dBFS vs dB SPL: What Your Audio Meter Measures
- What is the difference between dBFS and dB SPL?
- What does dBFS tell you?
- What does dB SPL tell you?
- How do the common labels compare?
- Why is there no universal dBFS-to-SPL conversion?
- Can you read this paper-only example?
- What should you write beside a meter reading?
- Does a green meter mean the sound is safe?
- Can a phone app replace the project meter?
- What takes priority over interpreting the display?
- Sources
What is the difference between dBFS and dB SPL?
dBFS describes a digital signal relative to full scale; dB SPL describes sound pressure relative to an acoustic reference. They do not share a universal conversion. A low digital peak or an unclipped file cannot establish a safe level at your ears. Keep playback low and stop if sound causes discomfort or hearing symptoms; seek professional advice. Sudden hearing loss requires immediate medical attention, not another meter check. Read the unit, reference and measurement location before interpreting a number.
The most useful question is not “How many decibels?” It is “Decibels relative to what, measured where?” This guide is a reading exercise. No test tones, speaker calibration or louder playback are needed.
What does dBFS tell you?
The letters mean decibels relative to full scale. The Federal Agencies Digitization Guidelines Initiative glossary describes this reference for digital systems with a maximum available peak level.
For ordinary fixed-point PCM audio, 0 dBFS is the full-scale peak reference. A sample-peak reading of -12 dBFS is 12 decibels below that reference; it does not mean “negative sound” in a room. PCM, or pulse-code modulation, represents audio with numerical samples.
Keep the measurement type attached. A sample peak, an average reading and a loudness measurement answer different questions. “The file measures -12” is incomplete even before anyone plugs in headphones.
There is also a format qualification. Sound Devices' explanation of 32-bit float describes floating-point files that can represent values above 0 dBFS. Do not turn the fixed-point reference into a claim that every internal processing stage clips at zero. None of that extra numerical range establishes hearing safety.
What does dB SPL tell you?
SPL means sound pressure level. For sound in air, the usual reference pressure is 20 micropascals, written 20 μPa or 0.00002 Pa. This is a reference, not zero physical pressure and not a promise that every person hears the same threshold.
Measurement-equipment manufacturer NI explains the pressure reference and how sound-pressure measurements depend on the receiver's position and environment. Reflections, other sounds and the source-to-receiver relationship matter.
A microphone-based measurement therefore needs context: instrument, location and settings. “At the listening chair” and “beside the loudspeaker” are different locations, even if the same song is playing.
An audio editor's project meter does not acquire that information merely because its display says dB. It would need an appropriate acoustic measurement path, not a renamed digital scale.
How do the common labels compare?
| Label | What it refers to | What it cannot establish alone |
|---|---|---|
| dBFS sample peak | Sample amplitude relative to digital full scale | Sound pressure at a listener's ears |
| dB SPL | Acoustic pressure relative to the stated pressure reference | A complete exposure assessment without settings, position and duration |
| dBA | An A-weighted acoustic level | Interchangeability with an unweighted, differently weighted or differently timed reading |
| Gain in dB | A change relative to an input or previous level | An absolute file level or acoustic level |
The NIDCD's sound-level explanation describes A-weighting as accounting for hearing's differing sensitivity across frequencies. Keep the A in your notes. Similarly, retain labels such as peak, maximum or average instead of shortening every result to “dB.”
A gain setting of 0 dB normally describes no gain change in the documented control. It is not the same statement as a file reaching 0 dBFS or an acoustic measurement of 0 dB SPL. Confirm the control's function in its manual.
Why is there no universal dBFS-to-SPL conversion?
The digital file is only one part of playback. Consider a conceptual path:
File and processing → playback controls and output hardware → speaker or headphones → listener
A meter observes a particular point along that path. It does not automatically observe every later stage.
Audacity's Playback meter documentation, checked 8 September 2026, makes a useful product-specific distinction: its playback-volume control changes what you hear without changing the project audio or exported file.
