> ## Content Index
> Fetch the complete content index at: https://soundforgepro.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# How to Use the Spectral Analysis Tool in Sound Forge Pro
- URL: https://soundforgepro.com/how-to-use-spectral-analysis-tool-sound-forge-pro/
- Published: 2026-04-13T13:50:42.000Z
- Updated: 2026-09-20T20:59:42.000Z
- Author: Erick Finn
- Tags: Sound Forge Restoration Guides

Sound Forge calls its spectral analysis window **Spectrum Analysis**. Open it from **View → Spectrum Analysis**. The spectrum graph shows frequency on the horizontal axis and level in dB on the vertical axis. The Sonogram adds time, so you can see when a hum, click, resonance, or burst of hiss occurs. Both views measure audio; neither edits it. RX calls the same view a spectrogram; see [how to read the RX spectrogram](https://izotoperx.net/how-to-read-izotope-rx-spectrogram/).

*Part of the [Sound Forge restoration hub](https://soundforgepro.com/tag/sound-forge-restoration-guides/). For the editor, pricing, and current Boris FX lineup, start with the [complete Sound Forge guide](https://soundforgepro.com/).*

Quick answer: select the passage, open **View → Spectrum Analysis**, choose **Normal Display**, and click **Refresh**. In Settings, start with FFT 4096 or 8192 and one slice. Store a baseline with **Set → 1**. For a static before-and-after comparison, apply the processing to a working copy and refresh the same region. An EQ that is only playing through the Plug-In Chain needs output monitoring instead; refreshing the unprocessed source is not an EQ check.

## Spectrum Analysis, Spectroscope, and WaveColor serve different jobs

Sound Forge Pro has three related frequency displays. Their names overlap, but their jobs don't.

**Spectrum Analysis** is the detailed FFT window. It can analyze a selection or the audio after the cursor, show precise frequency and amplitude values, switch to a Sonogram, monitor a signal, and store four comparison snapshots. Use it when a decision needs evidence you can reproduce.

Spectroscope divides the signal into [frequency ranges](https://soundforgepro.com/audio-frequency-chart/) and displays the level in each band. The [Spectroscope help](https://cdn.borisfx.com/borisfx/Documentation/soundforge/2026/en/content/spectroscope.htm) also documents appearance presets. Use that band-level view for an overview; use Spectrum Analysis for the four-slot comparison workflow below.

[**WaveColor**](https://soundforgepro.com/how-to-use-wavecolor-in-sound-forge-pro/) colors the waveform itself. According to the [current WaveColor help](https://cdn.borisfx.com/borisfx/Documentation/soundforge/2026/en/content/wavecoloring.htm), hue represents pitch range and saturation indicates how tonal or noise-like the signal is. It helps you locate suspicious regions before opening the analyzer, though it won't give you a numeric frequency readout.

A practical sequence: WaveColor to locate, Spectrum Analysis to measure and compare, then listening to decide. Keep Spectroscope visible when a continuous band-level view is more useful than a stored before-and-after graph.

## A six-step Spectrum Analysis workflow

The analyzer becomes useful when every comparison is repeatable. Change one variable at a time and keep the source region and analysis settings fixed.

1. **Select a representative region.** Choose the passage containing the peak, hum, hiss, or tonal balance you need to inspect. Keep its start and end points fixed. Preserve an untouched original before processing a working copy.
2. **Open Spectrum Analysis and refresh.** Choose View → Spectrum Analysis, select Normal Display, and click Refresh for a static reading. Auto Refresh follows selection changes; it is not the switch for monitoring playback.
3. **Choose the FFT settings.** Click Settings. Start at FFT 4096 or 8192 for general diagnosis and set Slices displayed to 1 for snapshots. Keep the FFT overlap and smoothing window fixed; Blackman-Harris is a low-leakage option, not a universal default.
4. **Read the axes and cursor value.** Read frequency on the horizontal axis and level in dB on the vertical axis. Hover over a repeatable feature for its frequency and level, then confirm what it sounds like by listening.
5. **Store the baseline with Set and a slot.** Click Set, then snapshot 1, 2, 3, or 4\. Keep the region, FFT size, overlap, window, channel, and display range unchanged. Snapshots store curves, not an audio backup.
6. **Measure the processed copy and listen.** Apply one EQ or repair change to the working copy, then refresh the same region and compare its graph with the baseline. If the effect is only in the playback chain, use Monitor Output in the spectrum graph and start playback from the Plug-In Chain instead. Level-match processed and original audio before deciding whether the change helps.

