Page 1 of 1

dba (always at 1 metre?)

Posted: Fri Jun 12, 2009 7:01 am
by twinbee
Simple question which I should know really: Whenever companies give ratings in dba for hard drives, PSUs etc., does that always means at a distance of 1 metre unless otherwise specified?

Posted: Fri Jun 12, 2009 8:19 am
by dhanson865
no, unless they specify the distance you should assume the distance varies and the method of measurement varies so much that the value is meaningless.

20dba from SPCR may be quieter than 10 dba from fan company B.

Or to put it another way

House Rules: Quotes and Sayings from Doctor House
Quotes on Lies by Dr. Greg House


* "Everybody lies." First said: [episode 1],

o "It's a basic truth of the human condition that everybody lies. The only variable is about what." [episode 21]

Posted: Fri Jun 12, 2009 9:35 am
by twinbee
Oh I know - there should be a standard, and penalties should apply for false advertising etc.

But *should* it be at 1 metre in theory? Is it technically wrong in acoustic measuring generally to state the dba without giving the distance as well?

Posted: Fri Jun 12, 2009 1:59 pm
by jessekopelman
twinbee wrote:Is it technically wrong in acoustic measuring generally to state the dba without giving the distance as well?
Yes. Without giving the distance the dBA rating is meaningless. This very meaninglessness is exactly why it is so popular for marketing use. It is like saying "there is no better time to buy than now!" Notice they never say in what regard is there no better time. Hell, they don't even say for whom there is no better time -- one should assume the seller!

Posted: Fri Jun 12, 2009 8:56 pm
by twinbee
Thanks thought as much - a bit depressing to hear. Perhaps there are other ways to skew the data, but the @1m bit is the most important it would seem.

Interestingly, even when honest measurements are given (such as the brilliant Nexus Value 430 PSU), they still don't mention @1m. SPCR on the other hand tags on the "@1m" bit quite often.

Be nice for the @1m to be enforced if not in writing, then at least in practice when it comes to computer noise generally.

One often sees those charts which go from 30 to 130db to explain what each noise level is equivalent to (30 = whisper, 60 noise in a restaurant, 120 = jet engine etc. etc.), and often they never give any distance either. Odd, since that's obviously not used for marketing purposes.

Posted: Sat Jun 13, 2009 8:48 am
by Ch0z3n
My guess is with that scale is how much sound pressure is put out at the source.

Posted: Sun Jun 14, 2009 5:30 pm
by jessekopelman
twinbee wrote:Thanks thought as much - a bit depressing to hear. Perhaps there are other ways to skew the data, but the @1m bit is the most important it would seem.
Doesn't have to be @1m, just has to indicate whatever the actual measurement distance was. The idea is to figure out what the SPL will be at your actual listening distance and this can be extrapolated, but you need a starting point. Typical listening distance for a PC is 0.6-1m which is why that is what MikeC tends to use (0.6m for desktop mount application, 1m for floor mount).
twinbee wrote:One often sees those charts which go from 30 to 130db to explain what each noise level is equivalent to (30 = whisper, 60 noise in a restaurant, 120 = jet engine etc. etc.), and often they never give any distance either. Odd, since that's obviously not used for marketing purposes.
For this purpose the actual measurement distance doesn't matter, because we know it is the listening position. The problem with the SPL numbers given in advertisements is that you have no idea whether the distance used is actualyl a viable listening position. If an advertiser lists a part having an SPL of 12dBA, but got that rating by measuring from 3m away, how is that valid if you expect to put it inside your desktop PC? Also, since they didn't even specify the distance, we can't extrapolate a value for our expected listening position. We also can't compare it with any other manufacturers ratings, as they are likely to be taken from a completely different and also unknown distance.

Posted: Sun Jun 14, 2009 9:41 pm
by MikeC
Because 1m is the most widely accepted distance for SPL measurements, it is what we have to assume when the distance is not given. Even companies that are normally quite obsessive about complete technical data sometimes don't cite mic distance.

I have been using the ISO 7779 "user position" SPL measurement more recently -- mostly because so many small systems and cases are meant to be placed on the desk near the monitor and the user. This is different from typical mid tower (and larger) systems that are almost always below or beside the desk, not on it, making the PC a bit farther from the user, with the noise path being somewhat blocked by the desk.

The distance for "user position"is 0.5m horizontal from the PC, and 45cm above. The PC must be on a table built to precise specs and 75cm above the floor. Hence my ~0.6m distance.

Here's the illustration from the ISO 7779 spec:

Image

The table in the chamber is not exactly what ISO 7779 calls for but it's very close, and the distances are right.

