How to Hire an Audio Engineer

When you hire an audio engineer, you are buying judgment as much as hours. Two engineers can open the same raw files and deliver very different results: one produces a clean, even, intelligible program at the right loudness for its platform, with every edit invisible; the other produces something technically "finished" that sounds processed, cuts off breaths mid-word and pumps every time a guest laughs. The software they use is often identical. What differs is listening skill, knowledge of delivery standards and the discipline to follow a spec.

This guide is for anyone who has to make that hire: podcast producers, marketing teams publishing branded audio, video producers who need audio post against picture, course creators, agencies and in-house teams that have outgrown doing it themselves. It walks through the process in order, from defining the work to measuring whether the hire is working, with the criteria, questions and pitfalls at each step.

A good hire, in short, looks like this: the engineer understands what you are making and who will hear it, asks about your delivery spec before starting, returns work on the turnaround agreed, handles your files and naming the way you asked, explains trade-offs when the source material limits the result, and gets faster and more consistent as they learn your show or brand. Everything below is designed to find that person and confirm you have found them.

What a good audio engineer actually delivers

"Audio engineer" covers a wide range of work, and a strong hire in one area may be the wrong hire in another. Before you look at anyone, it helps to know what the finished result should be and what skills produce it.

The core craft

• Editing. Removing false starts, long pauses, crosstalk, filler and tangents so the program flows, while preserving the natural rhythm of speech. Good edits are inaudible. Bad ones clip breaths, create unnatural pacing or change the meaning of what someone said.
• Cleanup and repair. Reducing noise, hum, room reverb, mouth clicks and plosives. Repair tools are powerful, but past a point they stop helping and start producing watery, artifact-laden speech. Knowing where that point is matters more than owning the tools.
• Levels and dynamics. Making speakers sit at consistent levels relative to each other and across the program, using gain, compression and automation, so the listener never reaches for the volume control.
• Music and sound design. Editing music to fit, placing beds under speech without masking it, and designing transitions and effects where the format calls for them.
• Mixing and mastering to a target. Delivering a final file at the loudness and peak level the platform or broadcaster expects, in the right file format, with the right metadata.

Delivery standards they should know without looking up

A capable engineer knows the common loudness targets and asks which one applies. Podcasts are commonly delivered around -16 LUFS integrated for stereo, with true peak kept below about -1 dBTP. US broadcast television follows ATSC A/85 at -24 LKFS, and European broadcast follows EBU R128 at -23 LUFS. Streaming video platforms and ad networks often publish their own specs. The candidate does not need to recite these, but they should immediately ask "what is the delivery target?" and treat the answer as a requirement, not a suggestion.

Specializations worth separating

Podcast editing, audio post for video, commercial and corporate audio, music production and live sound overlap, but they are different jobs. An engineer who mixes dialogue against picture thinks in timecode, stems and M&E versions. A podcast editor thinks in story, pacing and episode consistency. If your need is mostly podcasts, our more specific guide on how to hire a podcast editor goes deeper into that role.

The hiring process at a glance

The steps below run in order. Skipping ahead, for example shortlisting before you have written a brief, is the most common reason hires disappoint: you end up choosing on portfolio polish instead of fit for your specific work.

Step 1: Define the work before you hire an audio engineer

Most mismatches start here. "We need someone to do our audio" can mean light cleanup of well-recorded interviews or heavy editorial restructuring of three-hour recordings with poor microphones. The two jobs need different people and very different budgets.

Questions to answer internally

• What is the output? Podcast episodes, audio for video, voice-over finishing, ads, audiobooks, course audio, sonic branding elements.
• What is the volume and rhythm? One episode a week, a batch of forty course videos once, an unpredictable flow of marketing pieces.
• What is the raw material like? Runtime of recordings versus finished length, number of speakers, whether each speaker is on a separate track, the recording environment.
• How much editorial judgment is needed? Are you asking for cleanup, or for the engineer to shape the story, cut tangents and restructure?
• What are the delivery specs? Loudness target, file format, sample rate, stems, alternate versions, transcripts.
• What turnaround do you need? And how often will you need rush work?

