The key point I try to make is that objective measurement is critical to understanding the soundscape ("soundscape" being the cool term for the sounds in the environment around us). All too often sound measures are expressed using measures like "dB(A)" and "dB LAeq" without anyone explaining what these terms mean ... and they may not mean as much as officialdom would like you to think: In this Wikpedia article on decibels R Hickling is cited as saying that "Decibels are a useless affectation, which is impeding the development of noise control as an engineering discipline".
Here are my comments on the Noise Strategy including my attempt to explain why dB(A), dB LAeq and friends are insufficient and even misleading (and please, if you are an expert in this field please do let me know if I am misrepresenting anything here):
____
I have read through Section 5 - pp46-60, and (briefly!) through the Mayor of London's "Sounder City" paper.
The key, I think, is the text quoted in the City document taken from the Mayor's document, in particular that what we need " include noise mapping, measurement ..." (p47). We do need subjective and aspirational goals, but we need those goals expressed as objective measurable targets too. BTW, the Policy 78 they mention is on page 213 of the Mayor of London's document.
A lovely parallel is drawn in the City document (p47 & 48) between a landscape (cityscape?) and a soundscape. Both are complex multi-faceted things, both are measurable in many ways - for example a landscape has spatial dimensions but also colour and brightness and reflectivity. A soundscape also has spatial dimensions but instead of color and brightness it has sound frequency and volume. Taking this parallel just a bit further, imagine a landscape representation which muted colour (i.e. light frequency) and presented everything as a grayscale. The equivalent to this omission of colour in a soundscape would be to mute the information about sound frequency, and that's exactly what the dBA and dBC (the most common measures of sound levels) do, they are both averages of sound across all frequencies in just the same way that grey mushes together colors (light frequencies).
In context ... the recent question of plant noise being 10% below background noise and yet still being a nuisance is explained by thinking of sound as colour. Background sound is measured in dB(A) and give us the equivalent of a certain brightness of grey. Now think of plant producing a low noise. Low frequency light is red, so think of the new noisy plant machinery being as a fairly bright red light which is *on average* 10% dimmer than the current background grey, but of course the new bright red light stands out ... because it is *much* brighter than the other red sources in the background.
I see the document mentions "dB LAeq" (p50). This measure is even worse than dBA and dBC because it averages over frequencies *and* time. To go back to the light analogy, imagine a disco light that flashed lots of different colours. dBA and dBC would show the light as just grey, but you could still see it flashing. Now think about the disco light flashes, sometimes the light is dim and sometimes suddenly really bright, well "dB LAeq" would average the flashes over time and show a steady even brightness. How would that reflect the reality of standing near a disco light? It would not, of course. How can a "dB LAeq" measure reflect the reality of being on the heathrow flightpath or near a noisy pub? It would not, of course.
Perhaps to labor the point: Imagine the difference between watching the film in the cinema with all the richness of colour and light (bright daylight scenes, dark nighttime scenes). OK, now run that film though a "dB LAeq" filter and you end up watching a boring sheet of grey at a constant or very slowly varying brightness.
In our soundscape do we wish to perceive a grey slowly changing mush of average grey noise, or do we want to measure at greater resolution so we can pick out the loud crashed, the annoying plant noise, the high pitched squeal etc etc? dBA, dBC and dB LAeq are fine if you don't mind losing all the detail - and people making noise would be very happy to have the detail of their noise lost in a grey mush. I don't think we want to lose the detail.
(BTW, There is nothing wrong with having dBA, dBA and dB LAeq *as well* since these can be handy for measuring changes over longer periods, like months or years.)
The CITY SOUNDSCAPE MANAGEMENT PROCEDURE on page 51 says the right kind of things, but the "Establish unambiguous soundscape objectives" need to be expressed in objective terms, even if they are informally expressed as things like "Must be able to have a conversation at certain locations/places without having to shout.". For example, how loudly can the participants of the imagined conversation speak before this is deemed shouting? How good are their ears? How old are they? Instead how about something like "A tone generator at 10 meters distance must be detectable across all human audible frequencies using <a measurement device> with a difference from background at each 1/8th octave of n dB (dB at this resolution is fine :-)) ..." ... or something like that.
Just to repeat, there is nothing at all wrong with the suggested subjective measures on p52 & p52, they just need to be backed up by objective measures too.
p54. "Bell Happenings". Love it :-)
OK - so - those are my initial thoughts :-)
The absolute key to this establishing objective goals (largely missing from the document) on the basis of subjective aspirations (of which there are plenty in the document, and mostly very fine too), and then measuring the soundscape properly. The ideal would be a network of fixed monitoring instruments across the which could gather sound information continuously providing a living map of the soundscape.
