play sound data

I read in sci.electronics.design that Peter wrote (in ) about 'play sound data', on Fri, 14 Oct 2005:

Setting aside a problem which you may not have noticed, they are just 10 dB apart in level. So the louder one will sound roughly 'twice as loud'.

But the problem is that you have given A-weighted values for levels at discrete frequencies. How are we to interpret those levels then? Do we add or subtract the A-weighting to/from all the numbers except those for

1 kHz or do we disregard the A-weighting?

Incidentally, 'dB(A)' means 'decibels referred to 1 A' (IEC 60027-3), although it's very widely misused. The better abbreviation for A-weighted levels is 'dB A'.

Now, are the numbers to be interpreted as the levels of eight discrete sinusoidal tones, or as levels of the spectrum of random noise? That makes a lot of difference. For tones, you need to build 8 oscillators and an 8-input mixing stage, while for noise you need a graphic equalizer and some patience, because adjacent (and even non-adjacent) controls interact. Then you just need to switch a 10 dB attenuator in and out to hear the difference.

I think that's enough 'insufficient data' flags for one article. And unfortunately I won't be able to follow-up, because I'm off newsgroups from noon BST tomorrow for a while.

--
Regards, John Woodgate, OOO - Own Opinions Only.
If everything has been designed, a god designed evolution by natural selection.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

Actually, due to ear's non-linear hearing sensitivity at high and low frequencies, only the mid-band frequencies will follow the 'sound twice as loud' rule.

Graham

try this

or you should be able to manufacture the noises/tones

As JW says you maybe incorrect in specifying dB(A). More info please

martin

I read in sci.electronics.design that Pooh Bear wrote (in ) about 'play sound data', on Fri, 14 Oct

2005:

What was it about 'roughly' that you did not understand? We can't tell how the subjective spectra will differ, because that depends on the absolute sound levels of the reproduction, not the relative levels.

But the difference is likely to be small anyway; most of the F-M (or rather R-D) curves 10 dB apart are almost parallel.

--
Regards, John Woodgate, OOO - Own Opinions Only.
If everything has been designed, a god designed evolution by natural selection.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

First of all I would like to thank all of you for the answers, secondly I would like to appologize for the non sense I had written ;-)

This time I'll try to be more specific. The units we make produce noise and we measured the dB values per frequency band on each side of our units at a certain distance. Per side we calculate the sound pressure level (in dB A). Below is an example of some measured values. The first set of data is without attenuator and the second set of data is with attenuator. What I would like to achieve is that we can play a sound so that our audience can "hear" the difference. I've tried cooledit96 (because I could not download a more recent version) but unfortunately I was not able to produce a sound based on the following values.

unit WITHUT attenuator

side1 side2 side3 side4 top

63 73.0 65.0 70.0 70.0 68.0 125 72.0 65.0 67.0 67.0 67.0 250 72.0 57.0 58.0 58.0 66.0 500 68.0 60.0 59.0 59.0 68.0 1000 66.0 57.0 57.0 57.0 66.0 2000 65.0 53.0 53.0 53.0 63.0 4000 64.0 51.0 50.0 50.0 60.0 8000 60.0 46.0 45.0 45.0 59.0 dB(A) 72.0 62.0 62.0 62.0 71.0

unit WITH attenuator

side1 side2 side3 side4 top

63 68.0 62.0 68.0 68.0 63.0 125 63.0 59.0 58.0 58.0 59.0 250 54.0 50.0 44.0 44.0 52.0 500 49.0 53.0 46.0 46.0 49.0 1000 45.0 50.0 42.0 42.0 46.0 2000 43.0 46.0 39.0 39.0 49.0 4000 44.0 44.0 37.0 37.0 46.0 8000 40.0 41.0 32.0 32.0 47.0 dB(A) 53.0 55.0 49.0 49.0 55.0

Does anyone has a clue how we can realise this? Thanks a lot, Peter

Not at all ! Very much not so in fact.

In any event 44dB spl @ 63 Hz is inaudible according to my curves here.

Graham