Why the raw material matters so much

Runtime and raw material drive the effort more than anything else. Three hours of recording for a thirty-minute episode is an editorial job, not a cleanup job. Recording quality sets both the cost and the ceiling: one track per speaker in a quiet room is straightforward, while four people on one microphone in a reverberant room is expensive, and the ceiling on the result is lower. How much repair is needed also matters, because noise reduction, de-reverb and clip repair are slow, and past a point they stop helping. If your recordings clip regularly, read our guide on repairing clipped audio to understand what is and is not recoverable, and consider fixing the recording setup before you hire.

Step 2: Freelancer, agency or dedicated engineer

There is no universally right model. The choice depends on how much work there is, how steady it is, how much consistency matters and how much risk you can accept if one person becomes unavailable.

Freelancer

A single independent engineer can be an excellent choice for steady, moderate volume where one consistent ear matters, such as a weekly show. You get a direct relationship and, over time, someone who knows your voice intimately. The risks are capacity and continuity: illness, vacations or a bigger client can leave you without cover, and quality depends entirely on one person.

Studio or agency

A studio brings a team, a documented process, backup capacity and usually a broader range of skills, such as sound design, music editing, transcription and audio post under one roof. It suits variable or high volume, multiple formats and teams that need predictable turnaround. The trade-offs are that you may work with more than one engineer, so consistency depends on the studio's process, and you are paying for some management overhead. Our guide to choosing an audio studio covers how to evaluate one.

Dedicated engineer

A dedicated engineer, whether an in-house hire or an engineer assigned exclusively to you through a studio, works in your session format, your naming and your delivery spec. It suits teams with continuous volume or live-adjacent workflows where the engineer needs to be embedded in production. It is the most consistent model, but it is only efficient if there is enough work to keep them usefully occupied.

What it costs

Typical US market ranges give a sense of scale. A podcast episode runs $75 – $400 per episode for edit, cleanup, levels, music and mastering to a podcast target, priced per finished episode. Commercial and corporate audio, meaning voice-over finishing, sound design and mixing to a broadcast or platform spec, runs $400 – $2,500. Audio post for video runs $50 – $300 per finished minute for dialogue editing, effects, music and a mix delivered against the video's loudness target.

For comparison, our published starting rates are: podcast episode from $165 per episode, covering edit, cleanup, levels, music and a master at the right platform target, with turnaround of 2–4 business days; audio post for video from $95 per finished minute, covering dialogue, effects and music mixed against the picture and its loudness spec, with turnaround of 3–5 business days; and a dedicated engineer from $85 per hour, working in your session format, your naming and your delivery spec, with a start within 48 hours. These are starting prices, not guaranteed totals; the final figure depends on the cost drivers below. You can see how that work is scoped on our audio editing and production services page.

Whoever you hire, the same factors move the price: runtime and raw material, recording quality, how much repair is needed, sound design, the number of versions and transcript accuracy. Different loudness targets, stems and an M&E version for localization are each a deliverable. A raw automatic transcript is cheap; a corrected one with names and jargon right is a person's time.

Step 3: Write a brief candidates can actually respond to

The brief is the single document that makes candidates comparable. Without it, each one quotes for a different imagined job, and you end up comparing numbers that do not describe the same work.

What to include

• Context. What the program or project is, who the audience is and where it will be published or broadcast.
• Scope. The deliverables, the volume and the expected rhythm, such as one episode per week or a one-time batch.
• Source material. Recording setup, typical runtime of raw recordings, number of speakers and tracks, and links to the representative samples you collected.
• Editorial expectations. How much cutting and restructuring you expect, and what must never be cut, such as sponsor reads, legal disclaimers or particular segments.
• Sound references. Two or three published episodes or pieces, yours or others', that sound the way you want.
• Delivery spec. Loudness target, true peak ceiling, file format, sample rate and bit depth, stems, alternate versions, file naming and delivery location.
• Turnaround and review. When raw files arrive, when the first edit is due, how many revision rounds are included and who gives feedback.
• Rights and confidentiality. Music licensing responsibilities, who owns session files and any confidentiality needs for unreleased material.

If music, archival clips or third-party material appear in your productions, spell out who is responsible for clearance. Our guide to audio rights and clearance records explains what records to keep.

