Introduction to adding sound levels correctly
Decibel levels are logarithmic ratios, so ordinary arithmetic addition does not describe the combined sound from several sources. This calculator first converts every entered level to a relative linear quantity, adds those quantities and converts the result back to decibels. It can use the normal incoherent energy method, the special coherent in-phase method, or logarithmic subtraction to remove a measured background level.
For independent machines, traffic, fans, voices and most environmental sounds, select Incoherent energy sum. Two equal independent sources produce a rise of 3.01 dB, while ten equal sources produce a rise of exactly 10 dB. Select coherent addition only for signals with a fixed phase relationship that arrive in phase at the same point. That uncommon condition produces a 6.02 dB rise for two equal signals.
Every input must describe the level at the same receiver position and use the same frequency weighting. The weighting menu labels the answer as dB SPL, dBA or dBC; it does not convert between those systems. A broadband dBA reading therefore must not be mixed with an unweighted dB SPL reading.
The distinction between linear quantities and displayed levels is central to useful sound analysis. A difference of 10 dB represents a factor of ten in acoustic energy, not an extra ten units of energy. A source at 90 dB therefore carries ten times the mean-square contribution of a compatible source at 80 dB. When those two independent sources operate together, the result is only about 90.41 dB because the stronger source already dominates the total.
This behavior also explains why averaging decibel numbers arithmetically is generally wrong. The simple average of 60 and 80 dB is 70 dB, but an equal-duration energy average is about 77.03 dB. Addition, averaging and subtraction must all occur in the corresponding linear domain. Decibels are converted back only after the required linear operation has been completed.
Formulas for logarithmic decibel addition
A level compares one physical quantity with a reference quantity. In general:
Here and have the same units. The multiplier is 10 for an energy or mean-square ratio and 20 for a pressure-amplitude ratio. For sound pressure in air, both equivalent forms are:
The standard reference pressure in air is . The calculator accepts negative levels because a measured pressure can be below that reference. The reference itself cancels when compatible levels are combined.
For independent sources with levels through , the incoherent total is:
The individual linear energy ratio used in that sum can be written as:
The value is relative rather than an absolute watt measurement. That is sufficient because every compatible input uses the same reference. The calculator also evaluates the expression relative to the largest entered level. This numerically stable arrangement avoids unnecessarily huge intermediate values while producing the same mathematical answer.
Equal sources provide a useful check. If each one has level , the total is . The result can never be lower than the loudest input, and it cannot exceed that input by more than .
For two independent sources, it is often convenient to express the total using the higher level and the difference between them:
This difference form shows why a much quieter source barely changes the total. A 0 dB difference adds 3.01 dB, a 3 dB difference adds about 1.76 dB, a 10 dB difference adds only 0.41 dB, and a 20 dB difference adds about 0.04 dB. A source can still matter for tonality or annoyance even when its effect on the broadband numerical total is small.
The 3.01 dB and 6.02 dB addition regimes
Independent sources add mean-square pressure, giving the familiar equal-source increase:
Perfectly coherent, in-phase sources add pressure amplitudes instead:
For unequal coherent sources, the pressure ratios are summed:
Coherent addition is an upper-bound result at a particular frequency and location. Moving the receiver changes phase and can turn reinforcement into cancellation. Broadband machinery, road traffic, HVAC systems and unrelated loudspeakers normally belong in incoherent mode.
Real sources can occupy an intermediate state when they are partially correlated. For two signals, a correlation coefficient can appear in the mean-square pressure expression:
When the correlation coefficient ρ is zero, the ordinary energy result applies. When ρ is one and the signals are in phase, pressure amplitudes reinforce completely. Negative correlation or an out-of-phase relationship can reduce the combined pressure. This calculator intentionally offers the two clear limiting cases rather than asking for frequency-dependent phase and correlation data that ordinary sound-level surveys rarely provide.
A pair of loudspeakers driven from the same electrical signal is not automatically coherent at every listening position. Their acoustic paths have different lengths, their drivers introduce phase shifts, and reflections create additional arrivals. Coherent mode is best treated as a specifically justified in-phase estimate, not as the default simply because two devices share a program source.
Worked example: four packaging-line machines
Suppose four machines measure 88.5, 84.2, 81.0 and 78.4 dBA at the same operator position when tested separately. Because the machines are independent, their relative energies are:
The displayed linear values are rounded for readability. Using full precision, the combined level is:
The 88.5 dBA machine supplies most of the energy, but removing it does not remove all sound; the other machines remain. The contribution table produced by the calculator reports both the share and the maximum reduction available from eliminating each source.
