Sound Spectrum Analyzer

Choose an analysis method below.

--
-- Hz
Dominant Frequency
FROZEN

View & Settings

View:
Weighting:
Smoothing:
FFT Size:

Signal Stats

Peak
-- dB
RMS
-- dB
Crest
--
Correlation
--
Clipping
Low / Bass
20 – 250 Hz
-- dB
Mid / Vocal
250 Hz – 4 kHz
-- dB
High / Treble
4 kHz – 20 kHz
-- dB

Test Signal Generator

Need a dedicated tool? Try our Online Tone Generator.

Disclaimer: This tool provides a visualization of sound frequencies. It is not intended for professional acoustic analysis. A and C weighting are biquad approximations of IEC 61672 suitable for relative measurement and visualization, not a substitute for a calibrated SPL meter.

How to Use the Sound Spectrum Analyzer

This tool helps you visualize sound by breaking it down into its individual frequencies:

  1. Start Analysis: Choose "Start Mic Analysis" for live microphone input or "Upload Audio File" to analyze a file.
  2. Grant Permission: For mic analysis, allow microphone access when prompted. No audio is recorded or stored.
  3. Read the Spectrum: Frequencies run from 20 Hz (bass, left) to 20 kHz (treble, right). Height indicates loudness. Hover to see exact frequencies.
  4. Dominant Note: The tool detects the strongest frequency and displays it as a musical note — perfect for tuning or identifying hum.
  5. Spectrogram: Switch to Spectrogram view to see how frequencies change over time — great for analyzing voice patterns or intermittent noise.

Frequently Asked Questions (FAQ)

What is a sound spectrum?

A sound spectrum is a graph showing the different frequencies present in a sound and their intensities. Every sound is made up of a unique combination of frequencies. This analyzer provides a visual "fingerprint" of the sounds in your environment.

What do the frequency bands represent?

Frequencies are measured in Hertz (Hz). The human ear typically hears 20 Hz to 20 kHz:

  • Low (20–250 Hz): Bass and sub-bass. Drums, engines, low piano notes. Spikes at 50/60 Hz usually indicate electrical hum.
  • Mid (250–4,000 Hz): Most speech and instruments. Critical for clarity. Boxy mixes often have issues around 300–500 Hz.
  • High (4,000–20,000 Hz): Cymbals, sibilance, air, and detail. Gives sound its "sparkle."

How does note detection work?

The analyzer finds the highest-energy frequency bin and maps it to the nearest musical note (A4 = 440 Hz). It also shows cents sharp or flat from the ideal pitch — useful for tuning instruments or identifying hum frequencies.

Is my audio being recorded?

Absolutely not. All processing happens in your browser. Audio data is analyzed in real-time and immediately discarded. Nothing is sent to any server.

What is A/C/Z weighting?

Z is flat — raw dBFS with every frequency treated equally. A approximates how the human ear perceives loudness and is the standard weighting for environmental and workplace noise; by design it suppresses bass and treble extremes, so quiet rumble and air read lower than they do on a flat meter. C is much flatter at low frequencies and is used for peak and impact noise where bass content matters.

Why is the dominant note now more accurate?

The analyzer now uses a Harmonic Product Spectrum (HPS) to lock onto the true fundamental rather than getting fooled by louder harmonics, and it refines the bin position with sub-bin parabolic interpolation. The result is stable enough to be used as a basic tuner — pluck a low string and you'll get the right octave, not the second harmonic.

Can I analyze audio from a music app or YouTube tab?

Yes. Click Capture System Audio (Chrome, Edge, or Opera on desktop). In the share dialog, pick the tab or window and tick "Share audio" — most first-time users miss this checkbox, so the analyzer will warn you if no audio track was shared. On mobile, Safari, or Firefox, use Upload Audio File instead. When the share dialog opens, pick the Chrome Tab option (not Window or Entire Screen) — tab capture is the most reliable across Windows and Mac. Then tick the 'Share tab audio' checkbox at the bottom before clicking Share.

What are octave bands?

Sound is grouped into bands of equal proportional frequency (each octave is a doubling). 1/1 octave gives 10 broad bands; 1/3 octave gives 31 finer bands. Used in room acoustics, building noise, and noise standards.

What is Leq, L10, and L90?

Leq is the energy-equivalent average sound level over the measurement period. L10 is the level exceeded 10% of the time (loud peaks). L90 is the level exceeded 90% of the time (background noise floor). The gap between them tells you how variable the noise is.

What's the right measurement duration?

30 seconds works for steady noise (HVAC, traffic). 5 minutes is better for variable noise (offices, restaurants). 15 minutes is the standard for environmental noise surveys. Continuous mode runs until you stop.

What is the test signal generator for?

Pink noise has equal energy per octave and is the standard signal for room and speaker measurement — pair it with mic capture to see your room's frequency response. White noise has equal energy per Hz and is useful for system stress tests. A sine wave at a chosen frequency is handy for hearing checks, single-tone level calibration, or driving a speaker at one specific note. A log sweep walks from 20 Hz to 20 kHz over 10 seconds and lets you spot resonances or rolloff. Lower your speaker volume before pressing Play — high-level signals can damage tweeters and your ears.

How do I compare two measurements?

Click Save Ref while measuring scenario A — the spectrum is captured as reference A (cyan). Make your changes (move the mic, swap a speaker, tweak EQ) and click Save Ref again to capture B (magenta). Both reference traces overlay the live spectrum so you can see the difference immediately. A third Save Ref rolls the oldest off and replaces it. Use the ✕ on a chip to drop a single reference, or Clear All to drop both.

What can I do with the CSV exports?

The Export menu writes CSV files for the current spectrum, octave bands, the Leq session log, and any saved references. Open them in Excel, Google Sheets, Python, or R for further analysis. Each file includes the timestamp, sample rate, FFT size, and weighting in commented header lines so the measurement is fully reproducible. PNG screenshot is still in the same menu.

Can I generate test tones to play through speakers?

Yes — try our Online Tone Generator in another tab. Play a sine wave at any frequency between 20 Hz and 20 kHz, then watch the spectrum here pick it up. Useful for speaker testing, room mode hunting, or just hearing what different frequencies sound like.

Advanced Features

RMS, Peak, and Crest Factor

Peak is the absolute loudest sample. RMS measures average power and better represents perceived loudness. Crest Factor is Peak minus RMS — high values mean dynamic audio, low values mean heavy compression.

Mid/Side Stereo Analysis

The Mid channel is mono content (identical in L+R). The Side channel is stereo difference (what makes sound "wide"). A weak side signal means mono-compatible audio; strong side means wide stereo.

Freeze & Save

Freeze pauses the visualization for detailed analysis. Save captures a PNG screenshot — useful for before/after EQ comparisons or documentation.