Icon

Updated weekly

OW

OW

Icon

Updated weekly

OW

Output Weekly

Output Weekly

Tools, workflows and honest rankings for people who make things with AI

Tools, workflows and honest rankings for people who make things with AI

Gear

6 best microphones for AI voiceover work in 2026

Danielle Kerr

A black dynamic microphone on a boom arm against a dark background

Key takeaways

  • The Shure SM7B is the top pick for recording a voice to clone in an ordinary room. As a dynamic cardioid mic it hears less of the room than a condenser, but it needs an interface with plenty of clean gain.

  • If you want one cable and no interface, the Shure MV7+ ($299) and Rode PodMic USB ($209) are dynamic mics with both USB-C and XLR outputs.

  • Condensers like the Rode NT1 5th Gen pick up more detail and more room. They suit a treated or well-damped space, and the NT1's 32-bit float USB output makes clipping a take close to impossible.

  • ElevenLabs asks for 1 to 2 minutes of clean audio for an Instant Voice Clone and at least 30 minutes for a Professional Voice Clone, recorded between -23 dB and -18 dB RMS with a true peak of -3 dB.

  • The room and the routine count as much as the mic. Same mic, same distance, same settings, and the same space for every session keeps a clone consistent.

Which microphone is best for AI voiceover work?

For most people recording at home, a dynamic cardioid mic is the safer choice, and the Shure SM7B is the one we'd rank first. Dynamics pick up less of a bare room's echo and noise than condensers, which is what a voice-cloning model needs, because it copies the room along with the voice.

ElevenLabs' Professional Voice Cloning guide asks for good equipment "as the AI will clone everything about the audio," so reverb, hiss, and mouth clicks end up in the clone. Dubbing over AI video has the same problem, since a noisy voice track sounds wrong next to clean generated footage.

We ranked six mics on speech clarity, how well they reject an untreated room, the extra gear they need, and list price on each maker's US store in November 2026.

The six mics at a glance

Mic

Type

Connection

Polar pattern

List price

Best for

Shure SM7B

Dynamic

XLR

Cardioid

$439

Untreated rooms, with a good interface

Shure MV7+

Dynamic

USB-C and XLR

Cardioid

$299

One-cable setups with onboard processing

Rode PodMic USB

Dynamic

USB-C and XLR

Cardioid

$209

Dynamic sound on a mid budget

Rode NT1 5th Gen

Condenser

XLR and USB-C

Cardioid

$249

Treated rooms and detailed clone samples

Elgato Wave:3 MK.2

Condenser

USB-C

Cardioid

$169.99

Plug-and-play USB recording

Rode PodMic

Dynamic

XLR

Cardioid

$99

The lowest-cost route into XLR

1. Shure SM7B: best overall for untreated rooms

The SM7B is a dynamic cardioid broadcast mic with an XLR connector and a 50 Hz to 20 kHz frequency response. Shure lists it at $439.

Strengths: The cardioid pattern rejects sound from the sides and rear, so a desk, a window, or a flat wall behind you gets into the recording less than with a condenser. Shure says the mic has "improved rejection of electromagnetic hum," including from computer monitors, and an internal shock mount cuts desk noise. Bass rolloff and presence boost switches can tame a boomy room before anything is recorded. A built-in pop filter and the included A7WS windscreen deal with plosives, which helps at the close distance this mic wants.

Limits: Its output is low. Shure lists sensitivity at -59 dBV/Pa, about 27 dB below the Rode NT1's -32 dBV/Pa, so a budget interface may run near maximum gain and add hiss. The SM7dB ($549) adds a built-in preamp with +18 dB or +28 dB of boost, powered by the interface's 48V phantom power.

Best for: Recording clone samples or dubbing tracks in a spare room, an office, or anywhere without acoustic treatment, if you have or plan to buy an interface with strong gain.

2. Shure MV7+: best USB mic for messy rooms

The MV7+ is Shure's podcast mic, a dynamic cardioid with both USB-C and XLR outputs and a 50 Hz to 16 kHz response. It lists at $299.

Strengths: The USB side runs Shure's processing on the mic. Voice Isolation Technology separates your voice from background sound, the Real-time Denoiser removes ambient room noise, Auto Level Mode adjusts gain for distance and volume, and the Digital Popper Stopper removes plosives. Shure describes it as built for "recording perfect audio in less than perfect spaces." The free MOTIV Mix app controls all of it.

Limits: Processing that changes from take to take works against a consistent clone, and ElevenLabs warns that noise and artifacts get copied too. Pick your settings once, set gain by hand for clone sessions if Auto Level makes levels drift, and keep everything the same between recordings.

Best for: Creators who want a dynamic mic's room rejection without buying an interface, and who'd like the option to move to XLR later.

3. Rode PodMic USB: best value dynamic mic

The PodMic USB is a dynamic cardioid broadcast mic with USB-C and XLR outputs, priced at $209 on Rode's US store.

Strengths: The USB output records at 24-bit and 48 kHz through Rode's Revolution Preamp, which Rode calls "ultra-low-noise, high-gain," along with APHEX processing you can set up in Rode's apps. That solves the gain problem that dynamic mics usually bring. Over XLR, its -57 dBV sensitivity is a little hotter than the SM7B's.

Limits: At 900 g it's heavy, so it needs a sturdy stand or boom arm. Its USB output tops out at 24-bit and 48 kHz, with no 32-bit float option like the NT1's.

Best for: Anyone who wants the SM7B's style of room rejection for about half the price, with USB now and XLR later.

