How do you get a radio voice on a microphone?
A radio voice is mostly mic distance, compression and delivery, not a deeper throat. The layers in order, a processing chain with real numbers, and how to get the sound live on calls and streams.
You record a thirty-second intro for your first podcast episode. MacBook open, laptop mic, you sitting back in the chair the way you sit on calls. Then, out of curiosity, you play it straight after a clip of a morning-show host. The host sounds like they are leaning into your ear: close, even, a little heavy in the low end, every word arriving at the same weight. You sound like someone talking in the next room, loud on the first word of each sentence and gone by the last.
The natural conclusion is that the host has a better voice. Usually that is the smallest part of the gap. What people call a radio voice is three things stacked on top of each other: how the microphone is used, how the signal is processed, and how the person talks. Only the third one lives in your throat. The first two are technique and settings, and most of them cost nothing.
So the answer to how to get a radio voice is to work through the layers in the order that matters: distance first, then compression and EQ, then delivery, then how to get the result live on a call or stream instead of only in an edited file. If what bothers you is not the missing broadcast polish but that your recorded voice sounds thin or strange in general, how to make your voice sound better on calls and recordings is the page for you: it covers why your own recording sounds like a stranger and the basic fixes. This one is for the reader who already sounds fine and wants to sound like a radio host.
What a radio voice is made of
Pull a broadcast voice apart and you usually find the same five ingredients. The table shows who controls each one and what it costs.
| Ingredient | What the listener hears | Who controls it | Cost |
|---|---|---|---|
| Close mic distance (proximity effect) | Warmth, intimacy, a fuller low end | Where the mic sits | Free with a directional mic |
| Compression and limiting | Every word at the same loudness, a dense, "finished" sound | Software or hardware processor | Free (Audacity, OBS) up to studio gear |
| EQ | No boom, no mud, clear consonants | Software, or switches on the mic | Free |
| A dry room | Voice without the echo of bare walls | Furniture, soft surfaces, panels | From free to moderate |
| Delivery | Pace, pauses, pitch movement, energy | You | Practice |
Notice that "a deep voice" is not on the list. Radio hosts come in every pitch. What they share is the close, even, dry sound, and that comes from the first four rows.
Layer 1: get close and let proximity effect do its job
Directional microphones have a quirk called proximity effect: the closer the sound source, the more bass they pick up. Shure's recording guide puts the zone at about a foot, roughly 30 cm, and notes that omnidirectional mics do not show the effect at all (Shure, microphone techniques). Inside that zone, your voice gains a low-end weight that no room and no distant mic will give it.
The Shure SM7B, a dynamic mic whose user guide lists radio announcing and narration among its jobs, comes with blunt placement advice in its user guide: speak directly into it from 1 to 6 inches (2.5 to 15 cm), move closer for warmer bass, move away for less (Shure SM7B user guide). That is the radio voice recipe in one sentence. A laptop mic a foot and a half away is far outside the zone, which is a large part of why the intro in the opening scene sounded like the next room.
Now the twist. The Electro-Voice RE20, a broadcast dynamic still on the maker's site and pitched there for the radio booth and the podcast desk, is built to do the opposite. Its Variable-D design holds the low end steady on and off axis, with "no bass-heavy proximity bump" in Electro-Voice's own words (Electro-Voice RE20). The classic radio mic suppresses the very effect people credit for the radio sound.
Why would broadcasters want that? Because proximity effect is a gift that is hard to hold. Lean 2 inches closer and you boom; lean back to read a note and you go thin. A host talking for three hours moves constantly, and a mic that keeps the tone steady matters more than a mic that adds bass. The lesson for you: the radio sound is not "maximum bass", it is the same closeness every second. With an ordinary directional mic you get that by fixing the distance, not by chasing the boom.
Practical rules for this layer:
- Pick a distance and keep it. About a hand's width (3 to 5 inches, 8 to 12 cm) is a sane start for most directional mics. Closer gives more warmth and more risk of boom and pops.
- Put the mic on an arm or stand, not the desk. A mic you can pull to your mouth makes the distance repeatable. If you are deciding whether you need one, see whether you need a boom arm.
- Aim slightly off your mouth. Pointing the mic at the corner of your mouth, or a little above it, keeps the bursts of air from P and B sounds from hitting the capsule head on. A pop filter or foam windscreen handles what is left.
- Use a directional mic. Proximity effect is a property of directional pickup patterns, so it works with cardioid condensers too. In a normal untreated room a dynamic cardioid is the easier choice because, used close, it hears less of the room. The full comparison is in dynamic vs condenser for a normal room.
One hardware note if the SM7B tempts you: its guide lists a sensitivity of -59 dB and says it needs more than 60 dB of preamp gain, or an inline booster (Shure). Shure now also sells the SM7dB, the same design with a preamp built in (SM7dB guide). If you already own a low-output dynamic, whether you need a Cloudlifter or FetHead covers the gain problem.