What to ask candidates to send back

• Relevant work samples, ideally with a before-and-after for at least one.
• How they would approach your worst sample, and what they think the ceiling on it is.
• Their pricing basis for your volume and what is included.
• Their standard turnaround, revision policy and rush options.
• How they handle backup, file storage and session archiving.

Step 4: Shortlist against the brief

Portfolios are marketing. They show the engineer's best work on material they chose, often recorded well. What you need to know is how they handle material like yours, delivered to your spec, on your schedule. Shortlist on that basis.

Listening to portfolios properly

Listen on the equipment your audience uses, typically earbuds and a phone speaker, as well as on good headphones. Pay attention to the moments where engineering is hard: transitions between speakers, laughter, crosstalk, music entering under speech, the end of sentences where edits often land. Listen for breaths that are cut unnaturally, background noise that swells and recedes between words, and voices that sound thin, metallic or overcompressed. A clean-sounding sample on beautifully recorded source proves little; a clean result on difficult source proves a lot.

Questions to ask on a short call

A fifteen-minute conversation with each shortlisted candidate fills gaps the written response leaves. Keep the questions practical and listen for specifics rather than reassurance.

• "Walk me through what you would do with our worst sample." A strong answer names the problems they hear, the order they would tackle them in and what they think the realistic result is.
• "How do you check loudness before delivery?" You want to hear about metering the whole file against the stated target and true peak ceiling, not "I trust my ears."
• "What happens if you are sick the week an episode is due?" Studios should describe their cover; freelancers should have an honest plan, even if it is simply early warning.
• "How do you prefer to receive feedback?" Engineers who have a clear preference, such as time-stamped notes in one document, have usually thought about making revisions efficient.
• "Where do our files live, and for how long?" The answer tells you whether they have a storage and archiving routine or improvise one per client.

Write the answers down against each name. When you compare candidates a week later, notes beat memory, and the specifics often separate two portfolios that sounded equally good.

Red flags at the shortlist stage

• Promises that any recording can be made to sound like a studio recording.
• No questions about your delivery spec or audience.
• Vague pricing that does not say what is included.
• Samples that are all music, when your need is spoken word, or the reverse.
• Reluctance to do a paid test on real material.

Step 5: Run a paid test that tells you something

A paid test is the most reliable predictor you have. Give each finalist the same real raw material, with the same brief, and pay each of them their normal rate for it. Paying matters: unpaid tests attract people with time to spare rather than the busy professionals you want, and they create an awkward question about who owns the work.

Designing the test

Use a real piece of work, not a sanitized one. Your "typical" sample is usually the right choice; your "worst" sample can be added if repair is a significant part of the job. For a podcast, a full episode is better than a clip because pacing and consistency across the whole runtime are part of what you are testing. For audio post, a short sequence with dialogue, effects and music against picture will show how they work with timecode and a loudness spec. If sync is part of the work, our guide on timecode and audio sync lists what to check.

Include a revision round. How a candidate responds to feedback, whether they understand it, apply it precisely and do not introduce new problems, is as important as the first draft. Give specific, time-stamped notes, the same notes to each candidate where possible.

Scoring the test

Score blind if you can: have someone strip file names so the reviewers do not know whose work they are hearing. Use the same criteria for every candidate, and measure delivered loudness with a meter rather than by ear. A simple scorecard keeps the decision grounded in the brief rather than in whose email was friendliest.

Treat spec compliance and turnaround as gates rather than scores. An engineer who misses the loudness target or the deadline on a test, when they know they are being evaluated, is unlikely to do better on a routine Tuesday.

Step 6: Agree terms that prevent the usual disputes

Most disputes in audio work are not about quality. They are about scope creep, revision counts, missed deadlines, and who owns what. Settle these in writing before regular work starts.

What the agreement should cover

• Scope per unit. What an "episode" or "finished minute" includes, and what is extra: additional versions, stems, transcripts, sound design or heavy repair.
• Pricing basis. Per episode, per finished minute, hourly or retained, and how unusual jobs are quoted.
• Turnaround. Standard turnaround from receipt of complete files, what counts as complete, and rush options and pricing.
• Revisions. How many rounds are included, what a round is and how quickly revisions come back.
• Files and ownership. Who owns the final masters, session files and stems, how long the engineer keeps them and in what format they are handed over if the relationship ends.
• Confidentiality. Especially for unreleased interviews, embargoed announcements or internal communications.
• Tools and AI. Whether AI-assisted tools such as voice cloning, generative fills or AI noise removal may be used, and on what terms.
• Termination. Notice period and handover obligations.

