Audio Dynamics Compressor: Control Peaks Without Changing the Song's Tempo
A loud snare hit can stand several decibels above the rest of a mix. If you raise the whole track to make quiet sections audible, that hit may clip. Dynamics compression reduces gain when the signal passes a chosen threshold. It gives you room to raise the quieter material with makeup gain afterward. Upload a file above, listen to the original, then try the default settings and compare. Your file is processed locally, so the audio never needs to reach a server.
Threshold and ratio: when compression begins
Threshold is the level where compression starts. The default is -18 dBFS. A threshold of -30 dBFS catches more of a typical recording than -10 dBFS, but it may also flatten details you wanted to keep. Ratio describes how much the level above the threshold is reduced. At 4:1, an input that rises 8 dB beyond the threshold produces roughly 2 dB of output rise above it after the compressor responds. That is a simplified steady-state example. The actual result depends on the signal's timing, detector behavior, and the compressor's knee. This tool uses a fixed 6 dB knee for a smoother transition near threshold.
Start with the lowest ratio that makes an audible improvement. A 2:1 ratio can steady a gentle performance. At 10:1 or higher, the effect is much more pronounced when peaks cross the threshold. Setting the ratio to 1:1 removes most steady-state gain reduction, although other settings and export steps still remain active. For a truly unchanged reference, choose Original in the preview section. Compression does not repair distortion that is already baked into a clipped recording.
Attack and release: how the gain changes over time
Attack controls how quickly gain reduction builds after the signal crosses the threshold. This slider spans 1 to 100 milliseconds. A fast attack can pull down the initial snap of a drum. A slower attack allows more of that transient through, which may keep the percussion lively. Release is the time the compressor takes to recover after the signal falls back. This slider spans 10 to 1000 milliseconds. An overly short release can make low notes sound uneven; a very long one may keep the next phrase quiet after a loud hit. Test both settings with the actual material rather than relying on a preset number.
The browser uses a DynamicsCompressorNode, not a custom studio compressor model. The MDN DynamicsCompressorNode reference documents its threshold, knee, ratio, attack, and release controls. Browser implementations follow the Web Audio specification. The controls are real processing parameters, not estimates of how much a commercial compressor plugin would move its gain-reduction meter.
Makeup gain, peaks, and the downloaded WAV
A compressed file can sound quieter even when its peaks are better controlled. Makeup gain adds between 0 and 18 dB after compression. Try 2 dB, listen, and increase only as needed. The offline render uses the browser-decoded sample rate and channel count, which may differ from the source file's rate. If its largest sample exceeds 0.99 in linear amplitude, the tool scales the entire output down to that value before encoding 16-bit WAV. This avoids sample clipping in the WAV, but it also means a high makeup setting may end with a quieter result than you expected. That peak safeguard is not a true-peak limiter and does not promise an inter-sample ceiling.
For example, take a vocal recording with a few loud consonants. Try a -20 dBFS threshold, a 3:1 ratio, 10 ms attack, and 150 ms release. Switch between Original and Compressed at the same listening level. If the voice loses articulation, lengthen the attack or raise the threshold. If quiet words still disappear, try a little makeup gain, then inspect the exported file with the LUFS loudness meter. LUFS describes perceived program loudness more usefully than a single sample peak.
On a long podcast, listen for breathing and room tone between sentences as well as the spoken words. Heavy compression can make background noise seem louder because speech peaks are reduced while quieter passages remain. If gaps become distracting, the noise gate may help attenuate the space between phrases. Set it gently so it does not cut off the ends of words.
Dynamics compression is not file-size compression
This page changes the loudness envelope of the sound. It does not make an MP3 smaller. Output is an uncompressed 16-bit WAV, so a short MP3 import can produce a much larger download. If your goal is to reduce megabytes for email or a web page, use the audio format converter to choose an MP3 bitrate. The sample rate calculator can estimate the size of uncompressed PCM before export. Those are separate jobs from controlling peaks.
Need to adjust tone before compression? Use the five-band equalizer to cut excess low-frequency energy that might otherwise trigger gain reduction on every kick. The note frequency calculator helps identify where a recurring bass note sits. This compressor accepts browser-decodable mono or stereo audio up to 100 MB, subject to a 128 MB decoded-memory guard. If your file will not load, try a shorter recording. Keep the original file so you can return to an unprocessed mix.
