Why do SNR ratings tend to be about 3 dB higher than NRR ratings? I.e., what difference(s) in the testing procedures account for this gap?

Question

I read on https://eardefender.co.uk/snr-vs-nrr/ (mirror):

SNR (Single Number Rating) is the EU’s standard for showing the attenuation of hearing protection. [...] NRR (Noise Reduction Rating) is the noise reduction standard used in the United States. [...]

The most obvious difference between SNR and NRR ratings is that SNR ratings tend to be a few decibels higher. An exact formula for converting NRR to SNR doesn't exist, but SNR ratings tend to be about 3 dB higher, so a product with an NRR of 30 will have an SNR of about 33. This is because the two ratings are determined by following different testing procedures.

Why do SNR ratings tend to be about 3 dB higher than NRR ratings? I.e., what difference(s) in the testing procedures account for this gap?

I note that sometimes the gap is much higher than 3 dB, e.g. this Mpow earmuff (mirror) is rated SNR 36dB and NRR 29dB (= 7 dB gap between the NRR and the SNR).

Answer 1

Do you read french? If not, here is a small synthesis of what comes out from this document:

First, the tests are not performed of the same samples of the population:

SNR and NRR are expressed in terms of the percentage of the population they are intended to protect: 98% for the NRR and 84% for the SNR. This partially explains the difference (translated by me…)

There are also some differences on the formulas which are used for calculating these two indexes - see page 12 and 13, I have to admit that I do not have the courage to reproduce them here but here are the main differences:

• SNR is evaluated between 63 and 8000 Hz, NRR between 125 and 8000
• as already said, these two tests do not account for the same portion of the population. It is interesting to note that these tests are realized on samples of people, and asking them to "rate" the sound reduction (which can be very different for the same protection!). So, the result is a statistical measurement where the thresholds on the proportion of people is should be protective (or perceived as protective) are not the same between the european test and the american one,
• NRR is taking 107.9dB(C) as a reference level, and then substracting 3dB, whereas the SNR is taking 100 dB as reference level.

So, it is actually hard to answer with precision, especially as, as you pointed out, the difference is not constant. The methods of measuring these two indexes are not the same, the formulas are not the same and so.

But still, as the main effect, it appears that SNR is more "tolerant" as it should protect "less" people. This obviously results in an higher SNR than NRR: if you can guarantee a sound reduction of 10dB to 98% of the population, you can very likely guarantee​ a bit more on only 84%!