48kHz/24 Bit Voice Over Specs Pros Use for Podcast and Broadcast

Record in WAV at 48kHz and 24-bit, with peaks between -12 dBFS and -6 dBFS and a room noise floor below -60 dB FS. For delivery, target around -14 to -16 LUFS for podcasts and streaming, about -23 to -24 LUFS for broadcast, and keep true peak at -1 dBTP. Drop to 44.1kHz / 16-bit only if a client specifically asks for it.

---

TL;DR:

• Use 48kHz and 24-bit WAV files for most commercial, video, and broadcast voice work, converting to 44.1kHz and 16-bit only if specifically requested.
• Record peak levels between -12 dBFS and -6 dBFS with room noise below -60 dB FS to ensure clean audio and headroom for editing.
• Aim for final loudness around -14 to -16 LUFS for podcasts and streaming, and -23 to -24 LUFS with a true peak limit of -1 dBTP for broadcast.
• Always set your session to the correct sample rate before recording and apply dither only during export when reducing bit depth.
• Outsourcing to vetted voice professionals can save time on technical requirements, especially for strict loudness and noise floor specifications.

---

Table of Contents

• Essential recording specs at a glance
• Sample rate, bit depth, and file format: what to set and why
• Loudness and true peak targets for delivery
• Gain staging and meter targets for tracking voice
• Room, microphone placement, and accessories that actually change specs
• Monitoring, audio interfaces, and headphone and latency setup
• Export and delivery checklist for every platform
• When hiring a pro beats managing the technical chain yourself
• What actually matters most in a home studio
• Consider hiring vetted professionals instead of managing specs yourself
• FAQ
• Sources

Essential recording specs at a glance

Before touching a single plugin, set these defaults in your session. They cover almost every voice over job you will take, from audiobooks to commercial reads to e-learning modules.

• Sample rate and bit depth: 48kHz / 24-bit for video, e-learning, and most commercial work.
• File format: WAV (linear PCM), never MP3 or any lossy format for tracking.
• Channel count: mono for dialogue-only reads, stereo only when the brief calls for it.
• File naming: project name, take number, and date, so editors never confuse versions.

Audiobooks are the main exception. The Library of Congress NLS specification calls for linear PCM at 44.1kHz, 16-bit minimum, which lines up with legacy CD-based audiobook delivery. Video, broadcast, and most commercial platforms expect the higher 48kHz / 24-bit standard instead.

Always bounce your master from the native session rate rather than converting after the fact, and apply dither only when you reduce bit depth on export, never during tracking.

Sample rate, bit depth, and file format: what to set and why

The 48kHz / 24-bit combination has become the default for video, e-learning, and most commercial delivery because it matches the sample rate used in video production and gives engineers more headroom to work with during mixing. A recommended REAPER setup for voice over work confirms this split directly: audiobooks commonly track at 44.1kHz while voice over for video and broadcast tracks at 48kHz.

Bit depth matters just as much as sample rate, though for a different reason. Sixteen-bit audio gives you roughly 96 dB of dynamic range, while 24-bit gives you around 144 dB. That extra range is not about final loudness, it is about safety margin during recording and editing. A quiet take that needs boosting in post will carry far less audible noise at 24-bit than at 16-bit.

Record only as linear PCM WAV, never as MP3, AAC, or any other lossy format. Compression discards data permanently, and once it is gone you cannot recover it for remixing or noise reduction later.

• Keep project settings locked at 48kHz / 24-bit before you record your first take.
• Apply dither only on export, and only when reducing bit depth for a legacy deliverable.
• Archive your original WAV masters separately from any trimmed or processed copies.

If a client needs a different sample rate, convert at export time rather than changing your session rate mid-project. Converting a finished 48kHz master down to 44.1kHz introduces far less artifact risk than recording at a mismatched rate and converting up later.

Loudness and true peak targets for delivery

Three numbers govern whether your file passes a platform's loudness check: integrated LUFS (the average perceived loudness across the whole file), true peak in dBTP (the highest point your waveform actually reaches, including peaks that hide between samples), and short-term or momentary loudness (a rolling measurement engineers watch during mixing to catch sudden spikes).

Targets vary by destination, and getting them wrong is the single most common reason a finished voice over bounces back for revisions.