## 1\. Select the passage before choosing settings

Select a passage that contains the issue without surrounding material hiding it. A sustained hum needs enough time to reveal its tone and harmonics; for a click, begin around the event and include some audio on either side. For mastering, inspect a representative loud or dense passage, then quieter sections separately. Keep an untouched original and make processing changes on a working copy.

Open **View → Spectrum Analysis**. Click **Refresh** for a fixed reading. Enable **Auto Refresh** when you want the graph to follow a changing selection. The [current Spectrum Graph documentation](https://cdn.borisfx.com/borisfx/Documentation/soundforge/2026/en/content/proonly/viewing%5Fa%5Fspectrum%5Fgraph.htm) states that when no selection exists, Sound Forge analyzes samples after the cursor. A deliberate selection is safer for before-and-after work because it gives you the same source every time.

## 2\. Read the graph before chasing a peak

![Spectrum graph axes with peak, noise floor, and cursor reading checks](https://storage.ghost.io/c/af/4b/af4b05d4-2e0b-4edc-8378-1e2a62fecc70/content/images/2026/08/spectrum-reading-desktop-v1.webp)

Read the graph in order: keep the region consistent, inspect peaks and the floor, then use the cursor for an exact frequency and level.

The horizontal axis is frequency and the vertical axis is amplitude in dB. A tall peak means the selected passage contains more energy around that frequency; it doesn't prove the peak is a problem. Move the cursor over it for the frequency and level readout, then loop the same passage and listen. If the tooltip is missing, right-click the graph and enable Show Position. Enable Show Notes for the nearest musical note; Show Statistics toggles the bottom readout. These displays report the cursor position, not necessarily the source’s fundamental pitch.

Treat the low, uneven line between prominent peaks as context rather than a target curve. A speech recording, a cymbal hit, and a mastered song should not share the same spectral shape. Compare the same source under the same settings. If a supposed improvement appears only after changing the selection, display range, or FFT size, the comparison is inconclusive.

## 3\. Choose FFT size for the question

Click **Settings** in the Spectrum Analysis window to choose FFT size, overlap, smoothing window, and slices. FFT size is the number of samples in each analysis window. Bin spacing is sample rate ÷ FFT size; window duration is FFT size ÷ sample rate. A larger FFT gives closer frequency bins but covers a longer stretch of time and needs more processing. Closely spaced bins do not guarantee that two tones become distinguishable: window shape, leakage, and relative levels matter too. The [Spectrum Settings documentation](https://cdn.borisfx.com/borisfx/Documentation/soundforge/2026/en/content/proonly/change%5Fspectrum%5Fanalysis%5Fsettings.htm) explains these controls.

| FFT size | Approx. bin spacing at 44.1 kHz | Practical use                                      |
| -------- | ------------------------------- | -------------------------------------------------- |
| 2048     | 21.5 Hz                         | Fast refresh and broad movement                    |
| 4096     | 10.8 Hz                         | General diagnosis                                  |
| 8192     | 5.4 Hz                          | Closer resonances and harmonics                    |
| 16384    | 2.7 Hz                          | Narrow features when slower response is acceptable |

![FFT size and smoothing window tradeoffs at a 44.1 kHz sample rate](https://storage.ghost.io/c/af/4b/af4b05d4-2e0b-4edc-8378-1e2a62fecc70/content/images/2026/08/fft-window-tradeoff-desktop-v1.webp)

At 44.1 kHz, a larger FFT narrows the frequency bins but uses a longer time window. Window choice changes leakage and peak shape.

Start with 4096 or 8192 for general EQ and noise diagnosis. Try 2048 when refresh speed matters more than narrow-bin detail, or 16384 when a narrow feature needs a closer look. These are starting points, not diagnostic presets. At 44.1 kHz, an 8192-sample window spans about 186 ms, so it is a poor way to pinpoint the timing of a very short click. Keep **FFT overlap** unchanged for comparisons; raising it asks the analyzer to perform more overlapping FFT calculations.

The smoothing window changes leakage and peak shape. Boris FX describes Blackman-Harris as the lowest-sideband-leakage option with rounded peaks. Rectangle keeps a sharp central peak but produces high leakage, and Hamming and Hanning are common audio windows. Blackman-Harris is useful when leakage would hide a nearby component; it isn't a universal best choice.

## 4\. Set the display without changing the audio

Use the arrow beside **Normal Display** to choose Line Graph, Filled Graph, or Bar Graph. The same data can look different as a line, a filled shape, or bars. If real-time refresh feels sluggish, Boris FX recommends trying Line Graph first.

