At listening levels around 80 dB, most people can detect a level change of about half a decibel, and below 40 dB they need three times that. Hold onto both numbers. They explain almost everything about why broad shelf moves are so hard to judge, and why so many of them go wrong.
A half-decibel shelf sits right on the edge of what you can hear. That is the point of working in steps that small. You are not sculpting anything, you are nudging a balance that was already close.
The trouble is that everything around the decision is louder than the decision itself. Your monitoring level, your gain staging, the order of your chain, even how long you have been listening. Get those wrong, and a half-decibel judgment is worthless.
Loudness wins that comparison every time. The louder version sounds fuller, more detailed, and more finished, regardless of whether the EQ move actually helped. This is why gain-matched bypass comparisons are not optional on a mix bus.
The fix is unglamorous. Use the output trim to pull the processed signal back until both states read the same on a loudness or RMS meter, then switch. Peak meters are no help here, because a broad shelf can move average energy without touching the peaks much.
Continuous controls make this worse in a quiet way. Hardware built for this job steps in half decibels, so you cannot dial in 0.37 dB and convince yourself it did something. A plugin will happily let you, and false precision is the enemy of a decision you can hear.
Even the people who designed the circuit did not trust ears alone. Dangerous Music's own comparison sessions paired critical listening with measurement plots and impulse captures, over four rounds, before anyone was satisfied.
The numbers are larger than most people expect. At around 35 dB SPL, a 100 Hz tone needs about 9 dB more level than a 200 Hz tone to sound equally loud. Push the level up to 120 dB SPL and that gap collapses to roughly 2 dB.
So your low shelf decision is partly a monitoring decision. Set it quietly, and you will overdo the bass, because you are compensating for hearing that is not there at that level. Set it loud, and you will underdo it.
Pick a level and calibrate to it. Somewhere between 70 and 85 dB SPL is the usual working range, and the specific number matters less than using the same one every day. Check quietly afterwards, because that is how most people will actually hear the record.
Put the same shelf after the compressor and none of that happens. You get the tonal change and nothing else. Neither position is correct, but they solve different problems, and choosing by accident is common.
If you want the boost in front without the pumping, filter the compressor's sidechain instead. That lets the low shelf shape the sound while the gain reduction keeps ignoring it.
There is a headroom question too. Broad shelves lift peaks as well as average level, so a top-end boost that measures small can still push your true peak into trouble downstream. Trim as you go rather than at the end.
One habit is worth building. Make the move, leave it, and work on something else for ten minutes before you judge it. Small tonal shifts are easiest to hear on the return, and adaptation is fast enough that a shelf you set five minutes ago has already stopped sounding like a change.
A half-decibel shelf sits right on the edge of what you can hear. That is the point of working in steps that small. You are not sculpting anything, you are nudging a balance that was already close.
The trouble is that everything around the decision is louder than the decision itself. Your monitoring level, your gain staging, the order of your chain, even how long you have been listening. Get those wrong, and a half-decibel judgment is worthless.
Level matching decides whether you hear tone or volume
A shelf that lifts everything above 8 kHz by a decibel raises the overall level of the mix. Not by a full decibel, but by enough. When you bypass it and switch back, you are comparing two different loudnesses as well as two different tonal balances.Loudness wins that comparison every time. The louder version sounds fuller, more detailed, and more finished, regardless of whether the EQ move actually helped. This is why gain-matched bypass comparisons are not optional on a mix bus.
The fix is unglamorous. Use the output trim to pull the processed signal back until both states read the same on a loudness or RMS meter, then switch. Peak meters are no help here, because a broad shelf can move average energy without touching the peaks much.
Continuous controls make this worse in a quiet way. Hardware built for this job steps in half decibels, so you cannot dial in 0.37 dB and convince yourself it did something. A plugin will happily let you, and false precision is the enemy of a decision you can hear.
Even the people who designed the circuit did not trust ears alone. Dangerous Music's own comparison sessions paired critical listening with measurement plots and impulse captures, over four rounds, before anyone was satisfied.
Your monitoring level changes the answer
Fletcher and Munson published the first equal-loudness contours in 1933, and the current standard was revised as recently as 2023. The finding has never changed. Your ears are not a flat measurement microphone, and the error is not constant.The numbers are larger than most people expect. At around 35 dB SPL, a 100 Hz tone needs about 9 dB more level than a 200 Hz tone to sound equally loud. Push the level up to 120 dB SPL and that gap collapses to roughly 2 dB.
So your low shelf decision is partly a monitoring decision. Set it quietly, and you will overdo the bass, because you are compensating for hearing that is not there at that level. Set it loud, and you will underdo it.
Pick a level and calibrate to it. Somewhere between 70 and 85 dB SPL is the usual working range, and the specific number matters less than using the same one every day. Check quietly afterwards, because that is how most people will actually hear the record.
Where the shelf sits changes what the compressor does
A shelf before your bus compressor is not the same tool as a shelf after it. Put a low shelf in front and the compressor sees that extra energy. It responds by pulling the whole mix down whenever the bass moves, so a 1 dB boost can end up costing you midrange.Put the same shelf after the compressor and none of that happens. You get the tonal change and nothing else. Neither position is correct, but they solve different problems, and choosing by accident is common.
If you want the boost in front without the pumping, filter the compressor's sidechain instead. That lets the low shelf shape the sound while the gain reduction keeps ignoring it.
There is a headroom question too. Broad shelves lift peaks as well as average level, so a top-end boost that measures small can still push your true peak into trouble downstream. Trim as you go rather than at the end.
One habit is worth building. Make the move, leave it, and work on something else for ten minutes before you judge it. Small tonal shifts are easiest to hear on the return, and adaptation is fast enough that a shelf you set five minutes ago has already stopped sounding like a change.
