Audio-Technica AT2020 Review: A $99 XLR Condenser Microphone With a 15 mm Diaphragm

The listing calls this a large-diaphragm microphone. The field two rows below gives the diaphragm as 15.2 mm, and a comparison table embedded on Audio-Technica's own product page puts the AT2020 in a 15.4 mm diaphragm class, against 24.3 mm for the two models above it in the same line. The same shop prints the identical label on the AT2035, whose capsule is in that 24.3 mm class, which is the clearest evidence that the words carry no size at all. $99.00 buys a cardioid electret condenser with an XLR output, 144 dB of headroom measured at a stated distortion figure, but nothing on the body to switch. It fits a first interface and a quiet room. But it's the wrong microphone for anyone who needs a pad, a filter or a USB socket.

15 mm is not a large diaphragm, whatever the field says

Diaphragm size decides how much air the capsule moves for a given pressure. The conventional line for "large" sits near 19 mm, and at 15.2 mm on the shop's figure, or 15.4 in the maker's own comparison, this capsule is under it. The label matters less than what it hides. Two things follow from a smaller diaphragm, and both are visible in the numbers. One is less output for the same pressure, and more self-noise. Audio-Technica's own AT2035 at 24.3 mm publishes 12 dB of self-noise against this one's 20, and costs $60 more. But neither figure makes the microphone bad at what it does. What the label does is place it beside capsules half again its size. That's a comparison it loses on every axis that scales with area.

20 dB of self-noise, on a scale the maker never names

Audio-Technica prints one line, Noise, 20 dB SPL, and no weighting scale comes with it, on this model or on the AT2035's 12 dB. The shop adds the scale on both, printing 20 dB SPL (A-Weighted) and 12 dB SPL (A-Weighted) against figures the maker left bare. And that matters because the scale is the whole comparison. 30 microphones here carry an A-weighted self-noise figure, twenty of them below this one and one exactly level with it, and the row median is 17 dB. Every one of those positions rests on the dealer's attribution rather than the maker's.

Self-noise is the hiss the microphone makes with no sound reaching it, and that's the floor under everything quiet. Twenty decibels is audible on a whispered voice recorded close, but inaudible on anything at conversational level or above.

The comparison worth having is inside the same electret group. A Shure SM81 publishes 16 dB and states both the scale it used and a second figure on another, which is what a specification looks like when the maker expects it to be checked. This one gives a bare number. Everything after that is someone else's reading of it.

144 dB at 1 kHz and 1% THD

Maximum SPL is 144 dB, and the condition is printed with it: 1 kHz, 1 per cent total harmonic distortion. That's one of the few figures in this collection that can actually be compared, because the threshold is named. Comparing needs the threshold on both sides: a Neumann TLM102 reaches the same 144 dB at half a per cent, which is the stricter of the two conditions in common use.

But the number is high for the money. A cardioid condenser that takes 144 dB will sit in front of a guitar cabinet or a snare without a pad. That's unusual at $99, and it's the reason this microphone turns up on drum kits as often as on voices. A Sennheiser MKH416 stops at 130 dB at $1,049, for reasons that have nothing to do with headroom.

There's no pad on the body. So 144 dB is also the ceiling, and above it the distortion figure stops being 1 per cent and the microphone hasn't got a way to be turned down before the preamp.

14.1 mV/Pa, which is the middle of its own class

Audio-Technica publishes −37 dB re 1 V at 1 Pa, and prints the conversion itself, 14.1 mV/Pa, which is exactly the median of the 5 electret condensers collected here that publish the figure, 14.125 mV/Pa.

Against a dynamic the difference is the whole point of buying a condenser. A Shure SM7B puts out 1.12 mV/Pa, so this microphone delivers about twenty-two decibels more signal for the same sound, and a $99 interface preamp will be working well inside itself rather than at the top of its range.

