Which mic pickup pattern is best for a noisy room?

Your teammate unmutes and says she can hear your dishwasher. You reach for the Blue Yeti, notice the little knob on the back is pointing at the circle symbol, and turn it to the heart-shaped one. The dishwasher gets quieter on the call. It does not go away. Then you type a reply in the chat while you talk, and the keyboard is as loud as before.

Both results are the polar pattern doing exactly what its chart says. The circle is omnidirectional, so the mic was hearing the kitchen as loudly as your mouth. The heart is cardioid, which turns down sound arriving from behind the mic and leaves alone sound arriving from the front, and your keyboard sits right in front of it.

The short answer for a noisy room: use a cardioid or supercardioid pattern, point the mic at your mouth so the loudest noise source lands behind it or in its blind spot, and move the mic closer than feels natural. The pattern buys you a few decibels. The distance buys you more. The rest of this page shows the numbers so you can decide which of the two your room needs.

If your actual question is broader, how to stop a mic picking up noise by any means (settings, software, the room itself), the background noise guide covers all of it, and this page is the deep dive on one lever only: which way the mic listens.

What a polar pattern is, and how to read the chart

A polar pattern is a map of how sensitive a microphone is to sound arriving from each direction. The spec sheet draws it as a circle seen from above. The mic sits in the middle, its front points to 0° at the top, and the line shows how loud a sound of the same strength comes out from every angle around it. The rings are usually marked in decibels: where the line crosses the -6 dB ring, sound from that angle comes out 6 dB quieter than sound from the front.

Three habits make the chart useful instead of decorative:

• Look at where the line dips to its minimum. That angle is the null, the direction the mic is closest to deaf. It is where you want your noise.
• Look at more than one curve. DPA notes that a pattern can change with frequency, which is why many charts show several lines, one per band. If your noise is a low hum from a fridge or an air conditioner, the low-frequency curve is the one that predicts what happens.
• Check whether the mic is side-address or end-address. On a side-address mic like the Yeti, 0° points out of the side of the body with the logo, not out of the top. Talk into the top of a side-address cardioid and you are at 90°, which on a cardioid costs you about 6 dB of your own voice.

Omnidirectional microphone: hears the room as loudly as you

An omnidirectional microphone is, in Shure's words, equally sensitive to sound arriving from all angles. Its chart is a plain circle. Point it anywhere and nothing changes.

That sounds like the worst possible choice for a noisy room, and if the mic is on the desk half a meter from your face, it is. But omni has real strengths that explain why it still shows up everywhere:

• No proximity effect. Directional mics boost bass as you get close; Neumann explains that the effect happens only on pressure-gradient designs, and a pure omni is a pressure design, so it has none. Your low end stays the same whether the mic is 5 cm or 30 cm away.
• Less trouble from breath and wind. DPA says directional mics are far more sensitive to wind and pop noises than omnis.
• No off-axis coloring. Sound from the side of a directional mic comes in with its tone changed; DPA calls it the curtain effect. An omni hears the side and the front alike.

This is why most clip-on mics are omni. The Hollyland Lark M2, for one, lists its polar pattern as omnidirectional. On a lavalier the mouth is typically 15 to 20 cm away and stays there however you turn your head, so the closeness does the work that a pattern would do on a desk mic. The same goes for headset booms that sit a couple of centimeters from your lips: at that distance your voice is so much louder than the room that the pattern barely matters.

Where omni loses is the classic home setup: a USB mic on its stand, 40 cm away, in a room with a fan, a street, or another person in it. Then the mic gets the room at full volume, and you are only a little louder than everything else.

What is a cardioid microphone?

A cardioid microphone is most sensitive in front, a little less at the sides, and close to deaf directly behind. The name comes from the heart-shaped chart. DPA's figures for the ideal shape: about 6 dB quieter at 90°, the null at 180°. Six decibels down means half the sound pressure, so a fan exactly at your mic's side comes through at half strength, and a fan exactly behind it comes through much weaker than that.

Most voice mics you have heard of are cardioid. Shure's SM58 and SM7B are both listed as cardioid, the SM7B with a pattern Shure describes as uniform with frequency and symmetrical about the axis. Multi-pattern USB condensers such as the Logitech Yeti and the HyperX QuadCast 2 S offer cardioid as one setting among four (along with omni, bidirectional and stereo), and cardioid is the one you want for one person talking.

