We Tested 12 Earbuds as Work Microphones. Most Were Terrible.

Last updated: 2026-09-14

Wireless earbuds are designed for listening. Their microphones are an afterthought — a necessity for phone calls and video meetings, but rarely the focus of engineering investment. The acoustic engineering goes into the drivers, the noise cancellation algorithms, and the spatial audio processing. The microphone gets whatever space, power budget, and signal processing are left over. We wanted to quantify exactly how much this affects real-world call quality, so we tested the microphone performance of 12 popular wireless earbuds in controlled conditions and compared the results to a dedicated USB headset and a standalone desktop microphone.

How We Tested

Each earbud was tested in a sound-treated room using a calibrated reference microphone (Earthworks M30) as the baseline. A recorded speech sample was played through a reference loudspeaker at a fixed distance of 50cm from the earbud microphone, simulating the mouth-to-earbud distance during normal wear. The earbud's output was recorded at 48kHz/16-bit via its native Bluetooth codec (AAC or SBC, depending on the earbud) on both a Windows laptop and an iPhone, simulating a typical video call connection.

We then introduced background noise — 60 dB of pink noise from a speaker positioned 1 meter behind the listener — to simulate an open-plan office environment. This is the scenario where microphone quality matters most: the listener on the other end of the call is hearing your voice mixed with office noise, and the earbud's noise-reduction algorithms must separate the two.

Recordings were evaluated both objectively (frequency response measurements, signal-to-noise ratio calculations, spectral analysis) and subjectively (a blind listening panel of five participants rated each recording for clarity, naturalness, and intelligibility on a 1-10 scale).

The Results in Quiet Conditions

In the quiet room (no background noise), all 12 earbuds produced intelligible speech. The differences were in frequency response — how accurately the microphone captured the full range of the human voice (roughly 85 Hz to 8,000 Hz for speech intelligibility, with harmonics extending to 12,000 Hz for naturalness).

The three earbuds that scored highest — Apple AirPods Pro 2, Samsung Galaxy Buds2 Pro, and Jabra Elite 85t — captured speech frequencies from 100 Hz to 8,000 Hz with relatively flat response (within plus or minus 3 dB). Voices sounded natural, with clear consonants (which occupy the 2,000-8,000 Hz range) and adequate warmth from low-frequency content. The blind panel rated these recordings 7.5-8.2 out of 10, compared to 9.1 for the reference microphone.

The three earbuds that scored lowest — a $30 budget earbud, a $50 fitness-focused model, and surprisingly, a $250 premium model — had significant frequency response irregularities. The budget earbud rolled off all frequencies below 300 Hz (removing chest resonance and vocal warmth) and boosted 3,000-5,000 Hz by 6 dB (making sibilants harsh and voices thin). The $250 premium model had a narrow notch at 2,500 Hz that made certain consonants (particularly "s" and "f" sounds) nearly inaudible. The blind panel rated these recordings 4.2-5.5 out of 10.

The Results With Background Noise

Background noise is where microphone quality separates dramatically. All 12 earbuds use some form of computational noise reduction — algorithms that attempt to identify and suppress non-voice sounds while preserving the speaker's voice. The effectiveness of these algorithms varied enormously.

The AirPods Pro 2 performed best in noise, using what Apple calls "computational audio" — a combination of beamforming (using multiple microphones to focus on the voice direction), voice isolation (machine learning models trained to distinguish speech from non-speech sounds), and adaptive noise reduction (adjusting suppression intensity based on the noise environment). In our 60 dB noise test, the AirPods Pro 2 reduced background noise by approximately 25 dB while maintaining voice clarity with only minor artifacts (occasional brief dropouts when noise and voice frequencies overlapped).

The Samsung Galaxy Buds2 Pro and Jabra Elite 85t produced similar results, with 20-22 dB of noise reduction and slightly more noticeable processing artifacts — a faint "underwater" quality during heavy noise suppression that is characteristic of aggressive computational noise reduction. These artifacts were audible when listening critically but did not significantly affect intelligibility.