AI-assisted tools raise their own ownership and consent questions, particularly anything that synthesizes a voice. Our guide on copyright for AI-assisted audio explains the issues to settle in the contract.

How a typical project runs

For reference, this is how our own projects run: assessment is same day for most material; the first edit takes 2–4 business days; revisions take 1 business day per round; mix and master takes 1–2 business days; and rush is overnight, priced accordingly. Use a schedule like this as a template for the timings you agree with any engineer, adjusted to your own volume.

Step 7: Onboard the engineer properly

An engineer's first few pieces for you will be slower and less consistent than their tenth. Good onboarding shortens that curve. Treat it as a small project in its own right.

What to hand over

• Reference material. Two or three finished pieces that represent the sound you want, and any you consider mistakes.
• Templates. Intro and outro music, stings, sponsor read formats, and session templates if you have them.
• Style notes. What to cut and what to keep, how to handle filler words, how tight the pacing should be, how guests' names are pronounced.
• Spec sheet. Loudness target, true peak, format, sample rate, naming convention and delivery location, written down in one place.
• Access. Shared folders, upload locations, project management tools and the right contacts.
• Review process. Who gives feedback, in what form and by when.

Feedback quality determines how fast an engineer improves. Time-stamped notes, collected in one place from one person, beat scattered comments from several reviewers. Our guide on reviewing audio with clients covers how to give notes an engineer can act on.

Archiving from day one

Agree at the start how sessions, stems and masters are stored and named, and who keeps them. Six months later, when you need to recut an episode or pull a clip for a trailer, you will be glad you did. See our practical guide to archiving audio sessions for a structure that works.

A worked example: hiring for a weekly interview podcast

The following example is illustrative, built to show how the steps fit together; it is not a client story. A marketing team publishes a weekly interview podcast. Each episode is recorded remotely with a host and one guest, each on a separate track, for about 60 minutes, and the finished episode runs about 40 minutes. Recording quality is decent, with occasional guest echo from untreated rooms. They need one mastered MP3 per episode at a podcast loudness target, plus a corrected transcript.

Defining the work. The team identifies this as light editorial work with moderate cleanup, not heavy repair. They collect three samples: a clean episode, a typical one and one where the guest was in a reverberant kitchen.

Choosing a model. Volume is steady at one episode a week, consistency matters and they want backup cover for vacations. They decide to invite both independent engineers and studios, and to let the test decide.

Pricing context. The US market range for a podcast episode of $75 – $400 per episode tells them what to expect; the corrected transcript is quoted separately by most candidates because it is a person's time. A studio quoting from $165 per episode sits within that range; a quote well below the range is a prompt to ask what is left out.

Shortlist and test. From a handful of responses, they shortlist three, run a paid test on the typical episode with one revision round, and score blind with the scorecard. One candidate delivers the cleanest edit but misses the loudness target, which fails the spec gate. Another is slightly less polished on the echo but meets every spec, applies revision notes precisely and explains exactly what could and could not be recovered from the kitchen recording. The team hires the second.

Terms and onboarding. The agreement defines an episode, sets one included revision round, specifies turnaround from complete files and confirms the team owns masters and session files. Onboarding includes the style notes, a pronunciation list for recurring guests and a shared folder structure for archiving.

How to measure whether the hire is working

A hire that passed a test can still drift. Set up simple, regular checks so problems surface early, and review the relationship formally on a schedule, for example after the first few pieces and then quarterly.

What to track

Listener-facing signals

Audience signals can support the picture, although they are influenced by many things besides audio. Listener complaints about volume jumps, hard-to-hear guests or background noise are direct feedback. Accessibility is part of quality too: accurate transcripts and captions widen the audience, and our guide to transcription and audio accessibility explains the choices involved.

When to change course

If spec misses or late deliveries repeat after they have been raised, or revision load is not falling after the first couple of months, have a direct conversation using the log you have kept. Often the issue is an input problem on your side, such as late files, changing notes or unclear style direction, and fixing it fixes the output. If not, the brief, the test material and the scorecard you already built make rehiring much faster the second time.