That is enough to show why a file reading alone cannot settle acoustic level. The same exported file can be played through different output systems and control settings.
Do not add an internet “calibration offset” to your file's peak and call the result an ear-level measurement. A relationship established for particular equipment and conditions is not a universal conversion. Ask a qualified audio-measurement professional about a measurement task that genuinely needs calibration.
Our EQ headroom guide addresses digital overload separately. Keeping that signal clean and managing sound exposure remain different responsibilities.
Can you read this paper-only example?
These entries are invented for this guide. They are not test results, playback targets or a prescribed listening session.
Assume one documented sample-peak meter observes the same channel and passage at the same point in a hypothetical project:
| Record | Given information | Defensible conclusion |
|---|---|---|
| A | Version A peaks at -18 dBFS | Its highest measured sample is 18 dB below full scale |
| B | Version B peaks at -12 dBFS | Its highest measured sample is 12 dB below full scale |
| C | Version B is unchanged; only a downstream listening control changes | The upstream sample-peak reading alone cannot tell you the acoustic result |
The numerical difference between B and A is:
-12 - (-18) = +6 dB
B's reported peak level is 6 dB higher. That does not establish that the complete passage sounds twice as loud, that every sample changed by the same amount, or that a listener received a particular sound pressure.
A peak comparison also does not tell you how long either passage remains near its maximum. Two files can share a maximum sample value without sharing the same sequence of samples.
For C, the responsible answer is “acoustic result not supplied.” Do not invent a number to finish the table. Identifying what is missing is the exercise.
What should you write beside a meter reading?
Use this original record card when examining documentation or an existing measurement report:
- Tool and version: the meter or software that produced the result.
- Measurement point: file, track, bus, output or acoustic location.
- Full label: dBFS sample peak, dBA average or the actual documented metric.
- Time information: passage boundaries, measurement duration or averaging setting.
- Channel and settings: which channel, weighting and meter mode were used.
- Conclusion boundary: what the reading answers and what remains unknown.
For the fictional B record, a useful note is “-12 dBFS maximum sample peak, specified channel and passage; acoustic level not measured.” The weak version is “volume 12.”
You do not need to fill a missing field with a guess. Write “not recorded” and ask the person or tool responsible for the measurement.
Does a green meter mean the sound is safe?
No. Colors describe the software's chosen display behavior.
Audacity's playback-meter page distinguishes peak, RMS and gradient styles. A color in one mode is not a universal signal-safety standard, and none of those project-meter colors certifies acoustic exposure.
Nor does a clear sample-peak indicator prove the entire playback chain is free of overload. FADGI notes that signals without samples reaching full scale can still clip during conversion because of intersample peaks. This is another reason to read the meter's actual scope, not to increase playback until you hear a problem.
For more control vocabulary, use sound basics. Keep the technical question separate from the hearing-safety decision.
Can a phone app replace the project meter?
It measures a different thing, and its limits still matter. The NIOSH Sound Level Meter app is intended for workplace-noise measurement on supported iOS devices. NIOSH lists different metrics and A, C and Z weighting options; its stated Type 2 compliance requires a calibrated external microphone.
Those are conditions for that tool, not a certification of every phone app or improvised setup. Follow its documentation, and use an occupational-noise professional for workplace assessment.
Do not treat holding a phone beside a headphone cup as a validated measurement of your ear's exposure. Ask the equipment manufacturer or a qualified measurement professional about appropriate methods instead.
What takes priority over interpreting the display?
Reduce sound exposure rather than using a favorable-looking number to justify continuing. NIDCD recommends lowering volume, moving away from loud noise and using appropriate hearing protection when needed.
Stop listening if sound hurts or you develop ringing, muffled hearing or other concerning symptoms, and seek advice from a hearing professional. Symptoms are not a calibration tool.
For a sudden loss of hearing, even in one ear, NIDCD says to seek immediate medical attention. Do not wait to see whether another EQ preset or a quiet-looking meter explains it.
The listening lab is for learning about sound. No display reading overrides those health boundaries.