Metrics › Tonality

Tone-to-noise ratio and prominence ratio (ECMA-418-1)

Unit: dB · Also known as TNR, PR, ECMA-418-1, ECMA-74 Annex D, prominent discrete tones, ANSI S1.13 Annex

Tone-to-noise ratio (TNR) and prominence ratio (PR) decide whether a discrete tone is prominent: TNR compares the tone with the noise in its critical band, PR compares the critical band with its neighbours. ECMA-418-1 took the method over from Annex D of ECMA-74 in 2020; its 3rd edition (2024) revised the TNR procedure.

Editions and who implements them

Project map

The project map, with only the lines for tone-to-noise ratio and prominence ratio between the projects listed for it. Agreement between a project and the code it was ported or adapted from only checks the port.

• ported or adapted from another project
• results checked against
• same contributor
• SQAT SQAT, MoSQITo, and sottek-hearing-model
• legacy legacy project

Implementations

Projects in bold and other established projects come first, then newly released, developing and legacy ones. Validation is what each project says about its own testing. The details are below, and the kinds of evidence are explained on the About page. The project map above shows which project was ported or adapted from which.

See also: Tonality from the Sottek Hearing Model (ECMA-418-2), Audibility of tones in noise (ISO/TS 20065).

How they were validated

As reported by each project; LISQM has not run the code. Two implementations that agree can still share the same error. Kinds of evidence.

MoSQITo · ECMA-74:2019 (17th ed.)

not stated

• The documentation marks it as available but not validated.

phonometry new · ECMA-418-1:2024 (3rd ed.)

standard or paper data (formula examples; no published TNR or PR value)

• Tests check worked examples printed in ECMA-418-1:2024: the critical band at 500 Hz and 1 kHz (Clause 10, Formula 2), its band edges at 1 kHz (Clause 12.2) and the proximity spacing at 150 and 850 Hz (Clause 11.6, Formula 14).
• TNR and PR of a 1 kHz tone in white noise are compared with the values expected from the standard's formulas (within 0.4 dB for TNR and 0.1 dB for PR); no TNR or PR value published in the standard is checked.

rota-order new · ECMA-418-1:2024 (3rd ed.)

standard or paper data (formula examples; no published TNR or PR value) also compared with MoSQITo

• Tests check worked examples printed in ECMA-418-1:2024: the critical bandwidth at 500 Hz and 1 kHz (Clause 10), the band edges at 1 kHz (Clause 12.2) and the proximity spacing at 150 and 850 Hz (Clause 11.6). The criteria are checked at 100 Hz, 1 kHz and 5 kHz.
• TNR and PR of tones at 140 Hz, 1 kHz and 3 kHz in white noise are compared with the values that the standard's formulas give for the known tone and noise powers, within 0.5 dB; on spectra built line by line they agree to within 1e-9 dB for TNR and 0.03 dB for PR. No TNR or PR value published in the standard is checked.
• For an order in a synthetic run-up, order_prominence gives the TNR within 1 dB of the analytic value in every speed bin, with a mean error below 0.4 dB.
• A test compares the TNR of tones at 130 Hz to 4 kHz in flat noise with MoSQITo's tnr_ecma_st and requires agreement within 1.5 dB. MoSQITo follows ECMA-74:2019, so the comparison crosses editions. The test runs only where MoSQITo is installed, which the project's continuous integration does not do.

Mosqito.NET new · ECMA-74:2019 (17th ed.)

compared with MoSQITo (its source)

Ported or adapted from MoSQITo. It was compared with the code it was ported from, which checks the port only.

• Every public function is compared with the output of Python MoSQITo for seven WAV files (three of them ISO 532-1 test signals), within 1 % (0.5 dB for levels); MoSQITo's unit tests are ported too.

MoSQITo-FDP legacy · ECMA-74:2019 (17th ed.)

own tests only

Ported or adapted from MoSQITo.

• The README lists TNR/PR as available but not validated; the tests use synthetic tones in white noise.

Before you compare numbers

Implementations of the same metric can give different numbers without either being wrong. Check these choices first.

What the projects state:

rota-order new

• Input in any unit is turned into sound pressure in Pa by an explicit calibration factor and resampled to 48 kHz.
• order_prominence analyses orders on the order-tracked (angle-domain) spectrum, so that an order stays one spectral line in every frame of a run-up. The docstring and a test show that a fixed time window long enough for the resolution the standard asks for at low frequency spreads a fast-sweeping order over the band and biases its TNR low, by about 19 dB in the test case.
• Where the text of the 3rd edition is inconsistent, the module notes how it is read: the secondary tone is added in Formulae (A.7) and (A.8), the tone spacing is |f_s − f_p|, 500 Hz takes arithmetic centring, the criteria cover 89.1 Hz to 11 200 Hz, and the misprints of Formulae (21) and (23) are resolved with their Annex B counterparts.

References

1. ECMA-74:2019 (17th ed.). Measurement of airborne noise emitted by information technology and telecommunications equipment. https://ecma-international.org/publications-and-standards/standards/ecma-74/
2. ECMA-418-1:2020 (1st ed.). Psychoacoustic metrics for ITT equipment — Part 1 (prominent discrete tones). https://ecma-international.org/wp-content/uploads/ECMA-418-1_1st_edition_december_2020.pdf
3. ECMA-418-1:2022 (2nd ed.). Psychoacoustic metrics for ITT equipment — Part 1 (prominent discrete tones). https://ecma-international.org/wp-content/uploads/ECMA-418-1_2nd_edition_december_2022.pdf
4. ECMA-418-1:2024 (3rd ed.). Psychoacoustic metrics for ITT equipment — Part 1 (prominent discrete tones). https://ecma-international.org/wp-content/uploads/ECMA-418-1_3rd_edition_december_2024.pdf