Layer 2: compression, the real "radio" in radio voice
If proximity effect gives the voice weight, compression gives it the even, dense quality that makes it sound broadcast. A compressor turns down the parts of the signal that go above a threshold, so the gap between your loudest and quietest syllables shrinks. Then make-up gain lifts the whole thing back up. The Audacity manual describes the main purpose plainly: reducing dynamic range so the audio can be amplified further without clipping (Audacity Compressor). The result is that the end of your sentence is as loud as the start.
Radio stations take this much further than any podcaster should. A station's output passes through a dedicated broadcast processor before it reaches the transmitter, and Orban's Optimod line is one family built for exactly that job on AM, FM and digital radio (Orban). The Optimod-FM 8700 offers five-band and two-band processing (Orban timeline), which means the signal is split into several frequency ranges and each one is compressed on its own. That processing is why FM voices come across dense and loud even through a cheap car speaker.
You do not need that, and for a podcast you do not want it. Here is a sane chain for spoken voice, in order, with starting values:
1. High-pass filter 80 Hz removes desk thumps and rumble, keeps proximity warmth
2. Noise gate/suppress only if needed, keep it gentle
3. EQ -2 to -4 dB around 200-400 Hz if the voice booms
+2 to +3 dB around 3-5 kHz for presence, if it sounds dull
4. Compressor ratio 3:1 to 4:1
threshold: set so normal speech gets 3-6 dB of gain reduction
attack 5-10 ms, release 60-150 ms
5. De-esser around 5-8 kHz, only if S sounds get sharp after compression
6. Limiter ceiling -1 dB
7. Finished podcast loudness-normalize to about -16 LUFS
These numbers are starting points, not laws: every voice and mic lands differently, so adjust by ear with headphones on. The last line comes from Apple, which recommends podcast audio around -16 dB LKFS (LKFS and LUFS are the same unit) with a tolerance of 1 dB and true peaks no higher than -1 dBFS (Apple Podcasts for Creators).
Where to do it for free on a Mac:
- Audacity. The Compressor ships with voice presets named Fat, Power and Make it Right, and it can run in real time on a track while you adjust it. Loudness Normalization handles the final -16 LUFS step (Audacity manual).
- OBS. The Compressor filter's defaults are 10:1 ratio, -18 dB threshold, 6 ms attack and 60 ms release (OBS knowledge base). OBS's own page calls 2:1 weak and 6:1 much stronger, so a 10:1 default is heavy for talking. Drop the ratio to 3:1 or 4:1 as a first move. The order of filters in OBS matters too, and which OBS mic filters to use, and in what order has a full chain.
What compression costs you
Compression has a price that many tutorials skip. By lifting everything quiet, it also lifts the room, the fan of the laptop, and your breaths. The Audacity manual notes that compressing the louder parts of a vocal makes breathing and other close-up sounds very audible. In a bare room with a mic at arm's length, heavy compression produces a louder room, not a radio voice.
That is why the layers come in this order. Close distance raises your voice relative to the room before any processing touches it. A drier room (a rug, curtains, a bookshelf behind you, or the options in whether acoustic panels make your voice sound better) gives the compressor less mess to amplify. Only then does compression sound like polish rather than noise.
Layer 3: EQ, and why you cannot EQ your way to a deeper voice
When people ask how to sound like a radio host, they often mean "deeper". Here the honest answer disappoints: EQ changes the balance of frequencies in your voice, not its pitch. Boosting 100 Hz makes a voice heavier, but a high voice with boosted bass is still a high voice, just a muddier one. Lowering the actual pitch takes a pitch shifter, which is a voice changer: it changes who you sound like, which is the opposite of what a host wants.
What does create the impression of depth:
- Proximity effect, from Layer 1. It adds real low-end energy from your own voice.
- Cutting mud, not adding bass. A small dip around 200-400 Hz often makes a voice sound both clearer and fuller at once, because it removes the boxy part that masks the low end.
- Presence. A gentle lift around 3-5 kHz lets the consonants through, so the voice sounds close even at moderate volume. The SM7B has this built in as a switch: its user guide describes a bass rolloff switch to reduce the bass and a presence boost switch for a brighter midrange that "can help improve vocal clarity".
And one warning that is about your health, not your sound. Pushing your speaking voice lower than it naturally sits, for an hour of recording, strains it. The US National Institute on Deafness and Other Communication Disorders advises supporting the voice with breath rather than the throat and avoiding vocal extremes (NIDCD). Speak at your natural pitch and let mic distance and processing add the weight.
Layer 4: delivery, the part no plug-in does
Processing can make a voice even and close. It cannot make it interesting. Put a perfectly compressed voice reading a script in a flat monotone next to an average mic with a host who means every sentence, and people will listen to the second one. These habits are what separates the two:
- Talk to one person. Radio hosts address "you", not "everyone listening". Picture one specific listener and your pitch starts to move the way it does in conversation.