Right-click the graph for **Logarithmic** frequency display: lower octaves get more space than on a linear axis, where equal distances represent equal differences in hertz. The graph extends up to the Nyquist frequency, half the file’s sample rate—22.05 kHz for 44.1 kHz audio. **Normalize dB** changes the vertical view to fit the displayed data; it does not normalize the audio file. For before-and-after comparisons, keep the amplitude range fixed so a rescaled graph does not look like a level change.

**Settings → Slices displayed** controls how many successive FFT slices the graph shows. More than one slice disables snapshots. Use multiple slices to inspect movement; return to **1** and **Normal Display** when you need stored overlays.

## 5\. Compare before and after with Set and slots 1–4

In **Normal Display** with **Slices displayed = 1**, click **Set**, then a numbered slot from **1–4**. For a static EQ comparison, process a working copy, select the same passage, and click Refresh again. Leave the baseline curve visible over the new graph. Do not just adjust a playback-only effect and refresh the unprocessed source: that measures the source again. The monitoring section below covers the live chain.

![Sound Forge Spectrum Analysis snapshot workflow using Set and slots 1 through 4](https://storage.ghost.io/c/af/4b/af4b05d4-2e0b-4edc-8378-1e2a62fecc70/content/images/2026/08/spectrum-snapshot-workflow-desktop-v1.webp)

Keep the audio region and analysis settings fixed, store a slot with Set, make one change, then refresh and overlay.

Keep the selection, FFT size, FFT overlap, smoothing window, channel, and display range identical. Otherwise you are comparing different measurements. Slot 1 can hold the original curve, slot 2 a restrained correction, and slot 3 another candidate. A different reference recording in slot 4 may help with broad tonal context, but its curve is not a target to copy: the performance, arrangement, and level can all differ.

Click a numbered snapshot button to show or hide its curve. **Set** followed by an occupied slot replaces that snapshot; **Clear All Snapshots** removes all four. These are measurement curves, not saved audio or an undo history. Snapshots are unavailable in Sonogram mode and when Slices displayed is greater than 1.

## 6\. Use the Sonogram when time is part of the problem

![Spectrum graph and Sonogram comparison with monitoring and snapshot limits](https://storage.ghost.io/c/af/4b/af4b05d4-2e0b-4edc-8378-1e2a62fecc70/content/images/2026/08/spectrum-vs-sonogram-desktop-v1.webp)

Use the spectrum graph for precise frequency comparison and snapshots; use the Sonogram to see when spectral events occur.

The Sonogram plots time on the X-axis, frequency on the Y-axis, and amplitude as color intensity. Steady hum can form horizontal lines, often around 50 or 60 Hz with harmonics above; a short click can form a vertical streak. Hiss is broadband rather than confined to one frequency. These are clues, not diagnoses: a sustained instrument also makes horizontal traces, and useful percussion makes short transients. Listen to the selected event before choosing a repair. The examples in [iZotope’s spectrogram guide](https://www.izotope.com/en/learn/understanding-spectrograms.html) illustrate noise patterns; its RX repair commands are not Sound Forge menu paths.

Select a passage and click **Sonogram**. The [Sonogram guide](https://cdn.borisfx.com/borisfx/Documentation/soundforge/2026/en/content/proonly/view%5Fa%5Fsonogram.htm) warns that long selections take longer to analyze and make time detail harder to see. For a blocky display, open Settings, enable **Set sonogram resolution**, and try increasing its value toward **200**, the help page’s example. This value is a count of FFT samplings, not 200 Hz and not the FFT size. More samplings improve horizontal detail at a processing cost; a larger FFT addresses frequency detail instead.

Use the menu beside Sonogram to choose Color or Black and White. In Color mode, drag the slider below the graph to change display intensity and read the dB color scale at the lower right. These are view controls; they don't process the selected audio.

There are two limits worth knowing. Snapshots don't work in Sonogram mode, and real-time output monitoring is unavailable there too. If you choose Monitor Output, the Sonogram shows a playback-position cursor rather than a live post-processing spectral update. Return to the spectrum graph when you need to watch the processed output or compare stored curves.

## Real-time monitoring: use the graph for processed output

Open the arrow beside **Real-Time Monitoring**. **Monitor Input** follows the Record device on the Audio tab of Preferences, and is available in both the spectrum graph and Sonogram. **Monitor Output** follows the Playback device; use the spectrum graph for its live frequency display. To analyze the post-processing signal, the graph help specifically says to **start playback from the Plug-In Chain**. Auto Refresh only follows changes to the selection—it does not select the processed output.