Posted: Sun Jun 14, 2009 11:56 pm
by ATWindsor
For a single piece of hardware hey should give levels in Sound Power (SWL), not pressure. It's easier to calculate and there is not the same problems, becuase the power level is the power level, regardless of distance.

AtW

Posted: Mon Jun 15, 2009 5:33 am
by MikeC
ATWindsor wrote:For a single piece of hardware they should give levels in Sound Power (SWL), not pressure. It's easier to calculate and there is not the same problems, becuase the power level is the power level, regardless of distance.

AtW
Better in terms of a repeatable value but definitely not easier... unless you have the gear, time, expertise. Sound power is an expression of the total acoustic power emitted while SPL is a measure of the sound intensity at a specific distance and position from the sound source.

SPL only requires a sound level meter; in an anechoic chamber when high precision and repeatability is needed. Sound power requires measuring the sound at 1m distance at multiple point all around the sound source. The resulting data is then used to calculate sound power. SWL usually requires a larger chamber.

Image
This photo from Dell, provided to The Silent PC and posted on its ISO 7779 information page, show's a typical Dell setup in a hemi-anechoic chamber for measuring sound power. It's a hemispherical array of 18 measurement mics so that all the acoustic data for SWL can be captured at once. (The cost of those mics alone could be more than SPCR's annual revenue. :lol: )

In my experience, SWL does not have very good correlation to perceived sound; the A-weighted 1m SPL has better correlation to what I hear. That's based on the sound power values cited for various gear by various brands -- because I cannot measure sound power in my lab -- versus the SPL I've measured for these devices over the years.

Posted: Wed Jun 17, 2009 10:46 pm
by ATWindsor
MikeC wrote:
ATWindsor wrote:For a single piece of hardware they should give levels in Sound Power (SWL), not pressure. It's easier to calculate and there is not the same problems, becuase the power level is the power level, regardless of distance.

AtW
Better in terms of a repeatable value but definitely not easier... unless you have the gear, time, expertise. Sound power is an expression of the total acoustic power emitted while SPL is a measure of the sound intensity at a specific distance and position from the sound source.

SPL only requires a sound level meter; in an anechoic chamber when high precision and repeatability is needed. Sound power requires measuring the sound at 1m distance at multiple point all around the sound source. The resulting data is then used to calculate sound power. SWL usually requires a larger chamber.

Image
This photo from Dell, provided to The Silent PC and posted on its ISO 7779 information page, show's a typical Dell setup in a hemi-anechoic chamber for measuring sound power. It's a hemispherical array of 18 measurement mics so that all the acoustic data for SWL can be captured at once. (The cost of those mics alone could be more than SPCR's annual revenue. :lol: )

In my experience, SWL does not have very good correlation to perceived sound; the A-weighted 1m SPL has better correlation to what I hear. That's based on the sound power values cited for various gear by various brands -- because I cannot measure sound power in my lab -- versus the SPL I've measured for these devices over the years.
Whan i say easier to calculate, i mean that its easier to do caculations when you know the sound power of the source, as compared to knowing the sound pressure at 1m.

Sound power level is independant of size, so it is no suprise that there is a difference in what we hear, a small object with the same sound power as a large will be noisier at 1m. Besides, sound pressure is what we actually sense. But that doesn't mean sound power shouldn't be used, just as we use the power of and oven, not the temprature. Sound power is easier to caculate with, and you get none of the problems with "how was this measuered " (as long as you can assume the sound power level i actually correct). Sound pressure should be introduced when you are measuring the real sound for the people near the equipment, for instance on the work-station from an assembled PC or similar. For single pieces of equipment, i would much rather have the sound power.

AtW

Posted: Wed Jun 17, 2009 11:04 pm
by Fayd
MikeC wrote:Because 1m is the most widely accepted distance for SPL measurements, it is what we have to assume when the distance is not given. Even companies that are normally quite obsessive about complete technical data sometimes don't cite mic distance.

I have been using the ISO 7779 "user position" SPL measurement more recently -- mostly because so many small systems and cases are meant to be placed on the desk near the monitor and the user. This is different from typical mid tower (and larger) systems that are almost always below or beside the desk, not on it, making the PC a bit farther from the user, with the noise path being somewhat blocked by the desk.

The distance for "user position"is 0.5m horizontal from the PC, and 45cm above. The PC must be on a table built to precise specs and 75cm above the floor. Hence my ~0.6m distance.

Here's the illustration from the ISO 7779 spec:

Image

The table in the chamber is not exactly what ISO 7779 calls for but it's very close, and the distances are right.
wow, there's a standard for everything :o

Posted: Thu Jun 18, 2009 5:07 am
by jessekopelman
Fayd wrote: wow, there's a standard for everything :o
And yet they often languish in disuse and misuse.