- Shorter sentences, longer pauses. A pause before the important word does more than any EQ. Compression makes pauses more noticeable, which works in your favor if they are deliberate.
- Land the end of statements. Let the pitch fall at the end of a statement instead of rising as if it were a question. Rising endings sound unsure, and on a compressed signal they stand out.
- Smile when the content allows. A team at the University of Portsmouth (Drahota and colleagues, 2008, in Speech Communication) found that listeners can tell different kinds of smile apart from the voice alone (study abstract). The smile travels through the microphone.
- Breathe low and keep water nearby. The same NIDCD page recommends staying hydrated, and a dry mouth produces clicks that compression will happily turn up.
- Hold your position. Every lean back undoes Layer 1. Mark the spot, or rest a hand on the desk to keep your head where it was.
A five-minute warm-up before recording beats any plug-in for the first sentence: hum on an "mmm" for a minute, read one paragraph aloud slowly with exaggerated consonants, then read it again at your normal speed.
Radio voice vs podcast voice vs call voice
The three get lumped together, but they are tuned for different listening.
Radio is built for a noisy kitchen or a car on the highway. Dense, loud, heavily processed, often multiband. Copying it fully for a podcast produces fatigue: an hour of every syllable at maximum density tires the ear.
A podcast voice is the realistic target for most people searching "radio voice". Close mic, moderate compression, -16 LUFS for Apple Podcasts, room kept quiet. It sounds broadcast-quality without sounding like a commercial break.
A call voice (Zoom, Google Meet, Teams, Discord) mostly needs consistency and clarity. The call app runs its own processing on top of yours, so a heavy compressor on your side can fight with its noise suppression and automatic gain. Moderate settings travel best. For the app side, see the best Zoom audio settings for a clear voice.
Getting the radio sound live, not only in an edit
Everything above is easy in a recording: record clean, process afterwards in Audacity, done. The harder case is live. OBS filters only affect what OBS outputs, so your Zoom call or Discord voice channel hears the raw mic. To make Zoom hear a processed voice, the processing has to happen before the call app, inside a virtual microphone that the app selects as its input. What a virtual microphone is, and why your Mac would need one explains the idea; Audio Hijack or Loopback covers the route where you build the chain yourself.
This is where TunedMic fits, so here is exactly what it does. TunedMic adds a virtual microphone to your Mac. You read a short passage aloud for about a minute (Sound Check), it takes more than 30 measurements of your voice, your mic and your room, and builds nine profiles from them. One of the Creator profiles, meant for streams, videos and podcasts, is called Broadcast. It is built from your own Sound Check rather than a stock preset, and like every profile it keeps you sounding like you: it does not shift your pitch or turn you into someone else. Pick TunedMic as the mic in Zoom, Meet, Discord or OBS, and switching profile changes the sound in all of them at once, with less than a tenth of a second of added delay. If you want to steer it, there are optional controls for noise, room echo, clarity, loudness and a 12-band EQ, but you never have to touch them.
What TunedMic does not do matters just as much:
- It does not fix distance. A mic at arm's length gets no proximity effect, and no software adds back closeness that was never recorded.
- It does not fix delivery. Its Voice Training section has short exercises for breathing, diction and intonation, where you record a take and play it back with a checklist of what to listen for, but the practice is yours.
- It needs a Mac with Apple silicon (M1 or newer) and macOS 14 or later. No Windows, no Intel Macs.
- If you only record and edit, Audacity's free compressor and loudness normalization cover you, and you do not need TunedMic.
Your 20-minute radio-voice session
Do this once, with headphones on, and you will know which layer your gap was in.
- Minutes 0 to 5. Mic on an arm or stand, about a hand's width from your mouth, aimed at the corner of your lips. Record the same thirty-second intro you started with.
- Minutes 5 to 10. Play it next to the original laptop-mic take. The difference you hear is Layer 1, and it cost nothing.
- Minutes 10 to 15. In Audacity, apply a high-pass at 80 Hz, then the Compressor at 3:1, then Loudness Normalization to -16 LUFS. Listen for breaths and room noise that got louder: that is the room telling you what to soften next.
- Minutes 15 to 20. Record the intro again, talking to one person, pausing before the key word, ending statements low. Process it the same way.
If the processed take is what you want and you need it live on calls and streams too, watch the 38-second video on the TunedMic home page first: it shows the Sound Check that sets the app up, then the tuned voice in Zoom, Google Meet and Discord. Then download it and try the Broadcast profile on your own voice during the free 7-day trial, which starts at first launch and needs no card or account. After the trial it is $19 once, and the pricing page has the details. If you would rather tune the chain yourself, how to make your mic sound better without buying a new one is the free route, setting by setting.