BSM-7 specs:
Frequency response: 20- 20,000 Hz
Signal to noise ratio: 62dB, 1khz at 1pa
Open circuit sensitivity : -35dB (5.6 mv) re 1v at 1pa
Dynamic range: 95dB, 1kHz
Maximum Input Sound Level: 105dB SPL, 1kHz at 1%
Self-noise level is not indicated above, but can be inferred from the highlighted specs. I'll do my best to decipher that for you, but the quantification of noise is complicated. If DSatz happens upon this thread, I humbly ask him to correct me anywhere I've gone astray..
A microphone's self noise is expressed in terms of
equivalent noise level, most commonly as an
RMS A-weighted measurement, less commonly as a
quasi-peak CIRR / ITU-weighted measurement, sometimes both. The two measures give different numbers. RMS is essentially the average level over a short time interval (good for characterizing consistent noise) and quasi-peak is the peak level within the audible range (better for capturing transient noise spikes which peak above the general average). A-weighting verses CIRR or ITU noise weighting refers to different filtering curves that attempt to provide a meaningful match with the human perception of noise at low levels. Due to the differences in curve weighting and time-interval, a CIRR or ITU quasi-peak self-noise measurement value will be significantly higher than a A-weighted RMS measurement. It's a more accurate estimate of actual human perception of the noise, but is less frequently reported because it doesn't look as good on a specification sheet. Okay, with that out of the way..
The
Signal to Noise Ratio of a microphone is a measure of the interval between the level of self-noise (RMS, A-weighted) and a 94dB (1 pa) reference SPL. The signal to noise ratio of BSM-7 is listed above as:
62dB, 1khz at 1pa. 94-62 = 32,
so we might assume an (optimistic) RMS A-weighted self-noise level of 32dB at 1kHz for the BSM-7. We can't determine the more realistic CIRR or ITU noise level rating from the specifications provided, but consider as an example the miniature low-voltage DPA 4060 omnis I use, which are small low-voltage powered microphones like the MM's, yet more costly. They perform very well and are popular with music tapers. Their self noise level is specified as:
Equivalent noise level, A-weighted: Typ. 23 dB(A) re. 20 µPa (max. 26 dB(A))
Equivalent noise level, ITU-R BS.468-4: Typ. 35 dB (max. 38 dB)^
As you can can see, in this case the ITU noise spec is 12dB higher than A-weighted.
Comparing self-noise A-weightings, the 4060 is significantly quieter than BSM-7 (nominally by 10dB but could easily be more), yet 4060 is not the quietest microphone available by far. It is generally considered quiet enough for professional use on actors in stage-productions and for close instrument mic'ing for live classical concert use, yet not quiet enough for professional classical recording with the mics further away in quiet performance conditions without an audience. In my experience it's self noise is not low enough for very quiet close-mic'd nature and environmental recordings that will be amplified and used at levels considerably higher than that of the original sound. In such a case the microphone's self-noise is amplified enough that it does become apparent. However, their self noise is low enough for all of my music recording needs including live classical music performed in specially built quiet halls with near-silent audiences. In that case, the acoustic noise floor of the hall is higher than the self-noise of the microphone so the noise of the hall determines the noise floor of the recording. The same would remain true if I substituted a quieter microphone.
[edit- I'm not proselytizing for DPA 4060 here, only using it as comparative example, both in terms of specs and in my personal experience with its self-noise level]A microphone's
dynamic range is the difference between it's self-noise level (RMS A-weighted) and its maximum SPL level. Since max SPL is always higher than 94dB, the mic's dynamic range will always be higher than its SNR.
BSM-7 has a specified dynamic range of 95dB @ 1kHz, and a max input level of 105dB SPL @ 1kHz, before distortion exceeds 1%. But 105-95= 10dB, not 32dB as figured by the SNR spec above, so what gives? I think that has to do with the 1% distortion constraint. The microphone can accept higher SPLs before eventually clipping, but only with increasing levels of distortion Doing the math the other way, a 32dB A-weighted noise floor and a 95dB dynamic range suggests outright clipping at an SPL of 127dB, which is reasonable for that type of microphone.
Whew! Apologies for the long post, but thanks for the mental exercise. I hope that sheds some light on mic self-noise.