mix error.

Plan the production around real deadlines

Audio post-production is part of a larger schedule. A narrator may deliver the final script on Monday, the engineer needs time for edits and revisions, a sponsor may supply late artwork, and the platform team may require specific filenames and loudness reports. Build milestones backwards from the release date, including a buffer for a technical check and one round of sensible feedback. Spoken-word releases can be exposed to the same service lead times as music because artwork, distribution, podcast hosting and advertising assets often depend on separate teams.

The supplied brief should name the platform, target duration, delivery format and whether the episode will be mono or stereo. For lossless delivery, many teams request 24-bit WAV at 48 kHz, but the platform’s own specification takes priority. Include the sample rate, bit depth, loudness target, true-peak ceiling, room-tone requirements and any music licence details. An Australian producer working with a sponsor or independent network should also allow time for client approval and technical standards.

Where an episode includes an original song, a sponsor’s audio or a live take, coordinate the release plan early. Reading release deadline planning can help teams avoid discovering that mastering, distribution or artwork will not be ready in time. Keep the spoken-word edit stable once music clearance and final timings are approved; otherwise a late track swap can force another full mix and delay the feed.

Find the right specialist and start well

Look for a dialogue mixer, not merely someone who mixes records. A useful portfolio includes interviews, narration, audiobooks, voice-over work or documentaries, and the provider should be able to explain how they manage breaths, sibilance, room tone and loudness without flattening the performance. Uptown Plug gives independent artists access to a broad network of service providers, so a creator can compare specialists in mixing, mastering, production, photography, publicity and digital marketing without committing every budget to the same company.

Ask for a sample, a clear revision policy and an estimate based on the real length and complexity of the episode. Hourly pricing can be useful for straightforward narration, while interviews with several speakers, remote stems and original music may need a fixed quote. In Australia, confirm whether the quote includes GST, whether a deposit is required and which currency applies if the engineer works overseas. These practical details are especially helpful for independent creators working from a limited pool of cash.

Browse the Uptown Plug marketplace, compare dialogue-focused engineers and send a brief that lists your platform, deadline, source files, desired sound and budget. You can then book a discovery call, share a short test passage and approve a mix direction before the full episode begins. Clear expectations at the start give the engineer room to protect the speaker’s voice and give the finished programme the polish listeners expect.