Cardioid has a cost: proximity effect. Shure describes it plainly, bass response increases as the mic moves closer to the source. At 5 cm that adds weight to a thin voice; at 2 cm it can turn into boom. The fix is a small step back or the mic's bass-cut switch, not abandoning the pattern.

Supercardioid and hypercardioid: narrower, with a catch behind

A supercardioid squeezes the front pickup narrower and moves the blind spots from directly behind to the rear sides. Shure gives a front pickup angle of 115° and least sensitivity at about 125°. Hypercardioid goes further, 105° in front and least sensitive at 110°.

The catch is a small lobe of sensitivity pointing straight back. Sound On Sound puts the rear attenuation of a supercardioid at almost 12 dB and of a hypercardioid at only 6 dB. In Shure's comparison of its own SM58 and Beta 58A: the cardioid rejects best from the back, the supercardioid better from the sides.

That one sentence decides most "supercardioid or cardioid" questions at a desk:

• Noise behind the mic (the PC tower under the monitor, the fan, the TV on the wall the mic faces away from) favors cardioid.
• Noise to the sides and rear sides (a partner on the sofa off to your left, a window to your right, desk speakers flanking the screen) favors supercardioid.

You will find supercardioid on handheld stage mics like the Beta 58A and on some USB desk mics: Logitech's Yeti GX is sold as a supercardioid dynamic. Logitech also announced the Yeti 2 in September 2026, $159.99 on its store as of October 2026; launch coverage lists supercardioid and hypercardioid among its patterns, plus a sensor that adjusts gain, EQ and pattern as you move. Hypercardioid is the narrowest of the four in front, and Sound On Sound adds that it typically has a slightly stronger proximity effect and more handling and wind noise than supercardioid.

Cardioid vs omnidirectional, and the rest, in numbers

These are the textbook figures for the first-order patterns. The directivity index tells you how much less diffuse room sound (reverb, hum filling the room from everywhere) the pattern picks up compared with an omni. The distance factor tells you how much further away a mic can sit and still get the same ratio of voice to room as an omni.

Figures from Shure, DPA, Sound On Sound and Production Expert. Real mics deviate from these ideal shapes, so read your own mic's chart too.

Look at the jumps. Omni to cardioid: 4.8 dB less room. Cardioid to supercardioid: another 0.9 dB. Cardioid to hypercardioid: 1.2 dB. Almost all of the benefit arrives with the first step. If you are already on cardioid, a "tighter" mic will not transform a noisy room.

The lever that beats every pattern: distance

The inverse square law says that in open space the level of a sound drops by 6 dB every time the distance doubles. Applied to a mic, halve the distance to your mouth and your voice comes in 6 dB louder, while the noise across the room stays roughly where it was, because the room did not move.

Put the two levers side by side for a typical setup: a mic 30 cm from your mouth in a room with an air conditioner humming.

Omni,      30 cm -> baseline
Cardioid,  30 cm -> about 4.8 dB more voice over room hum
Supercard, 30 cm -> about 5.7 dB
Cardioid,  15 cm -> about 4.8 + 6.0 = 10.8 dB
Cardioid,  10 cm -> about 4.8 + 9.5 = 14.3 dB

The 9.5 dB for 30 to 10 cm is the inverse square law for a threefold change in distance. These are idealised figures for diffuse noise, and a real room will not hit them to the decimal, but the order of magnitude holds: moving the mic from a forearm away to a fist-and-a-half away is worth several times more than buying a narrower pattern.

There is a second, less obvious gain. When the mic is close, you turn the input gain down to keep your voice at the same level, and everything quieter than your voice drops with it. A dynamic cardioid that needs you within 5 to 10 cm is "good in a noisy room" partly because of its pattern and mostly because it forces this habit. That trade-off, sensitivity against reach, is the subject of the dynamic vs condenser article, and it is a separate question from pattern: there are omni dynamics and cardioid condensers.

Aim the blind spot, not just the front

Most people aim the mic at their mouth and stop there. In a noisy room, the second half of placement is turning the mic so the worst noise sits in its null. Some common desks:

Fan or PC tower behind the monitor. Put the mic between you and the monitor, aimed at you. The noise now arrives from 180°, the cardioid null. This is the easy case.

Mechanical keyboard in front of you. The keyboard is between you and the mic, so it is inside the front lobe. A cardioid at 30° off-axis loses well under 1 dB; no pattern saves you here. What works is geometry: put the mic on an arm above the keyboard, angled down at your mouth, so the keyboard ends up 90° or more off-axis. A boom arm earns its price mostly for this.

