Fix Audio Too Quiet

Start by measuring the actual level. Open the audio file in Audacity or any DAW and look at the waveform height. A healthy recording should visually fill the waveform window to roughly 50–80% of the available height. A recording that is too quiet will appear as a thin line with lots of empty space above and below. Check the peak level using a loudness meter. If the peak is below -20 dBFS, the recording was made with insufficient gain. If the peak is at -6 to -3 dBFS but it still sounds quiet, the problem is dynamic range — the average loudness (RMS) is much lower than the peaks, which usually indicates inconsistent performance or a very dynamic source. Identifying which issue you have determines the fix.

If the recording peak is below -20 dBFS, you need to add gain. In Audacity, use Effect > Amplify to raise the peak level to -1 dBFS. Audacity calculates the exact amplification needed automatically. In any DAW, add a Gain plugin on the channel and increase until the peaks reach -6 to -3 dBFS. If you are amplifying significantly (more than 20 dB), the noise floor will amplify along with the signal. Check whether the resulting noise floor is acceptable for your purpose. If background noise is problematic after amplification, apply a noise reduction plugin (Audacity has a built-in one; iZotope RX is the professional tool) before or after amplification. For spoken word, noise at -50 dBFS or below is acceptable. For music production, noise should be at least -60 dBFS.

Normalization is the simplest way to bring a quiet recording to a standard level. It finds the peak sample in the audio and scales the entire file up so that peak reaches your target level (usually -1 dBFS). In Audacity: Effect > Normalize. In Adobe Audition: Effects > Amplitude and Compression > Normalize. Normalization works well when the recording is consistently quiet. It does not work well when the recording has a single loud moment (like a cough or a table knock) followed by much quieter content — the loud moment becomes the peak and everything else stays quiet. In this case, clip the loud anomaly to a lower level first, then normalize.

If your audio has extreme dynamic range — very loud and very quiet moments — normalization alone will not make it consistently audible. A dynamic range compressor reduces the level of loud peaks, which allows you to raise the overall gain without clipping. Use a compressor with moderate settings: threshold at -20 dB, ratio 3:1, attack 10 ms, release 100 ms. Apply makeup gain after compression to bring the overall level up. For voice recordings specifically, the Vocal Rider plugin (Waves) or the Auto Gain plugin in many DAWs automatically rides the volume to maintain a consistent level — this is faster than manual compression for long-form content. After compression, apply normalization to maximize the output level.

If your audio sounds too quiet specifically on YouTube, Spotify, or streaming platforms but seems fine in your DAW, the platform's loudness normalization is likely at work. Platforms normalize all audio to a target LUFS level. If your audio is louder than the target, it gets turned down. But if your audio is measured at, say, -20 LUFS (very quiet), some platforms turn it up — but only to their target, not beyond. The result can be audio that is still quieter than neighboring content if the neighboring content was mastered to exactly hit the target. Check your integrated LUFS using a free LUFS meter plugin (Youlean Loudness Meter is free and excellent). Target -14 LUFS for Spotify, -16 for Apple Music, -13 to -14 for YouTube, and verify your peak level stays below -1 dBFS true peak.