Output Type

Pick the output first. I2S digital (INMP441, SPH0645) gives the cleanest audio but needs an I2S peripheral. Analog audio (MAX4466, MAX9814) works with any ADC pin. Digital threshold (KY-037, LM393) only says “sound above the set level”.

Gain Control

Fixed gain clips on loud sounds or misses quiet ones. A trimmer pot (MAX4466) lets you tune for one environment, while automatic gain control (MAX9814) adapts on its own when levels change.

Supply Voltage

MEMS I2S mics are 3.3V-only and can be damaged by 5V. Electret amp boards run from about 2.4–5.5V. On an Arduino, power analog mics from the quieter 3.3V rail for less noise.

Noise & Frequency Range

Signal-to-noise ratio (SNR) decides how clean quiet speech sounds. Voice sits roughly between 100 Hz and 8 kHz, so a 50–60 Hz to 15 kHz MEMS mic covers speech recognition and most audio-reactive work.

Board Compatibility

Arduino Uno/Nano: analog or threshold modules. ESP32, RP2040, STM32, Raspberry Pi: anything, and I2S is the best match. Check for library support (ESP-IDF I2S driver, Arduino I2S, CircuitPython audiobusio) before you buy.

The MAX4466 is the classic electret mic amplifier, and this Weewooday listing gives you six purple GY-MAX4466 boards for roughly the price of one branded module. The MAX4466 op-amp has excellent power-supply rejection, so it is noticeably less hissy than a bare LM358 circuit, and a small trimmer on the back sets the gain anywhere from 25x to 125x. Output idles at half the supply voltage, so you can wire OUT straight to an ADC pin. It is a great fit for audio-reactive LEDs, VU meters, FFT spectrum analysers and simple voice changers where the sound level is fairly predictable.

👉 Check Price on Amazon →

Prefer a branded board with proper documentation? Adafruit’s SPH0645LM4H breakout is a purely digital I2S MEMS microphone with a frequency range of roughly 50 Hz to 15 kHz, a SEL pin for left/right channel selection, and a detailed Learn guide covering Raspberry Pi, Feather M0/M4 and CircuitPython. It is ideal for Pi-based voice projects and SAMD or nRF52 boards. One honest caveat: the SPH0645 has known I2S timing quirks on the ESP32 and needs its samples shifted in software, which is why most ESP32 users choose the INMP441 instead.

👉 Check Price on Amazon →

The SparkFun Sound Detector solves the problem most hobbyists actually have: “how loud is it, and did something just happen?” It gives you three outputs at once. AUDIO is the raw amplified signal, ENVELOPE is a smoothed analog voltage that tracks loudness (read it slowly with analogRead(), no sampling or FFT needed), and GATE is a digital pin that goes high when sound is present. That makes clap switches, noise loggers, classroom volume meters and “is the room occupied” sensors almost trivial. Gain and gate threshold are set by soldering a through-hole resistor; SparkFun also sells a version with headers pre-soldered.

👉 Check Price on Amazon →

The KY-037 is the module in almost every Arduino sensor kit, and it is the easiest way to learn sound detection. A large electret capsule feeds an LM393 comparator: the digital DO pin flips when sound crosses the threshold you set with the blue potentiometer, and the analog AO pin gives a rough level reading. The KY-038 is the same circuit with a smaller, less sensitive mic. Use it for clap lamps, knock detectors and “make an LED blink on noise” lessons. Do not expect clean audio or meaningful FFT from the AO pin, since there is very little amplification. Listing reviews are mixed (mostly about tuning the pot), so we score it lower than the others.

👉 Check Price on Amazon →

When all you need is a yes/no sound signal on a GPIO pin, the 3-pin LM393 sound detection module is the simplest option, and this Stemedu listing gives you five. It runs from 3.3V or 5V, so it is happy on an ESP32, ESP8266 or Pico as well as an Arduino, and the OUT pin goes low (with the onboard LED lit) when sound passes the threshold set by the potentiometer. The listing itself is clear that it detects only the presence of sound, not loudness or frequency, which is exactly right. Use it for sound-activated wake-up, interrupts in low-power devices, and classroom projects where you need several identical sensors.

