How does distortion work? Listen and try it

What distortion does to a signal

A linear gain change scales the waveform. A nonlinear waveshaper changes the relationship between input and output, so a simple sine can acquire harmonics: components at multiples of its original frequency. With complex audio, the result can include interactions between frequencies as well as harmonic changes. More drive usually makes the change easier to hear.

This demo uses a symmetric tanh soft-clipping curve and 4x oversampling. For a sine input, a symmetric shaper produces odd harmonics in the ideal mathematical case. Real musical signals contain many frequencies, so do not expect a clean series of isolated lines from every source. Oversampling helps reduce aliasing; it does not guarantee its complete absence.

This is an educational waveshaper, not an emulation of a particular analog device. Delay makes copies later in time and reverb creates a tail; distortion changes the signal itself. You can combine them, but their order affects what you hear.

Distortion controls compared

Drive applies 0 to 30 dB, initially 12 dB, before the nonlinear shaping. More drive pushes more of the signal into the curved part of the transfer function. Listen for extra density and changes to transients, but also watch the output level.

Tone ranges from 800 to 16,000 Hz, initially 6,000 Hz. Lower settings reduce brightness; higher settings retain more high-frequency content. Mix ranges from 0 to 100%, initially 60%, blending the dry and processed paths. A partial blend can retain more of the dry attack while adding some processed character.

Output ranges from −24 to 0 dB, initially −9 dB. Use it to adjust the processed level after changing drive. Output gain controls level; it does not undo distortion already created. Its setting is not a true-peak guarantee.

RMS matching compensates for output gain changes. Turn matching off to hear and compare the output control itself; turn it on again for a level-guided A/B comparison.

Four listening exercises

Begin with the initial settings and return each control after its comparison. Use the same source and listen at a moderate level. Recheck perceived loudness before judging the character; the source influences how obvious each change will be.

• Drive: compare 3 dB with 24 dB. Listen for added edge and a change in the shape of drum attacks or a sustained note.
• Tone: compare 800 Hz with 16,000 Hz. Focus on the brightness and harshness of the processed sound.
• Mix: compare 10% with 100%. Listen for how the dry attack and processed texture combine.
• Output: turn RMS matching off, then compare −18 dB with −6 dB. Hear the level change; output does not remove harmonics created earlier. With RMS matching on, compensation offsets this gain change.

Read the meters and waveform

The waveform can show how the processed signal differs from the source, and RMS gives a level measure over a period. A change in shape is not a quality score. Dense material can be hard to interpret visually, so use a simple source and listen as well as look.

The RMS and sample-peak readouts use a mono downmix, even when you listen to stereo audio. The sample peak is the largest sampled level in that downmix; it is not true peak. Stereo signals can combine or cancel in the downmix, so these readings do not describe each channel separately.

Compare at similar perceived loudness and use the meter as an aid. RMS does not establish a fixed LUFS value or guarantee a true-peak limit. Strong drive and a high output can produce uncomfortable levels; lower your playback volume before testing extreme settings.

Add distortion with a purpose

Select a source and decide what you want to hear: a bass that speaks on small speakers, a drum with more edge, or a voice with an intentionally rough texture. Start at low drive, raise it until the change is audible, then adjust output for a fair comparison. Check the result in the full mix.

Use tone to control brightness and mix to blend the processed signal with the original. Reduce drive if consonants become brittle or a bass loses its definition. A little can be useful; the correct amount depends on the source and musical context, not a universal number.

The three presets offer starting points. Open the expanded view to inspect the controls, or load a local mono or stereo file up to 25 MB and 120 seconds. Audio stays in the browser. This tool teaches listening and does not claim to master or export the result.

A clean source provides a useful reference, but it is not the only valid choice. Existing saturation in a recording affects how much additional processing you hear. Reset the controls before testing a different source, start quietly, and give yourself time to compare. You may prefer less processing once the other tracks are playing.

Start with a simple source

A steady note makes added harmonics easier to hear; a drum loop makes changes to attacks easier to judge. Use one source for your first comparison and set a comfortable monitor level. Raise drive slowly until you hear a change, then use output to bring the processed result closer to the original in perceived loudness.

Listen beyond the solo track

A bass with extra harmonics can be easier to follow on a small speaker, but too much processing may cover a guitar or vocal in the same range. Listen with those parts playing. If the sound works alone but crowds the arrangement, lower the mix or drive and compare again. Check the result on more than one playback system.

Try effect order intentionally

Place distortion before delay in your DAW and listen to repeats of the altered sound. Reverse the order and the distortion also processes the delayed signal; overlapping echoes can interact at the input to the nonlinear effect. Keep levels comparable when switching the order. The lab demonstrates one effect at a time, so make this comparison in a separate audio project.

Common distortion mistakes

Comparing a loud processed signal against a quieter dry signal can make extra level sound like an improvement. Match perceived level as closely as practical, then listen for the actual change in tone and attack. RMS can help, but equal readings do not ensure equal perceived loudness for every sound.

High drive can mask a vocal or make cymbals tiring. Reduce drive or mix, adjust tone, and check the result with the rest of the arrangement. A waveform that looks flatter is not automatically better or worse; decide whether the sound serves the part.

Distortion questions

Is saturation the same as distortion?

Saturation often describes a milder or more gradual form of nonlinear distortion. Product labels vary, so listen to the behavior rather than relying on the name.

Does distortion always make a sound louder?

It can change both peak and average level, but the final level also depends on drive, mix, output and the source. Set output for a useful comparison.

Does oversampling eliminate aliasing?

Oversampling can reduce aliasing from nonlinear processing. It does not guarantee that all aliasing disappears, and results depend on the implementation and source.

Should distortion come before reverb or delay?

Both orders are useful. Distortion before a time effect sends the altered source into its tail; distortion afterward also processes the tail. Compare both in your DAW at matched listening levels.