Octave Band Analysis
Doug
using is quite old so does not have the facility to do this
automatically. I'm trying to establish the current noise environment at
a particular site which is to be developed as an industrial facility.
i.e. establish a baseline noise level.
There are existing industrial facilities close by (other noise sources)
as well as passing road traffic which includes a high percentage of
heavy goods vehicles. I am doing a 15 minute LAeq measurement at 3
points on the site and also L10 and L90 measurements.
If I want to do a frequency analysis what is the correct procedure?
Some things which are not clear to me are:
1. How long should I measure for at each frequency? I have been advised
that 30s should be fine but two of my measurement points are adjacent
to the road and if a truck passes during that time surely the results
will be skewed. It seems to me that a measurement of say 15 minutes at
each frequency would be required, and then the Leq should be read.
However that would take a long time if every frequency was measured. Is
that necessary?
2. Should the measurement be linear or A weighted? I have heard
arguments put forward for both options.
3. Are there other things I haven't even thought about in doing this
measurement.
I'd appreciate advice and recommendations.
Doug
Tony
news:1161351281.5...@e3g2000cwe.googlegroups.com...
>I need to carry out an octave band analysis but the noise meter I am
> using is quite old so does not have the facility to do this
> automatically. I'm trying to establish the current noise environment at
> a particular site which is to be developed as an industrial facility.
> i.e. establish a baseline noise level.
> If I want to do a frequency analysis what is the correct procedure?
I think most consultants would agree that sequential analysis is not
suitable for varying noise. Real time analysers have been around for a long
time now, and one would expect a professional to use the right tool for the
job even if it could be done (badly) with whatever is to hand. If your time
is worth anything reasonable, the amount of time that you will spend in
getting results that could be argued about will equate to more than the hire
cost of a suitable meter.
--
Tony Woolf
My e-mail address has no hyphen
- but please don't use it, reply to the group.
Brian Marston
measurement, then its time to upgrade to newer measurement equipment.
When I first started out, we would have done this type of measurement by
tape recording long samples running those samples through an octave band
filter to chart paper, and then analysed off the chart paper - deleting
the vehicle passes. Newer equipment allows this to be done in post
processing, without the old labour-intensive steps.
In my region, we are required to get the LAeq values and the LA90
values. The truck/traffic noise can badly skew the results. If the
results are badly contaminated with traffic noise, our local regulatory
body says that the accepted LAeq,15min background is to be taken as the
LAeq,15min (raw results) minus 10dBA.
If there are sufficient gaps in the traffic the LA90 can be taken from
the raw results. The other industry in the area would be part of the
background?
Our local regulatory body has also set upper limits to the general LAeq
for each area so we are stuck with having to measure this parameter in
the 15 minute format. I now also provide them with an assessment of the
background LAeq between the cars & trucks, if the gaps are large enough
to identify.
If you are only doing spot 15 minute checks, be very wary of the diurnal
LA90 variations. Been there - done that. Hated the stress justifying
results. Best we could do at that time (1990) with the best equipment
available back then. Great fun with a QC leading the questioning!
If the background is relatively stable between the truck and cars, then
grab your LAeq samples for as-long-as-it-takes for a stable result in
each octave band on either linear or A-weighting.
Once again, its time to upgrade to newer measurement equipment.
Regulatory requirements move on and the old gears is just not up to new
requirements, or is that - old requirements misapplied.
Recently had a project where I was required to get octave band LAeq,
L01, L10 and LA90 background levels for outside a bar/hotel (600+
patrons) with 3am closing time, on a busy inner city highway. Was able
to extract a full 15 min of uncontaminated background out of 1.5 hours
of monitoring (1 hr post processing). The requirement: L10 noise limit
of "L90 plus 5 in each octave band" (without traffic noise).
Could you use the third monitoring point for your octave analysis
(provided it is well exposed to the general background of the area) and
only use the traffic affected locations as indicative of those other
locations.
Pity the poor Council officers for they know not what they ask for.
Noral Stewart
I would generally agree with Tony and Brian that getting good measurements
is hard enough with modern instruments and you see the problem with serial
octave measurements and varying sound.
You might notice that Tony in the UK and Brian in Australia often have
relatively well defined measurement requirements set by authorities (though
it appears to be a problem everywhere that the authorities do not always
understand what they are asking for). In many places in North America the
measurement requirements are not defined by local authorities, and it sounds
like that may be the situation you are in. That gives you more flexibility,
but you still have the problem of what your octave band measurements mean.
The advice you got about 30 second measurements was presuming stable sound
which you do not have
You said you would do 15 min LAeq measurements adn L10 and L90. Do I
presume the L10 and L90 would be 15 minutes also? I suggest that as a
minimum you need to repeat these measurements during several different times
of day, and if possible multiple days and multiple seasons as sound will
vary with time of day, weather conditions, and season of the year. Ideally,
the LAeq, L10, L90 and octave band measurements would be done simultaneously
over the same periods, and it is my practice to do the octave bands in
percentile as well as Leq. Does your instrument allow you to do the
percentile and Leq at the same time? Can you use the percentile feature
with the octave bands? If so, I would do each octave band for 15 minutes
and record percentile and Leq for each octave. As to whether you record
linear or A-weighted octave band levels, most criteria for octave bands are
expressed in terms of linear levels and it is customary to usually record
the linear levels. The exception would be if your primary purpose of
obtaining the octave bands is to see the contribution of each to the overall
A-weighted level. In that case, it would be more accurate to A-weight the
sound when measured..
"Doug" <doug....@pateltonra.com> wrote in message
news:1161351281.5...@e3g2000cwe.googlegroups.com...
Florian Clement - Acoustic Consultant (UK)
understanding of how to take appropriate measurements so it might not
be relevant):
Calculation of Road Traffic Noise (CRTN - ISBN 0 11 550847 3) issued by
the Department of Transport in 1988.
It describes a shortened measurement procedure to follow in order to
assess background noise levels.
Then, you might also want to look at BS4142 for assessing the impact of
the proposed industrial activity.
As for measuring 1/3 octave band or octave band data, most recent SLMs
have this function. The price can vary immensely, particularly between
a type 1 and a type 2. For environmental noise survey, a type 2 sound
level meter is deemed sufficient.
Hope this helps.