The remaining nine earbuds fell into two groups: four that provided 10-15 dB of noise reduction (enough to make speech intelligible in moderate noise, but with the background clearly audible), and five that provided less than 10 dB of noise reduction (effectively no useful noise suppression — the listener hears both the speaker and the full background noise). The five worst performers in noise were not all cheap — they included two models priced above $100.

Why Some Earbuds Have Bad Microphones

Three factors determine earbud microphone quality: microphone hardware (size, type, placement), beamforming capability (number and arrangement of microphones used for voice isolation), and computational processing (the algorithms that enhance voice and suppress noise).

Microphone hardware is constrained by the earbud form factor. Standalone microphones use large diaphragms (typically 15-25mm) that capture a wide frequency range with low noise. Earbud microphones use MEMS (Micro-Electro-Mechanical Systems) elements with diaphragms smaller than 1mm. The physics of small diaphragms limits sensitivity and frequency response, particularly at low frequencies where longer wavelengths require larger capture surfaces.

Beamforming requires multiple microphones. The AirPods Pro 2 uses three microphones per earbud — an outer microphone for noise detection, an inner microphone for voice capture, and a rear microphone for directional reference. This three-microphone array enables effective beamforming that isolates the voice direction. Budget earbuds typically use a single microphone per earbud, which provides no directional information and cannot perform beamforming — the microphone captures everything equally from all directions.

Computational processing depends on the earbud's onboard processor. Apple's H2 chip, Qualcomm's QCC5171, and Samsung's AKG-tuned processors include dedicated DSP (Digital Signal Processing) cores for real-time voice enhancement. Budget earbuds use general-purpose Bluetooth SoCs with minimal processing capability, leaving voice enhancement to the receiving device's software — which is inconsistent and often inadequate.

Wind noise is the final microphone challenge that separates good earbuds from bad ones. Outdoor use in even moderate wind (10-15 mph) produces turbulent airflow over the microphone ports that overwhelms voice signals. The AirPods Pro 2 and Jabra Elite 85t use mesh wind screens over their outer microphones that reduce wind noise by approximately 12-15 dB. Budget earbuds with exposed microphone ports showed wind noise levels that completely masked speech at wind speeds above 10 mph — making outdoor phone calls effectively impossible without cupping your hand over the earbud.

Our Recommendations

If you use earbuds for video calls more than once per week, microphone quality should be a primary selection criterion — not an afterthought after sound quality and noise cancellation. The AirPods Pro 2 (for Apple users), Samsung Galaxy Buds2 Pro (for Android users), and Jabra Elite 85t (platform-agnostic) are the three earbuds in our test that produce professional-quality microphone output in both quiet and noisy environments.

If your earbuds are primarily for music and you take calls infrequently, microphone quality is less important — but be aware that the person on the other end of your calls may be struggling to hear you. A $50 USB headset like the Jabra Evolve2 30 — which we tested as a reference — produced clearer microphone output than 9 of 12 earbuds, including models costing three to six times as much.

The most cost-effective solution for regular call-takers is to use earbuds for music and a dedicated headset for calls. A $60 Jabra Elite 85t (frequently on sale) provides 90% of the call quality of a $250 premium earbud and 100% of the call quality of a $50 USB headset, with the added benefit of portability and wireless convenience. For any earbud purchase, check microphone-specific reviews before buying — the overall star rating on retail sites reflects sound quality, comfort, and features, but rarely mentions microphone performance, which is the metric that matters most to the people who hear you.

Beam-Forming vs. Single-Mic: What the Hardware Actually Does

The microphone technology inside wireless earbuds falls into two broad categories, and the distinction explains most of the performance variation we measured. Single-microphone earbuds — common in models under $100 — use one omnidirectional MEMS (micro-electro-mechanical system) microphone, typically mounted on the earbud's outer shell or stem. This microphone picks up sound from all directions with roughly equal sensitivity, relying entirely on software-based noise reduction to separate the wearer's voice from ambient sound.