• Podcasts and streaming: roughly -14 to -16 LUFS, with -14 LUFS the common target for platforms like YouTube.
• Broadcast: EBU R128 recommends -23.0 LUFS in Europe, while US broadcast under ATSC A/85 commonly targets -24 LUFS.
• True peak: -1 dBTP is standard across most professional specs; some theatrical and broadcast chains require -2 dBTP.

EBU R128 sets programme loudness at -23.0 LUFS with true peak capped at -1 dBTP, and recommends measuring with a loudness meter compliant with ITU-R BS.1770. That standard underpins most of the broadcast and podcast targets above, since BS.1770 defines how loudness meters actually calculate LUFS in the first place.

True peak matters separately from integrated loudness because digital-to-analog conversion can create peaks between your measured samples that a standard peak meter never shows you. A file that measures -1 dBFS on a normal meter can still clip during playback if its true peak exceeds that threshold. The practical workflow: monitor integrated LUFS throughout your mix, then apply true-peak limiting as the very last step before bouncing your final file.

Gain staging and meter targets for tracking voice

Clean gain staging during recording saves hours of repair work later, and it starts before you record a single word of real copy.

1. Read a loud test phrase at the volume you will use for your loudest line, and watch your peak meter.
2. Adjust input gain so those peaks land between -12 dBFS and -6 dBFS, per standard gain staging guidance.
3. Record a few seconds of room tone with no speech to confirm your noise floor sits below -60 dB FS.
4. Run a full practice take of your actual script, watching for any peaks that creep toward 0 dBFS.
5. Adjust and repeat until your loudest moments sit comfortably inside that -12 to -6 dBFS window.

Average RMS level matters for long reads too. The NLS audiobook specification calls for RMS between -24 and -16 dB FS for spoken text, a useful reference even outside audiobook work since it reflects how natural, sustainable speaking levels should sit on a meter over time.

Pro Tip: Gain for your loudest line first, not your average line, since a single clipped peak can ruin an otherwise perfect take.

Room, microphone placement, and accessories that actually change specs

Your room matters more than your microphone budget. A noisy or reflective space will undercut even a premium condenser mic, while a well-treated closet can make a mid-range dynamic mic sound broadcast-ready.

• Background noise target: below -60 dB FS for general voice over work, with stricter specs like NLS audiobook contracts requiring as low as -72 dB FS.
• Microphone type: condenser mics capture more detail and suit treated rooms; dynamic mics reject more ambient noise and suit untreated spaces.
• Placement: 6 to 12 inches from the capsule, angled slightly off-axis to reduce plosives without losing clarity.
• Accessories: a pop filter is non-negotiable, and a basic reflection filter behind the mic cuts room reflections cheaply.

Condenser microphones need phantom power, which almost every audio interface supplies at a standard 48V, so confirm your interface has that switch enabled before you start a session. A practical look at noise floor in home recording points out that treating a room for noise and reflections often produces a bigger audible improvement than swapping a decent microphone for a far more expensive one.

Pro Tip: Hang moving blankets or acoustic panels on the wall directly behind your mic position first. That single reflection point causes more smearing than any other surface in a typical home setup.

Isolation matters as much as absorption. Recording inside a closet full of clothes, a blanket fort, or a dedicated booth controls both reflections and outside noise bleed at the same time.

Monitoring, audio interfaces, and headphone and latency setup

Your interface is the one piece of gear standing between your microphone and your session file, so a few checks matter more than the price tag on the box.

• Preamp noise: choose an interface with a low self-noise rating so quiet passages stay clean.
• Phantom power: confirm 48V phantom power is available if you are running a condenser mic.
• Driver stability: use manufacturer-supplied drivers rather than generic ones to avoid dropouts mid-take.
• Direct monitoring: enable it on your interface so you hear your voice with zero latency while recording.

For headphones, closed-back designs with a neutral frequency response work best for tracking, since open-back cans leak sound into your microphone. Keep your monitoring level moderate. Tracking for hours at loud headphone volume causes ear fatigue that leads to inconsistent read levels later in a session.

