EDUKATE SINGAPORE · GOLD STANDARD PRACTICAL CASE STUDY 200
A −6 dBFS digital peak does not prove a recording sounds half as loud.
A fictional education studio measures a peak near −6 dBFS and calls the resulting track ‘half as loud’. The amplitude ratio is approximately one half, but digital peaks, signal power and perceived programme loudness answer different questions. This case shows the correct comparison and the proper end-to-end audio-quality gate.
All studios, recordings and outcomes are fictional. The quantitative example explains signal relationships, not hearing safety or a prescribed listening level.
The facts before the details
- Recorded peak: approximately −6 dBFS.
- Digital amplitude ratio: about 0.501 relative to full scale.
- Squared ratio: approximately 0.251 under relevant power assumptions.
- Not established: that listeners perceive the whole programme as half as loud.
- Needed: an appropriate loudness or playback test for the actual purpose.
Chapter navigation
- 1. The recording’s peak level becomes a claim about what listeners hear
- 2. The decibel is a logarithmic ratio, not an ordinary percent
- 3. Worked example: a six-decibel reduction is near half peak amplitude
- 4. dBFS identifies digital full scale, not sound pressure level
- 5. ITU and EBU distinguish loudness from peak metering
- 6. The peak of one short sound does not describe the whole programme
- 7. A wrong loudness claim can mislead the accessibility review
- 8. The immediate correction preserves the valid signal measurement
- 9. Build an audio acceptance process around real purposes
- 10. Practice: near-half peak amplitude is not half perceived loudness
1. The recording’s peak level becomes a claim about what listeners hear
A fictional education media studio prepares audio for a museum-style history programme. Its engineer measures the digital recording and finds the highest sample peak near −6 dBFS, using the agreed meter and format. The programme manager sees the numeral six and tells educators that the recording is ‘half as loud as a full-scale recording’. The website repeats that statement as a technical assurance.
The engineer’s level measurement can be valid while the interpretation is unsupported. A peak level relative to digital full scale is not itself a measurement of human-perceived loudness or the sound pressure at a listener’s ears. The Gold Standard asks which physical or digital property was measured, what comparison is legitimate, and what the audience is being promised.
2. The decibel is a logarithmic ratio, not an ordinary percent
Decibels express a ratio using a logarithmic scale. For a digital signal amplitude ratio, a value in decibels is twenty multiplied by the base-ten logarithm of the amplitude relative to its reference. A label such as −6 dBFS identifies the relation between a particular digital sample peak and the system’s full-scale reference.
It does not mean minus six percent or ninety-four percent. Nor does it automatically provide a direct statement about the acoustic sound pressure after a speaker, headphone or amplifier reproduces the signal. The digital measurement, electrical reproduction chain, speaker and listening environment must be distinguished. GS-R02 defines the quantity; GS-R08 requires the reported claim to stay within that quantity.
3. Worked example: a six-decibel reduction is near half peak amplitude
Use the standard amplitude relationship: amplitude ratio = 10 raised to the power dB/20. At −6 dB this gives 10^(−6/20), approximately 0.501. A reference peak amplitude of 1.0 full-scale unit would correspond to about 0.501 of that reference under the simplified sample-peak comparison.
A difference of exactly −6.0206 dB would correspond to half the amplitude. For a comparable waveform and applicable energy model, the square of a 0.501 amplitude ratio is approximately 0.251, or one quarter of the reference signal-power ratio under those assumptions. Those mathematical relations do not directly determine how loud a real programme seems to a listener.
The distinction matters because an amplitude ratio, a power ratio and a psychological loudness comparison are different outcomes. Multiplying one of them by a convenient percentage creates a fluent sentence without demonstrating the corresponding measurement.
4. dBFS identifies digital full scale, not sound pressure level
The abbreviation dBFS means decibels relative to digital full scale. Zero dBFS is a defined reference associated with the maximum digital sample-amplitude scale in the applicable representation. Values below zero describe a smaller sample peak relative to that reference. They do not say how a specific listener will experience the audio in a room.
A recording at −6 dBFS can be played with different gain settings, headphones or loudspeakers. The real acoustic output can differ enormously even if the stored digital file is unchanged. The studio should therefore not label its −6 dBFS sample peak as −6 dB SPL, a measured sound-pressure quantity that has a different reference and requires appropriate acoustic measurement.
GS-R07 is the handoff from digital editing to playback. A completed audio export does not justify a hearing-safety or perceptual claim unless the right measurement and context support it.
5. ITU and EBU distinguish loudness from peak metering
The International Telecommunication Union’s ITU-R BS.1770-5 recommendation provides algorithms for measuring audio programme loudness and true-peak level. Those are distinct descriptors used for different technical questions in sound production and distribution.
The European Broadcasting Union’s EBU R 128 loudness recommendation describes average programme loudness in LUFS alongside true-peak and loudness-range information. The EBU loudness overview explains why peak-based normalisation alone can leave programmes with different perceived loudness. These documents concern professional audio practices, not a certification of the invented museum recording.
Adobe’s Audition normalisation guidance also distinguishes digital full-scale normalisation from other volume-matching operations. The Gold Standard draws the practical lesson: choose the meter that answers the audience-facing question, rather than attach a loudness promise to a peak-sample indicator.
6. The peak of one short sound does not describe the whole programme
Suppose one fictional track contains a brief loud bell followed by several minutes of soft narration, while another contains continuous narration close to its peak level. Both could have a maximum sample peak near −6 dBFS. They would not necessarily have the same average programme loudness, listening comfort or intelligibility.
