The decibel (dB) is a logarithmic unit that expresses the ratio between two values of a power-related quantity, such as sound pressure, electrical voltage or acoustic intensity. In acoustics and audio, it is the standard way to state levels, gains and losses, because the range of sounds that human hearing handles spans a pressure ratio of about one million to one.
A decibel value is always relative to a reference. Without a stated reference, a number such as 60 dB only describes a ratio. Suffixes such as dB SPL, dBFS or dBu fix the reference and turn the ratio into an absolute level.
How the decibel scale is calculated
The decibel is one tenth of a bel, a unit named after Alexander Graham Bell. For power quantities, such as acoustic intensity or electrical power, the level in decibels is 10 times the base-10 logarithm of the ratio between the measured power and the reference power.
For field or root-power quantities, such as sound pressure, voltage or current, the formula uses 20 times the logarithm of the ratio. The factor differs because power is proportional to the square of these quantities, and the logarithm of a square equals twice the logarithm. Both formulas give the same decibel value for the same physical change.
The following table shows how ratios translate into decibels. These relationships are the most useful mental shortcuts for working with levels.
| Change in level | Power ratio | Pressure or voltage ratio |
|---|---|---|
| +3 dB | About 2 times | About 1.41 times |
| +6 dB | About 4 times | About 2 times |
| +10 dB | 10 times | About 3.16 times |
| +20 dB | 100 times | 10 times |
| -6 dB | About one quarter | About one half |
| 0 dB | Equal to the reference | Equal to the reference |
Because the scale is logarithmic, decibels add where ratios multiply. An amplifier with 20 dB of gain followed by a cable loss of 2 dB produces a total gain of 18 dB, which is far easier to calculate than the equivalent multiplication.
Reference levels and decibel suffixes
In practice, the decibel is almost always used with a suffix that names the reference. Mixing up these references is one of the most frequent sources of error in audio work.
| Unit | Reference value | Where it is used |
|---|---|---|
| dB SPL | 20 micropascals of sound pressure | Acoustic measurement, noise and hearing |
| dBA | dB SPL with A-weighting filter | Noise regulation and occupational exposure |
| dBu | 0.775 volts RMS | Professional analog line levels |
| dBV | 1 volt RMS | Consumer line levels and microphone sensitivity |
| dBFS | Digital full scale | Digital audio levels and meters |
| LUFS | Loudness relative to full scale | Broadcast and streaming loudness |
The reference of 20 micropascals corresponds roughly to the threshold of hearing at 1 kHz for a young listener. The detailed use of this reference belongs to the neighboring Term sound pressure level, which applies the decibel scale specifically to acoustic pressure.
Decibels and perceived loudness
A change in decibels is not the same as a change in perceived loudness. Hearing is roughly logarithmic, which is why the decibel scale is useful, but the relationship is not exact and depends on frequency and level.
| Level change | Typical perception | Practical meaning |
|---|---|---|
| 1 dB | Barely noticeable in direct comparison | Fine mixing adjustment |
| 3 dB | Clearly noticeable | Doubling of acoustic power |
| 6 dB | Distinctly louder | Halving the distance to a point source in free field |
| 10 dB | Roughly twice as loud | Ten times the acoustic power |
Human sensitivity also varies strongly with frequency. Equal-loudness contours, standardized in ISO 226, show that quiet bass and very high frequencies need more decibels to sound as loud as midrange tones. This is the reason A-weighting is applied when sound levels are measured for noise assessment.
Typical sound levels in everyday acoustics
Typical levels give the numbers a physical meaning. The values below are approximate A-weighted levels at the listener position and vary with distance, environment and source.
| Source or situation | Approximate level | Comment |
|---|---|---|
| Quiet recording studio | 20 to 30 dBA | Background noise target |
| Normal conversation at 1 m | About 60 dBA | Comfortable listening |
| Busy city traffic | About 80 dBA | Long exposure becomes a concern |
| Loud concert near the stage | 100 to 110 dBA | Hearing protection advised |
| Threshold of pain | About 120 to 130 dB | Risk of immediate damage |
Occupational rules commonly use 85 dBA over an eight-hour day as an action or limit value, with permitted exposure time halving for every 3 dB increase under many international guidelines.
Common misconceptions
The decibel is simple in principle but easy to misuse. Most errors come from forgetting the reference or applying linear thinking to a logarithmic scale.
- Adding levels directly: two equal sources of 80 dB produce about 83 dB, not 160 dB.
- Ignoring the reference: 0 dBu and 0 dBFS describe completely different things.
- Assuming 0 dB means silence: it only means the level equals the reference.
- Mixing the 10 and 20 rules: using 10 times the logarithm for voltage halves the real decibel change.
- Treating unweighted and weighted values alike: dB SPL and dBA readings of the same noise can differ substantially when low frequencies dominate.
Understanding it simply
The decibel is a way of comparing how much stronger one sound is than another. Our ears handle an enormous range, from a whisper to a jet engine. Writing those differences as plain numbers would need very long figures, so the decibel compresses them into a short and handy scale.
On this scale, every 10 decibels means ten times more sound energy. So 70 decibels is not a little more than 60. It carries ten times the energy, and most people hear it as roughly twice as loud.
Applied in practice
A sound engineer prepares a conference hall for a keynote. A meter shows a background level of about 45 dBA from ventilation. Speech intelligibility generally benefits from a speech level at least 10 to 15 dB above the background, so the engineer sets the sound system to deliver about 65 to 70 dBA at the back rows.
At the mixing desk, the microphone signal is set so that speech peaks reach about -18 dBFS in the digital console, leaving headroom for sudden laughter or applause. The engineer also checks that the front rows do not exceed comfortable levels, because those seats are closer to the loudspeakers. The result is clear speech throughout the hall without distortion or excessive loudness near the stage.