This example also demonstrates why the largest numerical level should usually receive early attention in a noise-control study. Treating the quietest 78.4 dBA machine alone cannot substantially reduce the 90.67 dBA total. Even perfect elimination of a small contributor leaves the larger energy terms unchanged. By contrast, reducing the leading machine may reveal that several secondary sources then become important.
Noise-control benefits should therefore be recalculated after each proposed change. If the dominant machine receives an enclosure that reduces its level by 10 dB at the operator position, its revised value would be 78.5 dBA rather than zero. Enter 78.5 with the three unchanged levels to estimate the new total. This approach represents the residual sound realistically and avoids claiming the unattainable saving associated with complete removal.
The result remains an estimate unless the individual measurements represent the same operating state. Production speed, material being processed, maintenance condition and microphone placement can all alter a machine’s level. Documenting these details makes a before-and-after comparison far more defensible than a set of unlabeled meter readings.
Reference decibel combinations
The following values provide quick checks for the arithmetic. The coherent column assumes complete in-phase reinforcement and should not be substituted for normal environmental or occupational source addition.
Checks for incoherent and coherent addition
| Sources | Energy sum | Coherent sum |
|---|
| 60 + 60 dB | 63.01 dB | 66.02 dB |
| 75 + 72 dB | 76.76 dB | 79.65 dB |
| 80 + 60 dB | 80.04 dB | 80.83 dB |
| Three sources at 80 dB | 84.77 dB | 89.54 dB |
| Ten sources at 70 dB | 80.00 dB | 90.00 dB |
The 80 and 60 dB row is especially instructive. Although both measurements are valid sources, the 20 dB separation means the quieter source has only one hundredth of the higher source’s relative energy. It consequently raises an incoherent total by only about 0.04 dB. That change is normally smaller than practical field measurement variation.
Conversely, many individually modest sources can create a substantial total. Ten independent 70 dB sources reach 80 dB, and one hundred such sources would reach 90 dB if all levels applied at the same receiver under compatible conditions. The logarithmic scale compresses a wide range of linear energy, but it does not make the accumulated energy disappear.
Formula for removing a background level
When a total is measured with equipment operating and a residual background is measured with it stopped, the source-alone level is:
The calculation requires . A difference under 3 dB is highly sensitive to normal meter uncertainty, a difference from 3 to 10 dB deserves caution, and a difference above 10 dB generally needs only a small correction.
The amount subtracted from the measured total can be stated as a correction determined by the total-to-background separation:
The correction C is negative because the source-alone level is below the measured total. At a 10 dB separation, the correction is about −0.46 dB. At 6 dB it is about −1.26 dB, and at 3 dB it is about −3.02 dB. As the separation approaches zero, the inferred source energy approaches zero and the correction becomes extremely sensitive to tiny changes in either reading.
Background subtraction assumes that the source and residual background are independent and that the background remains reasonably stable between measurements. It is not valid to subtract ordinary decibel numbers directly. For example, a measured total of 70 dBA with a background of 60 dBA does not imply a source level of 10 dBA. Correct logarithmic subtraction gives approximately 69.54 dBA.
A background reading above the measured total is physically possible when conditions fluctuate, but it cannot produce a valid source estimate through this equation. The likely causes include changing traffic, wind, occupancy, production activity or measurement uncertainty. Repeat the measurements under better-controlled conditions rather than forcing a numerical answer.
Frequency weighting and compatible sound measurements
Frequency weighting changes how a sound level meter responds across the audible spectrum. A-weighting reduces the influence of low and very high frequencies and is widely used for occupational and community-noise reporting. C-weighting has a flatter response over much of the audio range and is often useful for high-level or low-frequency-rich sound. Unweighted dB SPL should be identified with its bandwidth and measurement settings when those details matter.
The calculator’s weighting selector does not transform a dB SPL value into dBA or dBC. Such a conversion requires spectral information because the correction differs by frequency. Two sounds with the same unweighted level can have very different A-weighted levels if one is concentrated at low frequency and the other lies near the most sensitive range of hearing.
All entries in one sum should also use compatible time characteristics. A one-second maximum, a fifteen-minute equivalent level and an instantaneous fast response do not represent the same quantity. Combining them may produce a numerical result, but that result has no clear physical interpretation. Prefer levels measured over matching periods or source-specific levels that genuinely represent simultaneous operation.
When octave-band or one-third-octave-band measurements are available, add sources within each corresponding band first. The resulting band spectrum can then be weighted and combined using the appropriate standard corrections. Band-by-band work helps identify tones, low-frequency dominance and control opportunities that a single broadband number conceals.