Beam-forming earbuds use two or more microphones per earbud (the AirPods Pro 2 uses three) with an array processing algorithm that combines signals from multiple capsules to create a directional pickup pattern. The physics is straightforward: by analyzing the arrival time difference of sound at each microphone, the processor can determine whether a sound originates near the wearer's mouth or from the environment. Mouth-directed sound is amplified; environmental sound is suppressed. This approach solves the noise problem at the hardware level rather than leaving it entirely to software.

In our measurements, beam-forming earbuds achieved an average signal-to-noise ratio of 24.3 dB in the 60 dB noise test — a 9.7 dB improvement over single-mic models (14.6 dB average). That gap is perceptually enormous: a 10 dB improvement means the listener hears the speaker's voice at roughly double the perceived loudness relative to background noise. Every earbud that scored above 7.0 in our blind panel evaluation used beam-forming arrays. No single-mic earbud scored above 5.8.

The catch is that beam-forming quality varies. A two-microphone beam-former on a short-stem earbud (where the mics are close together) has limited spatial resolution — it struggles to distinguish between voice and noise arriving from similar angles. A three-microphone array on a longer stem (like the AirPods Pro 2's design) has wider baseline spacing, giving the algorithm more precise directional information. This is why stem length, often dismissed as a purely aesthetic choice, directly predicts microphone performance in our data set.

The Bluetooth Codec Bottleneck

Even an excellent microphone is throttled by Bluetooth audio transmission. During a phone or video call, Bluetooth operates in a fundamentally different mode than during music playback. Music streaming uses A2DP (Advanced Audio Distribution Profile) at data rates up to 328 kbps for AAC or 990 kbps for LDAC. But when the microphone is active, Bluetooth switches to HSP/HFP (Headset/Hands-Free Profile), which limits audio to the mSBC codec at approximately 64 kbps — or, on older devices, CVSD at 64 kbps with even narrower bandwidth.

The practical impact is dramatic. Under A2DP, earbuds stream music at up to 20 kHz frequency response. Under HFP with mSBC, the transmitted bandwidth drops to 50 Hz–8 kHz — barely half the range of a standard telephone. The listener's voice sounds thin and lacks the high-frequency consonant detail (sibilants, fricatives) that makes speech crisp and intelligible. This is why many users report that music sounds fantastic on their earbuds but call quality is mediocre — the limiting factor is the Bluetooth profile, not the driver.

Bluetooth LE Audio, standardized in 2022 but still rolling out in consumer devices, addresses this with the LC3 codec, which supports simultaneous bidirectional audio at higher bitrates and lower latency. In our testing, one LC3-capable earbud (the Samsung Galaxy Buds3 Pro) did demonstrate measurably better call audio when connected to a Galaxy S24 — 11.2 kHz effective bandwidth versus 7.8 kHz on mSBC with the same earbud connected to an iPhone. The improvement is audible but requires both the earbud and the phone to support LE Audio, which excludes the current iPhone lineup and most Android devices as of mid-2026.

Environmental Noise Rejection: Quantifying Real-World Call Clarity

Microphone specifications like "dual beamforming array" and "AI noise suppression" sound impressive but tell you nothing about actual performance in the environments where you take calls. We tested every earbud's microphone in three standardized noise environments: office HVAC (45 dBA broadband noise), busy cafe (65 dBA with speech babble from four pre-recorded conversation tracks), and urban street (72 dBA with traffic, construction, and wind). For each environment, a test speaker recited a standardized passage at a calibrated 68 dBA level, and we recorded the earbud's transmitted audio for analysis.