We should introduce the idea of measuring sound traveling through buildings, and through the fabric of buildings (e.g. the concrete of the Barbican)
Here are my comments on the Noise Strategy including my attempt to explain why dB(A), dB LAeq and friends are insufficient and even misleading (and please, if you are an expert in this field please do let me know if I am misrepresenting anything here):
____
I have read through Section 5 - pp46-60, and (briefly!) through the Mayor of London's "Sounder City" paper.
The key, I think, is the text quoted in the City document taken from the Mayor's document, in particular that what we need " include noise mapping, measurement ..." (p47). We do need subjective and aspirational goals, but we need those goals expressed as objective measurable targets too. BTW, the Policy 78 they mention is on page 213 of the Mayor of London's document.
A lovely parallel is drawn in the City document (p47 & 48) between a landscape (cityscape?) and a soundscape. Both are complex multi-faceted things, both are measurable in many ways - for example a landscape has spatial dimensions but also colour and brightness and reflectivity. A soundscape also has spatial dimensions but instead of color and brightness it has sound frequency and volume. Taking this parallel just a bit further, imagine a landscape representation which muted colour (i.e. light frequency) and presented everything as a grayscale. The equivalent to this omission of colour in a soundscape would be to mute the information about sound frequency, and that's exactly what the dBA and dBC (the most common measures of sound levels) do, they are both averages of sound across all frequencies in just the same way that grey mushes together colors (light frequencies).
In context ... the recent question of plant noise being 10% below background noise and yet still being a nuisance is explained by thinking of sound as colour. Background sound is measured in dB(A) and give us the equivalent of a certain brightness of grey. Now think of plant producing a low noise. Low frequency light is red, so think of the new noisy plant machinery being as a fairly bright red light which is *on average* 10% dimmer than the current background grey, but of course the new bright red light stands out ... because it is *much* brighter than the other red sources in the background.
I see the document mentions "dB LAeq" (p50). This measure is even worse than dBA and dBC because it averages over frequencies *and* time. To go back to the light analogy, imagine a disco light that flashed lots of different colours. dBA and dBC would show the light as just grey, but you could still see it flashing. Now think about the disco light flashes, sometimes the light is dim and sometimes suddenly really bright, well "dB LAeq" would average the flashes over time and show a steady even brightness. How would that reflect the reality of standing near a disco light? It would not, of course. How can a "dB LAeq" measure reflect the reality of being on the heathrow flightpath or near a noisy pub? It would not, of course.
Perhaps to labor the point: Imagine the difference between watching the film in the cinema with all the richness of colour and light (bright daylight scenes, dark nighttime scenes). OK, now run that film though a "dB LAeq" filter and you end up watching a boring sheet of grey at a constant or very slowly varying brightness.
In our soundscape do we wish to perceive a grey slowly changing mush of average grey noise, or do we want to measure at greater resolution so we can pick out the loud crashed, the annoying plant noise, the high pitched squeal etc etc? dBA, dBC and dB LAeq are fine if you don't mind losing all the detail - and people making noise would be very happy to have the detail of their noise lost in a grey mush. I don't think we want to lose the detail.
(BTW, There is nothing wrong with having dBA, dBA and dB LAeq *as well* since these can be handy for measuring changes over longer periods, like months or years.)
The CITY SOUNDSCAPE MANAGEMENT PROCEDURE on page 51 says the right kind of things, but the "Establish unambiguous soundscape objectives" need to be expressed in objective terms, even if they are informally expressed as things like "Must be able to have a conversation at certain locations/places without having to shout.". For example, how loudly can the participants of the imagined conversation speak before this is deemed shouting? How good are their ears? How old are they? Instead how about something like "A tone generator at 10 meters distance must be detectable across all human audible frequencies using <a measurement device> with a difference from background at each 1/8th octave of n dB (dB at this resolution is fine :-)) ..." ... or something like that.
Just to repeat, there is nothing at all wrong with the suggested subjective measures on p52 & p52, they just need to be backed up by objective measures too.
p54. "Bell Happenings". Love it :-)
OK - so - those are my initial thoughts :-)
The absolute key to this establishing objective goals (largely missing from the document) on the basis of subjective aspirations (of which there are plenty in the document, and mostly very fine too), and then measuring the soundscape properly. The ideal would be a network of fixed monitoring instruments across the which could gather sound information continuously providing a living map of the soundscape.
We should introduce the idea of measuring sound traveling through buildings, and through the fabric of buildings (e.g. the concrete of the Barbican)
No comments :
Post a Comment