A uniform gain change moves a signal’s true peak by the same number of decibels, provided nothing else processes the audio afterward. If you add 3 dB, a true peak of -5 dBTP becomes -2 dBTP. Simple enough, until a loudness target asks for more gain than the available peak headroom can tolerate.
Integrated loudness and true peak describe different parts of the same file. Loudness tells you how strong the program is over time, while true peak marks its highest reconstructed peak. You can therefore have a quiet master with very little spare peak room, or a louder one whose transient structure leaves considerably more.
The useful calculation starts before you reach for a limiter. Compare the loudness gain you want with the distance between the current true peak and the permitted ceiling. Whichever allowance is smaller becomes the practical limit for clean, uniform gain.
At that point the file could sit near -15 LUFS without changing its dynamics. Reaching -14 LUFS requires some other decision, such as limiting the peaks, changing the mix, accepting the lower loudness, or using a delivery system that handles upward normalization differently. Arithmetic can tell you where the conflict begins, but it cannot choose the least damaging fix.
This is where the difference between dBFS and dBTP measurements becomes practical rather than academic. A sample meter can make the remaining margin look safer than a true-peak meter does, especially on material with sharp reconstructed peaks. The headroom calculation is only useful when the peak value you feed into it matches the ceiling you are trying to protect.
Pure gain keeps the relationship predictable because every sample and every point on the reconstructed waveform is scaled by the same ratio. Insert a limiter, clipper, EQ, compressor, saturation stage, or sample-rate converter afterward and the old projection can stop matching the rendered result. Measure again after processing instead of carrying a convenient number forward.
Downward normalization is easier. Turning a master down by 5 dB normally lowers its true peak by the same 5 dB while also reducing integrated loudness by roughly the expected amount. Upward normalization is the awkward case because the loudness target may demand gain that pushes peaks past the allowed maximum.
One useful way to think about the file is its peak-to-loudness relationship. A track with wide separation between integrated loudness and maximum true peak has less room for upward gain than a denser track at the same measured loudness if its peaks already sit close to the ceiling. Loudness alone cannot tell you how much clean gain remains.
Measurement gating adds a smaller wrinkle. For ordinary finished music, a uniform gain move usually shifts integrated loudness by the expected amount, but BS.1770 loudness measurement includes gating. Extremely quiet material near a gate boundary can change which parts contribute to the integrated reading, so a calculator should be treated as a prediction rather than a replacement for a fresh measurement.
A -20 LUFS file at -6 dBTP aiming for -14 LUFS needs 6 dB of loudness gain. With a -1 dBTP ceiling, however, only 5 dB fits cleanly. The result can reach about -15 LUFS through uniform gain alone, and the last decibel would require a different compromise.
Run the final measurement on the rendered file, not merely on the live meter before export. Any processing after the gain stage can alter true peak, while gating or editing can shift the integrated loudness reading. The finished file is where the two constraints either coexist or collide.
Integrated loudness and true peak describe different parts of the same file. Loudness tells you how strong the program is over time, while true peak marks its highest reconstructed peak. You can therefore have a quiet master with very little spare peak room, or a louder one whose transient structure leaves considerably more.
The useful calculation starts before you reach for a limiter. Compare the loudness gain you want with the distance between the current true peak and the permitted ceiling. Whichever allowance is smaller becomes the practical limit for clean, uniform gain.
Simple gain math works until peak control intervenes
Take a master at -18 LUFS with a maximum true peak of -4 dBTP. Raising it by 4 dB would put the integrated loudness near -14 LUFS, but the true peak would also move to 0 dBTP. If your delivery ceiling is -1 dBTP, only 3 dB of untouched gain fits before the peak constraint wins.At that point the file could sit near -15 LUFS without changing its dynamics. Reaching -14 LUFS requires some other decision, such as limiting the peaks, changing the mix, accepting the lower loudness, or using a delivery system that handles upward normalization differently. Arithmetic can tell you where the conflict begins, but it cannot choose the least damaging fix.
This is where the difference between dBFS and dBTP measurements becomes practical rather than academic. A sample meter can make the remaining margin look safer than a true-peak meter does, especially on material with sharp reconstructed peaks. The headroom calculation is only useful when the peak value you feed into it matches the ceiling you are trying to protect.
Pure gain keeps the relationship predictable because every sample and every point on the reconstructed waveform is scaled by the same ratio. Insert a limiter, clipper, EQ, compressor, saturation stage, or sample-rate converter afterward and the old projection can stop matching the rendered result. Measure again after processing instead of carrying a convenient number forward.
Loudness targets do not guarantee enough peak room
A common mistake is to treat the difference between measured LUFS and target LUFS as permission to add exactly that much gain. It is only the loudness side of the calculation. Peak headroom is a separate constraint, and positive normalization can run out of it first.Downward normalization is easier. Turning a master down by 5 dB normally lowers its true peak by the same 5 dB while also reducing integrated loudness by roughly the expected amount. Upward normalization is the awkward case because the loudness target may demand gain that pushes peaks past the allowed maximum.
One useful way to think about the file is its peak-to-loudness relationship. A track with wide separation between integrated loudness and maximum true peak has less room for upward gain than a denser track at the same measured loudness if its peaks already sit close to the ceiling. Loudness alone cannot tell you how much clean gain remains.
Measurement gating adds a smaller wrinkle. For ordinary finished music, a uniform gain move usually shifts integrated loudness by the expected amount, but BS.1770 loudness measurement includes gating. Extremely quiet material near a gate boundary can change which parts contribute to the integrated reading, so a calculator should be treated as a prediction rather than a replacement for a fresh measurement.
The safest gain budget uses the finished file
The clean calculation needs four numbers. You need the current integrated loudness, current maximum true peak, desired loudness, and required true-peak ceiling. Desired gain comes from the loudness difference, while peak-safe gain comes from the distance to the ceiling. Use the lower positive value when you want to preserve the waveform without peak processing.A -20 LUFS file at -6 dBTP aiming for -14 LUFS needs 6 dB of loudness gain. With a -1 dBTP ceiling, however, only 5 dB fits cleanly. The result can reach about -15 LUFS through uniform gain alone, and the last decibel would require a different compromise.
Run the final measurement on the rendered file, not merely on the live meter before export. Any processing after the gain stage can alter true peak, while gating or editing can shift the integrated loudness reading. The finished file is where the two constraints either coexist or collide.