For slow live analysis, try Line Graph, a smaller FFT, or fewer displayed snapshots. For a repeatable static comparison, stop playback and measure the original and processed copies over the same fixed region. If the graph seems stuck after stopping, check **Settings → Maintain last monitored view**: it can deliberately keep the last live display visible. **Hold peaks** adds peak markers; those markers are not the current instantaneous spectrum.

## Turn a measurement into an audio decision

A spectrum analyzer can locate a repeatable feature, but it can't tell you whether that feature belongs in the recording. For a suspected resonance, loop the passage, read the frequency, make a small EQ move, and level-match Bypass before deciding. The [Sound Forge equalizer workflow](https://soundforgepro.com/how-to-use-equalizer-sound-forge-pro/) covers that correction step.

For hum or broadband noise, diagnose the pattern before processing. The [hiss and hum guide](https://soundforgepro.com/remove-hiss-hum-sound-forge/) separates steady electrical tones from broadband noise, while the [background-noise guide](https://soundforgepro.com/how-to-remove-background-noise-in-sound-forge-pro/) explains capture-and-reduction workflows. For final tonal checks, use Spectrum Analysis alongside Statistics and listening in the [Sound Forge mastering workflow](https://soundforgepro.com/how-to-master-a-track-in-sound-forge-pro/).

## What Spectrum Analysis cannot do

Spectrum Analysis is a viewer; it won't let you paint out a bird chirp, isolate one voice, or redraw a harmonic. The [current ARA integration documentation](https://cdn.borisfx.com/borisfx/Documentation/soundforge/2026/en/content/proonly/ara.htm) describes integrated editing with a separately installed compatible SpectraLayers VST3 plug-in. The [current Sound Forge plan comparison](https://www.soundforge.com/sound-forge/pricing/) lists VST2, VST3, and ARA support but does not list SpectraLayers as an included Sound Forge or Sound Forge Plus component. Check the license for the separate editor before planning that workflow.

The safest division is simple: Spectrum Analysis identifies and measures, a processor changes the audio, snapshots verify the measured difference, and a level-matched listening comparison decides whether the change is useful.

## Frequently Asked Questions

### Where is the Spectral Analysis tool in Sound Forge Pro?

Choose View → Spectrum Analysis. Sound Forge labels the window Spectrum Analysis. Select a short audio region and click Refresh, or enable Auto Refresh while moving the selection.

### What FFT size should I use in Sound Forge Spectrum Analysis?

Open Settings and start with FFT 4096 or 8192 for general diagnosis. Try 2048 for faster response or 16384 for narrow features. At the same sample rate, a larger FFT gives closer frequency bins but uses a longer time window and takes more processing. Window shape and leakage still affect whether nearby tones can be separated.

### Is Blackman-Harris the best smoothing window?

No single window is best for every measurement. Blackman-Harris has low sideband leakage and rounded peaks. Rectangle produces a sharper peak with high leakage. Choose the window for the question and keep it unchanged during comparisons.

### How do I compare before and after EQ in Spectrum Analysis?

Use Normal Display with Slices displayed set to 1\. Measure a fixed region, click Set and a slot from 1–4, process a working copy, then refresh that same region with unchanged FFT size, overlap, window, channel, and display range. For an effect that only runs in the playback chain, use Monitor Output in the spectrum graph and start playback from the Plug-In Chain instead. Level-match processed and original audio when listening.

### What is the difference between the spectrum graph and the Sonogram?

The spectrum graph plots level against frequency and supports snapshots plus real-time input or output monitoring. The Sonogram adds time and shows amplitude as color intensity. It supports input monitoring, but not snapshots or real-time output analysis; Monitor Output only adds a playback-position cursor there.

### Does Sound Forge Audio Studio have Spectrum Analysis?

Boris FX currently sells Sound Forge and Sound Forge Plus rather than a separate Audio Studio tier. Legacy Audio Studio features vary by version, so check View → Spectrum Analysis in that installation instead of assuming the Pro workflow is present.

### Why is Spectrum Analysis slow or sluggish?

Switch to Line Graph, reduce FFT size, or turn off snapshots. If live monitoring still lags, use Refresh on a short fixed selection. In Sonogram mode, reduce the number of samplings or FFT size when processing time is the problem.

Last fact-checked: September 8, 2026 against the Sound Forge 2026 Spectrum Graph, Spectrum Settings, Sonogram, Spectroscope, WaveColor, ARA integration, and current plan documentation, with iZotope’s explanation of spectrogram noise patterns. Menu paths here follow the documentation; this review does not claim a hands-on test of every Sound Forge build.