The highest sample value is a useful headroom and technical signal measure. The perceived strength of longer audio material involves time distribution, frequency content, listening context and other factors. A media centre choosing levels for an educational programme should evaluate the actual presentation conditions rather than treat two equal peaks as proof of equal listener experience.
One number can be correct within its limited scope and misleading when used to describe a different property. This is the same Gold Standard mechanism seen in scientific measurement and business reporting: the measured unit must match the intended user outcome.
7. A wrong loudness claim can mislead the accessibility review
The museum’s educational guide needs speech that listeners can understand using the intended playback system. A manager might assume that a digital peak at −6 dBFS automatically makes the audio suitable for every visitor. That would substitute a stored-signal value for real listening conditions, including interface controls, background noise and other access needs.
The centre should separate audio-file integrity, production headroom, programme loudness and appropriately evaluated playback usability. Relevant technical and accessibility specialists decide suitable acceptance criteria. A general article cannot issue one universally safe playback volume or infer hearing risk from a single stored peak number.
GS-R01 returns to the visitor’s actual purpose: receiving a clear, meaningful history programme. The correct question is whether the final medium supports that purpose under accepted conditions.
8. The immediate correction preserves the valid signal measurement
The studio need not erase its −6 dBFS report. The reading can be useful for the relevant digital-peak task. It should remove the unsupported ‘half as loud’ statement, retain the original source and exported versions, and identify which descriptions or lesson sheets carried the misleading interpretation.
A properly bounded update might say the recording’s measured sample peak was approximately −6 dBFS relative to digital full scale. The current evidence does not establish a particular ratio of perceived loudness or acoustic sound pressure. Any further claim requires suitable measurement of the relevant signal or reproduction outcome.
GS-R09 governs the exception and GS-R10 tests the repair. The correction should reach the web page, accompanying transcript or media specification wherever the original claim was reused.
9. Build an audio acceptance process around real purposes
The next studio specification distinguishes avoiding unwanted clipping in the file, achieving appropriate programme loudness for the intended distribution service, preserving dialogue intelligibility and providing useful visitor controls. Each condition may use a different method, meter or expert judgement under the actual media programme.
The engineer records which tool produced the measured digital peak and what type of peak it represents. A different analysis may examine integrated loudness or true peak under suitable standards. The content manager uses the approved wording without replacing LUFS, dBFS or dB SPL as though all measured the same quantity.
GS-R07 ensures the measurement’s meaning survives from engineering to copywriting. GS-R12 requires rechecks when exported codecs, playback arrangements or delivery platforms change.
10. Practice: near-half peak amplitude is not half perceived loudness
Exercise. A fictional audio file peaks at −6 dBFS. Calculate the amplitude ratio relative to full scale using 10^(−6/20). Explain what can and cannot be inferred about its power ratio under simple assumptions and listeners’ perceived loudness.
The result is approximately 0.501 of reference peak amplitude; its squared ratio is approximately 0.251 for a comparable signal-power model. Neither result establishes that visitors perceive the whole programme as half as loud or that a loudspeaker produces a specified sound-pressure level. The Gold Standard lesson is to name the exact measurement before promising an audience outcome.
Frequently asked questions
Does minus six dBFS mean fifty percent volume?
No. It is a logarithmic digital peak relation. The amplitude is near half of full scale, but perceived programme loudness is different.
Why is the amplitude ratio approximately 0.5?
The ratio is 10^(−6/20), about 0.501; exactly half would be roughly −6.0206 dB.
Does the squared ratio mean one quarter of perceived loudness?
No. Squaring is relevant to signal-power comparisons under defined assumptions, not directly to perceived loudness.
Is dBFS the same as dB SPL?
No. dBFS references digital full scale; dB SPL references sound pressure and requires acoustic measurement.
What is LUFS used for?
LUFS is used to describe programme loudness under a specified measurement standard, rather than merely one sample’s peak.
What should the museum publish?
A bounded technical peak claim and any separately verified programme-loudness or visitor-access information, without unsupported interpretations.
Audio-quality acceptance checks
- Identify the digital or acoustic unit and its reference.
- Distinguish sample peak from true peak and integrated programme loudness.
- Check the amplitude calculation against an accepted logarithmic definition.
- Avoid turning a peak reading into a listener-loudness claim.
- Evaluate the actual output and playback conditions for the intended users.
- Review the public wording with competent audio personnel before publication.
An educational listening demonstration can use synthetic audio clips under qualified supervision and an accepted playback arrangement. The engineer may show that two files have the same sample peak yet different content over time. Pupils can compare the displayed waveform shapes and learn why a maximum numerical value does not summarise all of the signal’s experience. It is not necessary to expose students to unusually loud audio to teach this relationship.
Authoritative audio sources and further case studies
- ITU-R BS.1770-5 · Audio Programme Loudness and True Peak
- EBU R 128 · Loudness Normalisation
- EBU · Loudness Measurement Overview
- Adobe Audition · Digital Peak Normalisation
- Shure · Microphone and Decibel Fundamentals
- Gold Standard · Core Requirements and Evidence
- Gold Standard · Evidence, Claims and Verification
- Case 194 · Choir Metronome and Wrong Beat Unit
- Case 195 · Astronomical Magnitude and Brightness Ratio
- Gold Standard Casebook
ITU and EBU recommendations define professional measurement practices; the appropriate distribution requirements depend on the actual programme. GS-R01–GS-R12 is eduKate’s educational management framework, not an external audio accreditation.