👉 Check Price on Amazon →

Voice assistant, TinyML or audio streaming on ESP32?

Get the INMP441: clean 24-bit I2S audio straight into the ESP32 with no ADC noise.

Arduino Uno or Nano audio project?

Use the Adafruit MAX9814 for changing sound levels, or the MAX4466 6-pack if levels are predictable.

Music-reactive LEDs or FFT spectrum display?

The MAX4466 with gain tuned for your room, or the INMP441 on an ESP32 for higher resolution.

Raspberry Pi or CircuitPython project?

The Adafruit SPH0645 has the best step-by-step guides for Pi and Adafruit boards.

Noise logger or loudness meter?

The SparkFun Sound Detector‘s envelope output gives a loudness voltage with no DSP code.

Clap switch or simple sound trigger?

A KY-037 for analog plus digital output, or the LM393 3-pin 5-pack for cheap digital interrupts.

What is the difference between a sound sensor and a microphone module?

A sound sensor such as the KY-037, KY-038 or LM393 board uses a comparator to answer one question: is the sound louder than the threshold I set? Its digital output is a simple high or low. A microphone module such as the MAX4466, MAX9814 or INMP441 delivers the actual audio waveform, either as an analog voltage or as digital I2S samples, so you can record it, measure its level properly, run an FFT or feed it to speech recognition. If you only need a trigger, a sound sensor is enough; for anything involving the sound itself, buy a microphone module.

MAX4466 or MAX9814: which should I buy?

Both are electret microphone amplifiers with an analog output. The MAX4466 has a fixed gain you set with a trimmer (25x to 125x), which is ideal when the sound level is predictable, such as a mic placed in front of a speaker for music-reactive LEDs. The MAX9814 adds automatic gain control, so it copes with both a whisper across the room and a shout up close without clipping. Choose the MAX9814 for voice and general-purpose recording, and the MAX4466 when you want consistent, unprocessed levels for FFT or VU meters.

Can I use the INMP441 with an Arduino Uno?

No. The INMP441 outputs I2S digital audio, and the ATmega328P in the Uno and Nano has no I2S peripheral and nowhere near enough RAM or speed for audio streams. Use it with an ESP32, ESP32-S3, Raspberry Pi Pico (RP2040/RP2350 via PIO), STM32, SAMD21/SAMD51 or a Raspberry Pi. On a typical ESP32 wiring, SCK goes to GPIO14, WS to GPIO15, SD to GPIO32 and L/R to GND, powered from 3.3V. If you are stuck with an Uno, choose the MAX9814 or MAX4466 instead. See our ESP32 development board guide if you need a board.

Can these modules measure decibels (dB)?

Only approximately. You can compute an RMS level from a MAX4466, MAX9814 or INMP441 and convert it to a relative dB value, and the SparkFun Sound Detector’s envelope pin gives a ready-made loudness voltage. To get real dB SPL or dB(A) numbers you must calibrate against a proper sound level meter, and the MAX9814’s automatic gain makes absolute readings unreliable. The KY-037’s analog output is too crude for meaningful dB values. Our ESP32 sound pollution monitor project shows the approach and its limits.

Why does my sound sensor output stay HIGH or never trigger?

On KY-037, KY-038 and LM393 boards the threshold potentiometer is almost always the cause. Power the module, watch the second (signal) LED and turn the pot slowly until the LED just goes off in a quiet room; then a clap should flash it. These pots are multi-turn on some boards, so it can take many turns. Also check that you are reading the DO pin with digitalRead() and AO with analogRead(), that the grounds are shared, and that a noisy USB supply is not keeping the comparator tripped. Our KY-038 sound sensor tutorial walks through calibration with code.