SPL and the 20 microPascal reference
Sound pressure level is a ratio. It compares the measured pressure against a reference of 20 microPascals, which is roughly the smallest pressure a young adult ear detects near 1 kHz in an anechoic chamber. The number is chosen, not derived, which is worth remembering because it means zero is a convention rather than an absence.
A tenfold increase in pressure is 20 dB. A hundredfold is 40 dB. The scale is logarithmic because human hearing covers roughly 1 x 10^-5 Pa to about 100 Pa, a span of ten million to one, and a linear scale of that range is useless in practice.
| Source or event | Approximate SPL | Note |
|---|---|---|
| Threshold of hearing, 1 kHz | 0 dB SPL | 20 microPascal reference |
| Quiet home studio | 25 to 35 dB | Good working noise floor |
| Open-plan office | 50 to 60 dB | Noticeably present |
| Busy café | 70 to 80 dB | Speech becomes hard |
| Mix at full tilt | 85 to 100 dB | Fatigue territory |
| Concert front row | 105 to 115 dB | Pain after extended exposure |
Two operational details catch people out. Meters are not calibrated by default, so an uncalibrated meter is showing dBFS and calling it dB SPL, which is the origin of most frightened forum posts. And a 6 dB rise is a doubling of pressure but nowhere near a doubling of loudness, which is the entire point of the next two scales.
A-weighting and dBA
The A-weighting curve exists because the ear is far less sensitive at low frequencies than at 1 to 2 kHz. It applies a defined attenuation per frequency, roughly minus 50 dB at 31 Hz, zero through the 2 to 4 kHz region, and about minus 10 dB at 8 kHz, so a broadband SPL reading becomes an A-weighted reading that better tracks perceived loudness at moderate levels.
- A-weighted readings read lower than unweighted ones for any source with energy at the frequency extremes.
- A kick-heavy mix can be 8 to 10 dB louder in dBA terms than the same peak in dB SPL suggests.
- A-weighting is a model, so it is wrong in detail, but it is the model regulators standardised on for exposure limits.
- C-weighting is used where low frequencies matter more, such as cinema and some environmental measurements.
A-weighting tells you about how the energy is distributed, not about how a specific listener will feel a specific track. It has no knowledge of your mix, and a brilliant top end with weak low end can read lower in dBA than a muddy one with the same peak.
Phons, sones and equal loudness
The phon scale does not come from a formula at all. It comes from listening panels. A sound is assigned the number of decibels SPL that a 1 kHz tone must reach to be judged equally loud, so ten phons means equal in loudness to a 10 dB SPL tone at 1 kHz. Every other frequency has its own equal-loudness contour derived the same way.
Those contours slope heavily at the extremes. At 20 phons, a 1 kHz tone and a 50 Hz tone differ by something like 50 dB in level for equal loudness. At 80 phons the same pair is maybe 25 dB apart. In plain terms, bass and treble get closer in perceived loudness as the level rises, which is why a mix translated up often sounds better than the same mix at a whisper.
The son scale is the odd one out and is rarely useful in a studio. It is a scale of perceived loudness growth: ten sones are judged roughly twice as loud as one sone, which is a very different relationship from the pressure ratio underneath it.
Using levels honestly in a session
- Calibrate once, against a known reference, and note the trim value on the meter or interface. Without this, no SPL number you quote means anything.
- Track your mix in dBFS for workflow and LUFS for delivery, and keep the two numbers in separate columns so they never get compared to each other.
- Work at a level where you are not adjusting volume mid-session. Around 73 to 79 dB SPL is common for extended critical listening.
- Never compare two sounds at different playback levels and conclude anything about the sound.
The trade-off nobody mentions is that a meter cannot tell you whether a mix is good. It reports an average of a signal across time and frequency, which is a useful engineering constraint and a poor aesthetic judgement. Keeping the three scales distinct is what stops a level debate from turning into a physics argument.
The bottom line
SPL measures pressure against the 20 microPascal hearing reference, dBA is a weighted version of that, and phons come from equal-loudness judgements. Only SPL is a physical measurement; the other two are derived scales, and treating them as interchangeable is where most level arguments go wrong.
A/B the same mix at a matched level
432Hz MASTER changes pitch, not gain. A/B it against your own file at one fixed level and the comparison stays honest.
Frequently asked questions
Is 0 dB SPL silent?
No. Zero dB SPL is the reference pressure of 20 microPascals, roughly the threshold of hearing for a young adult at 1 kHz in a quiet room. It is a chosen datum, not an absence of sound, and it can be exceeded.
What is the difference between dB SPL and dBA?
dB SPL is unweighted and flat across the spectrum. dBA applies the A-weighting curve, which reduces the measured value at low and very high frequencies to approximate how sensitive the ear is at moderate levels.
What is a phon?
The loudness level of a sound judged equally loud to a 1 kHz tone at that many decibels SPL. Ten phons means equal to a 10 dB SPL tone at 1 kHz, and it comes entirely from listening panels, not from a meter.
Does 6 dB mean twice as loud?
No. A doubling of SPL is 6 dB, and most listeners need roughly 10 dB to judge something clearly louder under controlled conditions. Loudness grows far more slowly than pressure.