Latency becomes a problem mainly when you are monitoring through your DAW rather than your interface directly. Lower your buffer size for tracking, then raise it again for mixing and plugin-heavy playback, where stability matters more than speed. External preamps and premium clocks can improve a vocal chain, but for most voice over work the gains are marginal compared to fixing your room treatment and gain staging first.

Export and delivery checklist for every platform

Export your master as WAV (PCM) at 48kHz / 24-bit, mono or stereo depending on the brief, and apply dither only if you are reducing bit depth for a specific legacy deliverable. This combination works as a safe default across nearly every platform you will encounter.

A complete audio delivery standards guide for film and television recommends exactly this fallback: when a platform's exact spec is not published, export at 48kHz / 24-bit with true peak limited to -1 dBTP, since that combination satisfies nearly every professional chain downstream.

• Name files clearly with project, take, and date so editors and clients never confuse versions.
• Tag basic metadata such as title and project name using ID3 tags if the deliverable is podcast-bound.
• Estimate storage at roughly 0.5 GB per hour for a mono 48kHz / 24-bit WAV file, more for stereo.
• Run final QA by checking integrated loudness, true peak, and listening once on headphones and once on consumer speakers.

If you are delivering video alongside your voice track, pair your audio export with a captioning pass. Partner workflows like transcription and captioning guidance are worth reviewing if your delivery needs accessible transcripts alongside the finished audio.

When hiring a pro beats managing the technical chain yourself

Tight deadlines, platform-compliant masters, and unfamiliar delivery specs are the three situations where outsourcing saves more time than it costs. An audiobook that needs to hit NLS loudness and noise floor requirements, or a broadcast spot that must land within -24 LUFS and -1 dBTP, leaves little room for trial and error.

This is where working with vetted voice actors through VoiceBros removes the technical guesswork. We connect clients with over 1,500 vetted voice actors across commercial, audiobook, e-learning, and narration categories, each put through an extensive vetting process before they ever take a booking.

What actually matters most in a home studio

Room treatment and noise control deliver more audible improvement than almost any microphone upgrade, a point borne out by how much engineers emphasize noise floor over gear price. Record at 24-bit for headroom, set your gain so peaks sit well below 0 dBFS, and measure loudness throughout your mix rather than guessing at the end.

If you remember one rule: fix your room before you fix your gear.

— Onur

Consider hiring vetted professionals instead of managing specs yourself

Getting sample rate, loudness, and true peak right takes practice, and not every project has time for a learning curve. When a deadline is tight or a platform's delivery spec is unclear, working with vetted talent through VoiceBros hands off that technical burden entirely.

• 1,500+ vetted voice actors across commercial, audiobook, e-learning, and narration categories.
• Talent matching and delivery handled together, so the final file arrives already formatted correctly.
• A practical fit for tight turnarounds or any project where offloading technical QA saves more time than it costs.

Compare rates and talent directly on our pricing page and see which voice fits your next project.

FAQ

Should I record in 44.1 or 48kHz?

Record at 48kHz for video, e-learning, and most commercial voice over work, since that rate matches standard video production formats. Use 44.1kHz only for legacy audiobook delivery, where the NLS specification sets that as the baseline with 16-bit as a minimum.

What dB should voice over be?

Aim for recording peaks between -12 dBFS and -6 dBFS during your loudest lines, which preserves headroom while keeping your signal well above the noise floor, per standard gain staging guidance. For final delivery, target -14 to -16 LUFS for podcasts and streaming or -23 to -24 LUFS for broadcast, with true peak capped at -1 dBTP.

How much to charge for a 30 second voiceover?

Pricing depends on the voice actor's experience, usage rights, and project category, so there is no single published rate. On marketplaces like VoiceBros, pricing is set per talent and project, with transparent terms shown before you book.

How many GB is 1 hour of audio?

A mono WAV file recorded at 48kHz / 24-bit runs roughly 0.5 GB per hour. Stereo files roughly double that figure, so plan storage accordingly for longer projects like audiobooks or multi-episode series.

Sources

• Loudness Normalisation and Permitted Maximum Level of Audio Signals (EBU R128)
• Audiobook mastering — Library of Congress (NLS) specification
• Basic gain staging overview (Berklee Online help)
• Audio delivery standards for film and television: A complete format guide (Tools for Film)