Instrument calibration and microphone conditions matter as well. Use an appropriate acoustic calibrator before and after an important survey, fit a windscreen where needed, avoid handling noise, and keep the microphone away from unintended reflective surfaces unless the measurement method specifies otherwise. A precise logarithmic calculation cannot repair inconsistent input data.
Equivalent levels, duration and changing noise
Source addition answers what happens when compatible sources contribute at the same receiver. Noise that changes over time often requires an equivalent continuous sound level instead. The energy-equivalent level over several time intervals is:
The duration terms distinguish time averaging from simultaneous source addition. If two independent machines operate together, their energies are added for that interval. If one machine operates for one hour and another operates later for seven hours, their interval energies are weighted by duration instead. Treating sequential events as simultaneous sources would overstate the level present during either period.
The OSHA and NIOSH durations shown with a calculated result are contextual comparisons, not a complete dose calculation. Occupational assessment must consider all exposure periods, hearing protection rules, instrument settings and the requirements of the governing standard. NIOSH generally uses an 85 dBA recommended exposure limit with a 3 dB exchange rate, while OSHA’s permissible exposure framework uses different criteria and a 5 dB exchange rate.
A-weighted equivalent level is also different from peak sound pressure. Short impulsive sounds can have important peak levels even when their energy-averaged contribution is modest. Use instrumentation and criteria suited to impulsive or peak assessment rather than trying to infer a peak from an ordinary broadband sum.
Distance, position and propagation assumptions
Every source level entered here should apply at the same receiver position. A level measured one metre from one machine cannot be combined directly with a level measured ten metres from another unless both are adjusted or remeasured for the intended receiver. Under ideal free-field conditions, the approximate change for a point source follows:
This ideal relation predicts a 6.02 dB decrease for every doubling of distance from a compact point source. It is not universally applicable. Line-like sources can decline more slowly, while barriers, ground interaction, atmospheric absorption, directivity and reflections alter propagation. Indoors, reverberant sound may prevent the expected free-field decrease.
Positions near a wall, floor or corner can also differ substantially from free-field positions. Measurement standards often specify microphone height, façade corrections and distances from reflective surfaces. Record the position carefully if results will support engineering, permitting or compliance decisions.
The combined result describes the selected observation point, not an intrinsic sound power level for the equipment. Sound pressure depends on the acoustic environment and distance. Sound power is a source property determined by specialized procedures and should not be confused with a local sound pressure reading.
How to use the decibel source list
Choose the physical addition mode, then choose the label matching the meter readings. Enter one source level per row. Blank rows are ignored, and the Add source button supports up to forty entries. Use background subtraction only with a measured total and a lower residual reading. Press Calculate combined level to show the total, rise above the loudest source, source count, exposure context and contribution chart.
Copy link stores the current mode, weighting and entries in the URL. Copy result copies the visible calculation, while Download summary creates a CSV file. Reset returns the page to three blank source rows. Keep extra decimal places during entry because the calculator rounds only displayed results.
For a source inventory, measure or predict each source at the receiver with only that source operating when practical. If individual shutdown tests are impossible, use a validated acoustic model or another method that isolates source contributions. Entering several levels measured while all equipment was already running would count the same acoustic energy more than once.
After calculating, compare the combined level with the loudest input. The increase is a useful reasonableness check. A total only a few hundredths of a decibel above the loudest source means the remaining entries are collectively minor in the selected linear domain. A rise near 3 dB often indicates two similarly important independent contributions. Larger rises require several substantial contributors or coherent reinforcement.
The contribution share reports each source’s fraction of the summed linear quantity. In standard mode it is an energy share; in coherent mode it is a pressure-amplitude share. The “saving if removed” column recalculates the total without that entry. It is an upper bound because real treatment normally reduces a source rather than eliminating it completely.
Use the contribution data as a screening aid, not as the only control-selection criterion. Feasibility, cost, worker access, spectral character, maintenance and the number of exposed people may make a smaller contributor worth treating first. A tonal source can also remain perceptually or regulatorily important even when its broadband energy share is limited.
Interpreting uncertainty in a decibel result
A calculator can retain many decimal places, but field sound measurements are not exact. Instrument tolerance, calibration drift, microphone placement, environmental conditions and source variability all contribute uncertainty. Reporting 86.374921 dBA from inputs known only to the nearest decibel implies unjustified precision. Two decimal places are useful for checking arithmetic, while a final report may appropriately round more coarsely.
