equal loudness, 1/3 octaves, and adding phons

Thread view:

Enable EMail Alerts

Start New Thread

Thread rating:

Brian Ravnaas - 21 Dec 2004 11:23 GMT

i am trying to piece together a spreadsheet to model noise (just boring
old 1/3 octave noise sources to be combined with 1/3 octave background
noise and/or transmission loss data, etc.)... nothing so formal, just
intended to grant an idea.

in addition to the regular dBA, i have implemented method A from ISO
532 (stevens). I decided that method B was beyond the scope of my
programming interest, so for purposes of utilizing method A, all the
data is combined into octave bands.

my first query is philosophical in nature - is applying something like
ISO 226 & Robinson-Dadson to this type of application (1/3 octave
generalized spectra intended to represent music, restaurant noise, the
noise in my bedroom, etc., etc.) an acceptable idea?

the pure tones -vs- dynamic broadband noise is a notable difference,
but can any real harm come from it, so to speak?

my second query is technical in nature - is the addition of phon levels
at various bands generally handled by conversion to sones, processing,
and then re-converting to phons? The goal would be to create a "total"
level of phons for an ISO-226 adjusted curve.

Does anyone ever add phons logarithmically ala dBA (it seems not)?
So in any case, thanks in advance,

Brian

Reply to this Message

Brian Ravnaas - 22 Dec 2004 04:44 GMT

to add a bit of food for thought:

the A, B, and C weighting scales are said to be based upon the 40, 70,
and 90 phon curves from Fletcher & Munson, and they are added
logarithmically. If one utilized Fletch & Munson (or another set of
equal loudness) to assess a "phon" level, perhaps this would be
analagous to a dynamic weighting scale, so to speak.

i 110% realize this is no official standard, but that's not the goal.

the deviations between that method and converting to sones and
estimating via 0.85*max + 0.15*sum, then converting back to phons are
not immense (at least in the curves i've peeked at so far), typically
within a few phons, but some noteworthy trends are observable.

maybe this is to peculiar or un-answerable to be of interest, but it
floated onto the brian the other day... :)

Brian

Reply to this Message

Natural Science Forum / Physics / Acoustics / December 2004

equal loudness, 1/3 octaves, and adding phons