Phantom power is 48 V at 2 mA typical. That matters only when several microphones share one supply. Output impedance is absent from both sheets. The maker's specification table hasn't got such a row and neither does the listing, so a figure that decides cable runs and preamp loading simply isn't published at this price.

The row is thin, too. Five of the electret condensers here publish a sensitivity with a stated reference unit, and 14.1 mV/Pa sits in the middle of those five rather than of a broad field.

XLR only, and why the USB version is a different microphone

There's one output, a 3-pin male XLR. The AT2020USB and AT2020USB+ are separate products with their own converters, their own headphone jacks and their own prices, but nothing on this page describes them.

That distinction is worth making. The two share a model number and a body. Buying the XLR version means buying an interface as well, and the total isn't $99. Buying the USB version means the converter is fixed. It can't be upgraded around.

122 ratings at $99.00

One hundred and twenty-two owners average 4.6, and four of them rated it one star or two, which is a 3.3 per cent tail on a large sample, and the shape carries more than the average does: 94 per cent of buyers put it at four stars or five.

What the two sheets answer about the AT2020

Is the AT2020 a large-diaphragm microphone?

The listing says so, though its own diaphragm field says 15.2 mm and the maker's comparison table places it in a 15.4 mm class against 24.3 mm for the AT2035 and AT2050.

What is the AT2020's maximum SPL?

144 dB, measured at 1 kHz at 1 per cent total harmonic distortion. There is no pad, so that figure is also the ceiling.

Does the AT2020 need phantom power?

Yes: 48 V at about 2 mA. It's a condenser and won't work on an unpowered input.

How much does an Audio-Technica AT2020 cost?

$99.00 on the listing these figures come from, which makes it the second cheapest of the 10 electret condensers in our own listings, and only a $59.99 USB model undercuts it.

Where a $99 condenser makes sense

A first XLR microphone in a room that's already quiet. The headroom is real, the sensitivity is ordinary for its class, but the absence of switches means there's nothing to set wrongly.

But the room decides the rest. A condenser at 14.1 mV/Pa hears everything a dynamic would ignore, so a hard-surfaced room, a computer fan or a street outside will be in the recording. And no switch on this body removes any of it. A RODE NT1 publishes 4 dB of self-noise at $141, which is what $42 more buys on the one figure a quiet room exposes.

Against its own siblings, though, the choice is straightforward. The numbers make it. Sixty dollars more moves the diaphragm from 15.4 to 24.3 mm and the self-noise from 20 dB to 12, and it also buys the two switches this body lacks: a 10 dB pad and an 80 Hz filter at 12 dB per octave, and the pad lifts the ceiling from 144 dB to 158.

Price and specifications

Sources

1. Audio-Technica's product page for the AT2020: pattern, frequency response, open circuit sensitivity with its conversion, maximum input sound level with its distortion threshold, dynamic range, signal-to-noise ratio, phantom power draw, weight and dimensions. The table carries no output impedance row. A comparison table embedded on the same page, supplied by Audio Test Kitchen rather than by the maker, gives the diaphragm classes for the AT2020, AT2035 and AT2050.
2. B&H Photo Video listing: price, availability, rating count and star distribution, the element type, the diaphragm field, the equivalent noise level with the weighting scale the shop adds, and the mount thread.
3. Audio-Technica's product page for the AT2035, the next model in the same line: its 12 dB self-noise, again with no weighting scale, its price, the 10 dB pad and 80 Hz filter, and the 148 and 158 dB ceilings.
4. B&H Photo Video listing for the AT2035: the same Large Diaphragm label applied to a 24.3 mm capsule, and the same A-weighted tag added to the maker's bare noise figure.
5. Medians, row sizes and positions come from listings of 200 audio products collected the same way. How those are converted and compared.

Shop data on this page was captured on 21 August 2026.

What we did not do

This microphone has never been in our hands. Every figure above is Audio-Technica's or its dealer's, and what we add is the ranking: where each number sits among the microphones collected here.