20–125 Hz · low end
Listen for deep bass, kick weight or low rumble. Small speakers may not reproduce the lowest energy well; removing it can also thin wanted bass.
Free tools for your next audio edit
Audio spectrum analyzer. See the frequencies in a moment.
Inspect the frequency content of a short MP3 or WAV without uploading it. Choose a start time and channel to view a short FFT snapshot and its strongest displayed frequency bin.
No account neededNo audio uploadWorks in your browser
Identify and reduce steady hum Analyze an audio spectrum
Free · localChoose a recording to inspect it locally.
One 8,192-sample Hann-windowed snapshot, with no averaging. The graph clips its display to −100–0 dBFS; peaks above 0 dBFS may be reported in the text. Bins, windowing and the chosen moment affect readings. No audio is uploaded, modified or saved.
Inspect, then listen
How to use the audio spectrum analyzer
Compare several moments before treating a peak as a persistent characteristic.
- 01 / OPEN
Choose a recording
Select a mono or stereo MP3 or WAV up to 50 MiB and five minutes. The browser decodes it locally at 48 kHz and displays the first snapshot.
- 02 / SELECT
Inspect a moment and channel
Play the recording to find a useful passage. Enter its start time in seconds, select left or right if available, and choose Analyze selection. Each snapshot covers approximately 0.171 seconds.
- 03 / COMPARE
Check other windows
A steady hum may appear at similar frequencies across quiet moments. Speech, music and transients change over time. Compare channels and listen before deciding what to edit.
Reading the graph
Frequency and level are different dimensions.
The horizontal axis represents frequency on a logarithmic scale from 20 Hz to 20 kHz; bass appears on the left and higher frequencies on the right. The vertical scale shows the level of individual FFT bins relative to full scale. A 0 dBFS bin corresponds approximately to a full-scale sine exactly centered on that bin. The graph is limited to −100 through 0 dBFS; a higher peak can still be reported in the text.
The analyzer uses 8,192 samples with a Hann window and no averaging. At the 48 kHz decoding rate, bin spacing is about 5.86 Hz. A frequency between bins spreads across them, and short events or changing tones can give different results from one window to another. The strongest bin is not necessarily the fundamental pitch, the loudest instrument, or evidence of a problem.
Unlike a spectrum, a waveform shows amplitude over time. Neither view alone tells you whether a mix sounds good or identifies which sound to remove.
A starting point for listening
Audio frequency range chart
These broad, approximate labels help you locate a region. Sources and their harmonics overlap across ranges; a peak is not a diagnosis or an instruction to boost or cut.
125–500 Hz · low mids
This region can contribute warmth and body, or a sense of muddiness in some recordings. Compare the voice and instruments together before cutting.
500 Hz–2 kHz · mids
Many voices and instruments have energy here. A narrow peak can be a note or resonance; a single snapshot cannot identify its source.
2–8 kHz · upper mids
Some consonants and attacks occupy this region. It can add presence, but boosting it may also emphasize harshness or sibilance.
8–20 kHz · high end
High-frequency content can add air or cymbal detail, while hiss may also appear here. Hearing and playback equipment vary, especially at the upper end.
These boundaries are a convenient guide, not hard divisions. For more examples and context, see iZotope’s EQ reference. In the analyzer, compare the same channel at several times, then listen at similar playback levels. The four-band equalizer can test broader changes on a copy; it changes all sources in each affected range.
Put a finding to work
Check a possible hum without removing wanted sound.
Compare several quiet sections. A narrow peak near 50 or 60 Hz and at multiples can suggest steady electrical hum, but music and other sources can also occupy those frequencies. The displayed nearest bin is approximate. Use the hum guide for a listening workflow and test a narrow cut only when appropriate.
The separate hum remover applies fixed notch filters, and the equalizer changes broader bands. Both affect wanted audio in those frequencies too. This analyzer does not filter, separate sources, measure calibrated sound pressure, or provide a full-file loudness or true-peak reading.
Audio is decoded and analyzed in your browser. No recording is uploaded or saved by this tool. Browser decoding and playback support for individual MP3 and WAV files can vary.
A few useful answers
Free audio spectrum analyzer questions
How do I analyze an audio spectrum online?
Choose a mono or stereo MP3 or WAV up to 50 MiB and five minutes. The first snapshot appears after local decoding. Enter a start time with at least 0.171 seconds remaining, select a channel if stereo, and analyze again. Playback and analysis do not upload the recording.
What does the spectrum show?
It shows a Hann-windowed 8,192-sample FFT of one channel over about 0.171 seconds, decoded at 48 kHz. The horizontal axis runs logarithmically from 20 Hz to 20 kHz; the displayed level range is −100 to 0 dBFS. Bin spacing is about 5.86 Hz.
Does the strongest bin show the song's pitch?
Not necessarily. The strongest displayed bin may be a harmonic, noise, or another source. Windowing and frequencies between bins affect the level and location. A single snapshot cannot establish the key, identify an instrument or diagnose an entire file.
Can I analyze stereo channels separately?
Yes. Choose left or right and analyze again. This tool does not average the channels or show their phase relationship. Compare the same moment in both channels before drawing a conclusion.
Does the analyzer remove noise or change my file?
No. It reads a local copy, displays a snapshot and plays the original. It does not apply EQ, filter hum or export edited audio. Keep your source file; the selection is not saved when you leave.
After preparing your audio