We quantified call clarity using two metrics: speech-to-noise ratio (SNR, measured in dB) and Perceptual Evaluation of Speech Quality (PESQ) score, an ITU-T standard that predicts subjective speech quality on a 1.0–4.5 scale. In the office environment, all earbuds in our cohort achieved PESQ scores above 3.5 (good quality) because the noise level was low enough that even minimal processing kept speech intelligible. The differentiation emerged in the cafe and street environments.

The Apple AirPods Pro (2nd generation) achieved a PESQ score of 3.2 in the cafe environment and 2.8 on the street—the best in our cohort by 0.3 and 0.4 points respectively. Its computational audio pipeline, which uses all six microphones (three per earbud) to isolate the wearer's voice from ambient noise, was particularly effective at suppressing speech babble—the most challenging noise type because it occupies the same frequency range as the target voice. The Sony WF-1000XM5 achieved PESQ scores of 2.9 (cafe) and 2.5 (street), with its AI-based noise suppression occasionally introducing a "underwater" artifact during consonant-heavy speech that our panelists found mildly distracting.

Wind Noise Handling: The Overlooked Microphone Challenge

Wind noise is the single most disruptive environmental factor for earbud microphones, and it is the one that software processing handles least effectively. Wind generates turbulent pressure fluctuations across the microphone diaphragm that the DSP cannot distinguish from voice because both are broadband signals originating at the microphone itself. We tested wind-noise rejection using a calibrated wind tunnel that produces laminar airflow at three speeds: 8 km/h (light breeze), 16 km/h (moderate wind), and 24 km/h (strong wind).

At 8 km/h, most earbuds in our cohort maintained intelligible voice transmission, with PESQ scores dropping by only 0.3–0.5 points from their no-wind baseline. At 16 km/h, the spread widened dramatically: the AirPods Pro maintained a PESQ of 2.6 (still intelligible, though degraded), while three earbuds dropped below 2.0 (unintelligible for most listeners). At 24 km/h, no earbud in our cohort produced reliably intelligible voice audio—even the AirPods Pro dropped to a PESQ of 1.8, confirming that strong wind overwhelms current noise-suppression technology regardless of price point.

The physical design of the microphone port is the primary determinant of wind-noise performance. Earbuds with recessed microphone ports and mesh wind screens (the AirPods Pro, Samsung Galaxy Buds3 Pro) outperformed those with flush or protruding microphones by 4–6 dB of wind-noise rejection at 16 km/h. This mechanical advantage compounds with software processing—a cleaner raw signal gives the DSP more headroom to work with—which is why microphone placement and wind-screen design matter more than the sophistication of the noise-suppression algorithm.

The Video Call Context

Our testing used 48kHz/16-bit recording, but real-world video calls compress audio significantly. Zoom uses the Opus codec at approximately 48 kbps for voice. Microsoft Teams uses the Satin codec at similar bitrates. Google Meet uses Opus at 32-128 kbps depending on network conditions. These codecs are designed to preserve speech intelligibility while aggressively discarding non-speech content — which means they amplify some microphone deficiencies and mask others.

We retested the three best and three worst earbuds through actual Zoom calls (not just raw recordings) to assess how codec compression interacts with microphone quality. The good news: the three best earbuds (AirPods Pro 2, Galaxy Buds2 Pro, Jabra Elite 85t) sounded nearly identical through Zoom as in raw recordings — the codec preserved their frequency response and noise reduction effectively. The bad news: the three worst earbuds sounded even worse through Zoom than in raw recordings. The codec's compression artifacts combined with the microphone's frequency response problems to produce garbled, robotic-sounding voice that multiple listeners described as difficult to understand.

The implication is straightforward: if your earbuds produce poor microphone quality in ideal conditions, video call compression makes it worse, not better. The codec cannot compensate for hardware limitations — it can only preserve or degrade whatever the microphone captures. Testing microphone quality in raw recordings gives you the best-case scenario. Real-world video calls will be equal or worse, never better.