Desk speakers either side of the screen. With the mic aimed at you, the speakers sit roughly 120 to 150° off-axis. That is supercardioid territory, close to its null. Better still, wear headphones: if your speakers are in the mic's lobe at all, the far end hears themselves back, which the echo article covers.

Someone talking on the sofa beside you. Side noise. A supercardioid helps more than a cardioid, by Shure's own comparison. If the person is behind you, you are out of luck with any pattern, because behind you is in front of the mic.

Street noise through a window. Traffic rumble is low frequency and, in a reflective room, it arrives from everywhere after bouncing. Pattern helps only by the directivity index, a few dB. Closing the gap between mic and mouth helps more, and closing the window helps more than any mic.

An empty room that echoes. Reverb is the textbook diffuse sound, so pattern and distance both help exactly as the table predicts. If the problem is that you sound distant rather than noisy, the muffled mic article is the better read.

For a Yeti in particular, three things settle most complaints: cardioid mode, speaking into the side with the logo, and the mic moved within about a hand's width of your mouth. The Blue Yeti settings guide walks through the gain knob and the rest of the mic's controls.

Which pattern should you pick?

The decision, as plainly as it gets:

1. One person, desk mic, noisy room: cardioid. It is the default for a reason, and the step from omni to cardioid is the biggest one available.
2. Noise mostly from the sides, mic close to your mouth: supercardioid, if you are buying anyway. Expect about one extra decibel of diffuse rejection and better side rejection, not a different room.
3. Clip-on or headset: omni is fine, often better. The distance is already doing the job.
4. Two people, one mic: figure-8 between them, with the sides facing the noise. Figure-8 is deaf at 90°.
5. Laptop or webcam mic: you have no pattern switch. Many laptops use mic arrays instead: Apple's spec page for the MacBook Pro lists a three-mic array with directional beamforming, which steers pickup in software rather than with a capsule shape.

Avoid omni on a desk stand in any room you would describe as noisy. That is the single setting that most often explains "they can hear everything in my house".

When the pattern runs out

Three situations defeat every pattern, so spot them before you buy a new mic:

• The noise comes from where your mouth is. A keyboard under your hands, a laptop fan in front of you, a child sitting across the table facing you. All in the front lobe.
• The noise fills the room. Hum, traffic, an air conditioner in a hard-walled room. The pattern gets you its 4.8 to 6 dB of directivity and nothing more.
• The mic has to stay where it is. A lavalier on a shirt, a laptop on the kitchen table, a call taken from a hotel room.

That is where software comes in. On a Mac, Mic Mode in the menu bar offers Voice Isolation in apps that support it; Apple notes that options vary by app. Call apps have their own suppression, and the trade-offs are laid out in the noise cancelling software comparison.

TunedMic is a different kind of tool for the same gap. It adds a virtual microphone to your Mac called TunedMic; you select it in Zoom, Meet, Teams, Slack, Discord, Telegram or OBS, and it processes the sound from your real mic before the app hears it. Setup is a one-minute Sound Check: you read a short passage aloud and it takes more than 30 measurements of your voice, your mic and your room, then builds its profiles from those. It cannot change which way your mic listens; it works on what the mic captured, and it does not promise to remove every sound.

The case it is built for is the one patterns cannot fix: the mic stays put and the room stays noisy. The 38-second video on the home page shows the Sound Check and the apps it works in: Zoom, Google Meet and Discord. Your room is not in it, so judge the result on your own mic during the free trial.

The limits are firm: Mac with Apple silicon (M1 or newer), macOS 14 or later, no Windows or Intel version. It costs $19 once, with a free 7-day trial that starts at first launch, opens every feature and needs no card or account. There are no refunds, which is why the trial comes first.

What to do in the next ten minutes

Check which pattern your mic is in right now; on multi-pattern USB mics the switch is often a small knob or a button with symbols. Set it to cardioid. Turn the mic so your loudest noise source is directly behind it, then bring it to about 10 to 15 cm from your mouth and lower the gain until your voice peaks where it did before. Record ten seconds of talking with the noise running, using the steps in our call-testing guide, and listen back on headphones.

If the noise is still there after that, the pattern has done what it can. At that point, install TunedMic from the download page to start the free trial and test it on the same mic in the same room, or see the pricing page.