How to Read Audacity Plot Spectrum FFT Graphs

Audacity’s Plot Spectrum tool uses a Fast Fourier Transform (FFT) algorithm to break down complex audio signals into their individual frequency components. This article explains how to read the resulting graph by understanding its axes, identifying fundamental frequencies and harmonics, spotting unwanted noise or resonance, and adjusting analysis settings for greater accuracy.

Understanding the Axes

The Plot Spectrum window displays audio information across two primary axes:

• Horizontal Axis (X-axis) – Frequency (Hz): This represents pitch, measured in Hertz (Hz) or kilohertz (kHz), spanning from the lowest bass frequencies on the left (e.g., 20 Hz) to the highest treble frequencies on the right (up to 20,000 Hz or the Nyquist limit of your sample rate). Human hearing generally spans 20 Hz to 20,000 Hz.
• Vertical Axis (Y-axis) – Amplitude (dB): This represents volume or energy level, measured in decibels (dB). In digital audio, 0 dB represents the maximum possible level before clipping. Values on this axis are negative (e.g., -20 dB, -60 dB, -90 dB); higher peaks indicate louder frequency content, while lower points represent quieter frequencies or silent regions.

Identifying Key Spectral Features

• Peaks: Distinct, sharp spikes indicate concentrated acoustic energy at specific frequencies. In musical recordings, the first prominent peak at the lower end usually represents the fundamental frequency (the perceived pitch of the note), while regularly spaced subsequent peaks represent harmonics (overtones that give an instrument or voice its timbre).
• The Noise Floor: The broad, relatively flat, or chaotic low-level activity running along the bottom of the graph represents background noise, such as microphone self-noise, room ambiance, or analog hiss.
• Narrow Spikes (Interference): Isolated, tall spikes that do not relate harmonically to the main signal often indicate unwanted interference. A sharp peak at 50 Hz or 60 Hz indicates electrical mains hum, while sharp peaks above 10 kHz often indicate digital clock noise or coil whine.
• Broad Curves: A smooth, wide shape across a range of frequencies indicates broadband sound, typical of percussive transients (like a snare drum hit), unpitched vocal sounds (like "s" or "f" consonants), or white/pink noise.

Configuring Display Settings for Better Accuracy

The controls at the bottom of the Plot Spectrum window directly alter how data is rendered:

• Axis Scale (Linear vs. Logarithmic): Switch the frequency axis to Logarithmic for musical or vocal analysis. Human hearing perceives pitch logarithmically (octaves), making a logarithmic scale more natural to read. Use Linear when hunting for evenly spaced technical artifacts or specific high-frequency clock signals.
• Size (FFT Window Size): This setting controls frequency resolution. A low size (e.g., 512) gives a broad, smoothed representation of the spectrum. A high size (e.g., 8192 or 16384) provides surgical frequency detail, allowing you to distinguish closely spaced musical notes or pinpoint the exact frequency of a hum down to a single Hertz.
• Algorithm & Function: Keep the algorithm set to Spectrum. The Function dropdown selects the windowing function (such as Hanning, Hamming, or Blackman-Harris). The default Hanning window provides an optimal balance between amplitude accuracy and frequency precision for most audio tasks.

Practical Application

By moving your cursor over the graph, Audacity displays the exact frequency and amplitude at that specific position, alongside the closest musical note. If your mix sounds muddy, look for excessive energy clustered between 200 Hz and 500 Hz. If a vocal sounds harsh, inspect the 3 kHz to 6 kHz range for prominent spikes. This data allows you to apply precise equalization (EQ) cuts or notch filters to correct technical flaws in your audio.