4. Rode NT1 5th Gen: best for treated rooms

The NT1 5th Gen is a large-diaphragm condenser with a 1-inch capsule, a cardioid pattern, and both XLR and USB-C outputs, at $249. It's one of the two mics ElevenLabs names in its Professional Voice Cloning guide, alongside the Audio-Technica AT2020.

Strengths: Rode lists self-noise at 4 dBA, which is very low, so quiet speech stays clean. Over USB, it records in 32-bit float, which Rode describes as an "unclippable" output, so an unexpected laugh or a loud line can be pulled back down later without distortion. A shock mount, pop filter, and both cables come in the box.

Limits: A condenser this sensitive also hears the room clearly, including reflections off bare walls and the hum of a nearby PC. Rode's APHEX processing only works at 48 kHz and 24-bit, not in 32-bit float. Over XLR, it needs 48V phantom power from your interface.

Best for: A treated booth, a closet full of clothes, or a room you've damped with blankets, where you want the most detail in a clone sample.

5. Elgato Wave:3 MK.2: best plug-and-play USB condenser

The Wave:3 MK.2 is a USB-C condenser mic with a 16 mm capsule and a cardioid pattern, priced at $169.99.

Strengths: Clipguard 2.0 runs on the mic and guards against distortion when your voice suddenly gets loud. Auto Gain sets levels for you, an LED ring shows your level as you speak, and a built-in pop filter handles plosives. It records at 24-bit and 48 kHz, has a 3.5 mm headphone jack, and needs nothing but a USB cable.

Limits: It's USB only, so there's no path to an interface later. Its frequency response runs from 70 Hz to 20 kHz, and like any condenser, it picks up more of an untreated room than the dynamics above.

Best for: Recording short Instant Voice Clone samples or quick dubbing passes at a desk, without an interface.

6. Rode PodMic: best budget XLR mic

The original PodMic is an XLR-only dynamic cardioid mic, and at $99 it's the cheapest pick here.

Strengths: It shares the broadcast shape of the USB version and doesn't need phantom power, its sensitivity matches the PodMic USB at -57 dBV, and its 50 Hz to 15 kHz response covers the range of speech.

Limits: It's XLR only, so add the cost of an interface. Like other dynamics, it needs a fair amount of preamp gain, and its 937 g weight calls for a solid stand.

Best for: A first XLR setup for clone samples on a tight budget, paired with an interface you already own or plan to buy.

What voice-cloning services ask for in a recording

ElevenLabs publishes recording guidance for both of its cloning modes, and it works as a baseline for other services too.

ElevenLabs guidance

Instant Voice Cloning

Professional Voice Cloning

Amount of audio

1 to 2 minutes, with little gain past 3 minutes

At least 30 minutes, closer to 2 to 3 hours recommended

Recording level

-23 dB to -18 dB RMS, true peak -3 dB

Same

Upload format

MP3 at 192 kbps or above

Same

Speakers

One voice, no background noise

One voice, no music or other voices

The Instant Voice Cloning docs say "How the audio was recorded is more important than the total length (total runtime) of the samples." They also ask for a consistent tone and audio quality across every sample, and note that a flat, monotone read produces a flat clone.

The Professional Voice Cloning guide goes further on gear. It says "Any microphone will work, but an XLR mic going into a dedicated audio interface would be our recommendation." It also asks for a pop filter and one speaking style throughout, and says the AI mimics long pauses and filler sounds too, so edit them out before you upload.

ElevenLabs says WAV uploads usually don't improve clone quality, so record and edit in WAV, then export the finished sample to MP3 at 192 kbps or higher.

Room and gain tips for clean voice samples

  • Mic distance: ElevenLabs suggests about two fists from the mic. Mark the spot on your desk or boom arm so every session matches.

  • Room treatment: ElevenLabs suggests a thick duvet or quilt to damp the space if you can't treat the room.

  • Noise sources: Turn off fans, move the mic away from your computer, and record when the house is quiet.

  • Gain staging: Set gain so your normal speaking voice lands in ElevenLabs' -23 dB to -18 dB RMS window with peaks below -3 dB. On a low-output dynamic like the SM7B, that may mean an interface with more gain or an inline booster.

  • Processing: Keep onboard processing the same in every session, or record clean and handle cleanup afterward with the steps in our AI voice audio cleanup guide. Go light on noise reduction, since a clone copies its artifacts.

  • Consent: Record only your own voice, or a voice whose owner has agreed. Our AI voice translators roundup covers the consent rules each dubbing service sets.

AI voiceover microphone FAQ

Is a dynamic or condenser mic better for voice cloning?

In an untreated room, a dynamic mic usually gives cleaner samples, because it picks up less echo and background noise. In a treated or well-damped room, a condenser like the Rode NT1 5th Gen records more detail.

Do I need an audio interface?

Not with the Shure MV7+, Rode PodMic USB, Rode NT1 5th Gen, or Elgato Wave:3 MK.2, which all connect over USB-C. The Shure SM7B and Rode PodMic are XLR only, so they need an interface or mixer, ideally one with plenty of clean gain.

How much audio do I need to record for a voice clone?

For an ElevenLabs Instant Voice Clone, 1 to 2 minutes of clean audio. For a Professional Voice Clone, at least 30 minutes, and ElevenLabs recommends closer to 2 to 3 hours.

Can a USB mic make a good voice clone?

Yes. ElevenLabs recommends XLR gear but says any microphone will work. A USB mic in a quiet, damped room can produce better samples than an XLR mic in an echoey one.