Uncertainty becomes particularly important in background correction. When total and background levels are close, the desired source contribution is the small difference between two much larger linear quantities. A minor change in either measurement can then cause a large change in the corrected source level. The calculator warns about small separations, but a warning cannot turn weak measurement conditions into reliable evidence.
Repeated samples can help characterize variability. Measure over representative operating cycles and note whether the source is steady, intermittent, cyclical or impulsive. Weather conditions should be suitable for outdoor measurements, especially where wind direction or atmospheric stability affects propagation. Do not average a few selected readings merely because they support a preferred conclusion.
For formal work, state the measurement method, instrument class, calibration information, frequency and time weighting, averaging duration, microphone location, operating state and background conditions. These records let another practitioner understand what the calculated total actually represents.
Limitations of this sound-level calculation
The calculator assumes compatible measurements taken at one position. It does not model distance, directivity, barriers, ground effects, air absorption, room reverberation or changing operating conditions. Fluctuating noise should first be represented by an equivalent continuous level over a defined period. Peak and impulsive pressure assessment requires different measurements.
A combined level is not automatically an occupational dose or a regulatory rating level. Exposure depends on duration, while environmental standards may require tonal, impulsive or time-of-day adjustments. The OSHA and NIOSH durations shown in the result are context, not a compliance determination. Obtain band-by-band data when spectrum matters and seek a qualified acoustician for safety-critical or legal decisions.
The coherent option assumes positive in-phase addition. It does not model destructive interference, frequency-dependent phase, spatial averaging or a complex coherence function. It should not be used to predict a general room response from multiple loudspeakers. Acoustic simulation or measured transfer functions are more appropriate for that task.
The source-removal saving assumes every other entered source remains unchanged. In real systems, shutting down or treating one component may alter loads, airflow, speed or operating schedules elsewhere. Noise controls can also create new paths through ventilation openings or structural connections. Re-measure after implementation instead of relying solely on a planning estimate.
Finally, dBA and dBC describe measurement weightings rather than subjective loudness in sones, hearing damage for an individual, speech intelligibility or annoyance. Those questions require additional information. The calculator performs logarithmic level arithmetic; it does not replace a complete acoustical assessment.
Common questions about decibel addition
Why does doubling independent sources add only 3.01 dB?
Doubling independent sources doubles energy. Ten times the base-ten logarithm of two is 3.0103 dB. The increase is added to the level of either equal source, so two 80 dB independent sources produce approximately 83.01 dB.
Can dBA, dBC and dB SPL values be mixed?
No. Frequency weighting changes the measured quantity. Combine only levels with the same weighting and compatible measurement conditions. Converting between weightings requires frequency-spectrum data rather than a universal correction.
Does rounding inputs affect the answer?
It can. Enter the available precision and round once at the end, especially when many similar sources contribute. The number of displayed decimals should still reflect the uncertainty of the underlying measurements.
Why is the total sometimes almost equal to the loudest source?
A source 10 dB below the loudest has one tenth of its relative energy, and a source 20 dB below has one hundredth. Such sources make only small changes to a broadband energy total, although their spectral or tonal characteristics may still matter.
Can I enter negative decibel levels?
Yes. A negative decibel level means the measured quantity is below the applicable reference; it does not mean negative acoustic energy. The inputs must still share the same reference and measurement definition.
Should several measurements of the same fluctuating source be added?
Not usually. Measurements taken at different times are not simultaneous independent sources. Use an energy-equivalent time average or another statistic appropriate to the measurement objective.
Can this calculator predict a noise-control reduction?
Yes, as a planning estimate. Replace an entered source level with its expected treated level and recalculate. Do not delete the source unless the control truly removes its contribution at the receiver.
Sources and standards for sound-level arithmetic
The equations and interpretation follow established acoustical definitions and measurement guidance. Primary references include:
- NIOSH, Criteria for a Recommended Standard: Occupational Noise Exposure, Publication 98-126, for level definitions, the 3 dB exchange rate and occupational guidance. Read the NIOSH publication.
- OSHA, 29 CFR 1910.95 — Occupational Noise Exposure, for permissible exposure and action-level requirements. Read OSHA 1910.95.
- Federal Highway Administration, Noise Barrier Design Handbook, for practical logarithmic sound-level combination guidance. Read the FHWA guidance.
- IEC 61672-1 for sound level meter specifications and ISO 1996-1 and ISO 1996-2 for environmental sound description and measurement.
Standards may define particular averaging periods, meteorological conditions, corrections and reporting conventions. Consult the current edition that applies to the project and jurisdiction. Equations on this page explain the arithmetic but do not supersede a required measurement protocol.