HERTZKIT / AUDIO PRACTICE
How to hear and adjust a compressor
Hear how compression changes a phrase’s attacks and recovery. Compare without letting loudness decide.
On this page8 sections
Hear what happens to the attack
The same performance, uncompressed and with two attack times. Follow the beginning of each note rather than overall level.
Choose a version to play.
Save prepared fileFirst listen: compare B with C. The stronger note entries help reveal the attack behaviour. Which version keeps more of the onset? Then hear A: did those attacks need controlling, or were they part of the intention?
Separate the decisions: only attack changes between B and C before loudness compensation. Do not choose C out of habit or B just because it seems more even. With your material, check whether the onset keeps the character you want and whether the tails become too exposed.
Settings and scope
Peak detection with linked stereo channels, a threshold of approximately −22.5 dBFS, 6:1 ratio and 150 ms release. B and C use 0.5 and 30 ms attack times. Integrated loudness was approximately matched after processing. This is a deliberately pronounced instrumental demonstration, without recorded speech. Another compressor or source can respond differently.
HertzKit material under CC0. Provenance, settings and permission · TXT
What do you want to control?
A compressor reduces level differences according to its detector and settings. Before enabling it, choose a goal: steady a vocal, control attacks or create a rhythmic effect. Every track does not need compression. If an entire phrase is too quiet, volume automation may solve it more deliberately.
Threshold, ratio and time
Threshold determines where compression acts; ratio describes how levels above it are reduced. In an ideal 2:1 example, an input 6 dB above threshold ends up 3 dB above it; knee, detector and timing can alter real behavior. Attack controls how quickly reduction engages and release how it recovers. Listen for preserved attacks and recovery that follows the phrase.
| Input | Output | Reduction |
|---|---|---|
| −24 dBFS | −24 dBFS | 0 dB |
| −18 dBFS | −18 dBFS | 0 dB |
| −12 dBFS | −15 dBFS | 3 dB |
| −6 dBFS | −12 dBFS | 6 dB |
Ideal example with a −18 dBFS threshold and no makeup gain. It does not model attack, release, knee or a specific compressor’s response.
A useful exercise
Use a repeatable passage. Lower threshold until you clearly hear the effect, change attack, then release. Return to gentler reduction if that suits your goal. Match output level with the original: makeup gain can make processing seem better simply by making it louder. If breaths become exaggerated, unwanted pumping appears or expression disappears, revisit the settings. Compare versions in HertzKit; the tool compares and does not apply compression.
One phrase, three versions
Choose your own take with a clear entrance and a softer ending. Process it in your editor or music software and keep the original. The aim is to hear attack and recovery separately, not to find a preset for every voice.
- A: original. Export the excerpt without compression.
- B: faster attack. Choose a threshold that produces reduction on louder passages. Keep ratio, release and output gain unchanged.
- C: slower attack. Change only the attack from B. Export from the same start with the same duration.
- Compare A with B, then A with C, at matched loudness. Listen to consonants, note entrances and the phrase ending.
If both compressed versions lose expression, you do not have to choose either: keep A or try less reduction. Louder breaths can result from the dynamic balance and gain compensation; they do not mean the microphone captured a better recording.
Then check recovery
With your chosen attack, change only the release. Listen for the level breathing between notes, the background rising in gaps or the next word arriving while reduction is still active. Pumping can be an artistic choice; if you did not intend it, revisit the threshold and the amount of compression too. Save the setting with an audible observation. Changing five controls at once makes it hard to tell which one helped.
Put the idea into practice
Compare my compression versions
Reference for further reading
You can continue with compare EQ or check stereo and mono.