Jnoisemeter is a small app designed to measure audio
test signals and in particular noise signals.
The simplest use is to measure the S/N ratio of your
sound card. If you can calibrate the input levels of
your soundcard it can also be used (with some external
hardware) to measure noise levels of any type of audio
equipment, including preamps and microphones.
There are various 'standard' ways to measure noise
(and S/N ratio). All amount to some combination of a
particular weighting filter and a particular detector.
Jnoisemeter has the following filters:
No filtering at all, the signal is passed directly
to the detector.
This is 4th order Chebyshev lowpass filter having
a noise bandwidth of exactly 20 kHz. This means
that if the input signal is white noise, the RMS
output level is the same as for a 'perfect' 20 kHz
lowpass filter. This is the normal filter to use for
'unweighted' measurements, as any noise measurement
should always use a well-defined bandwidth.
Future version may use a higher order filter.
IEC A and C
The well-known standard noise weighting filters used
to obtain dB(A) and dB(C) measurements.
This is a filter optimised for measuring low-level
background noise. It rises 6 dB/oct at low frequencies,
has a peak of around +12 dB at 6.3 kHz, and falls off
radiply after that. It should be used together with the
ITU-R468 detector described below.
ITU-R468 (Dolby variant)
This is the same filter as the previous one with around
5.6 dB less gain. See below for why this exists.
All filters except the last one have 0 dB gain at 1 kHz.
The Dolby version of the ITU filter has 0 dB gain at 2 kHz.
Apart from these filters jnoisemeter also provides a DC
blocking filter (first order highpass, 5 Hz). This may
be necessary when using the FLAT and 20KHZ filters, the
others are DC-blocking anyway.
Jnoisemeter has three detectors:
Root-mean-square (i.e. 'power') meter. The time constant
is 125 ms as per IEC standard, or 1 second in slow mode.
This measures the average of the absolute value. The one
used in jnoisemeter is actually a VU meter. A 10 times
slower version is also provided.
This is a 'pseudo-peak' detector designed specifically
to measure noise and S/N ratios. For a peak meter it is
quite slow, as it should be for noise measurements, but
at the same time it is much more sensitive to short
impulsive noise than its speed would suggest.
The original rationale for this was the presence of
impulsive noise (generated by the electromechanical
telephone exchanges of those days) on long analog audio
lines. Today long distance audio lines are all digital,
but a detector such as this one is also ideal to reveal
the typical short noise bursts and 'crackle' originating
in computers and other digital equipment.
All three detectors will show 0.0 dB for a 'digital full
scale' sine wave (i.e. peaking +/- 1.0).
The ITU-R468 standard
This uses both the ITU filter and detector, and is probably
the 'best' standardized way to measure noise. It is used by
e.g. the manufacturers of quality microphones, in particular
the European ones. It produces a result that is on average
about 10 dB higher than an A-weighted RMS measurement.
There is a 'variation' of this standard called 'ITU-ARM'.
This was devised by Dolby Inc. at the time they were selling
noise reduction technology for magnetic tape recorders.
The traditional A-weighted measurements would show very little
S/N ratio improvement when using Dolby-B. For this reason Dolby
wanted to adopt the ITU-R468 method (which shows the difference
quite clearly) but without the apparent 10 dB loss in S/N ratio
as this was deemed bad for marketing. The solution adopted was
to lower the gain of the filter, and use an average detector
instead of the pseudo-peak one. Despite the 'ITU-ARM' name
this is not an official standard, and not approved by the ITU.