Why is there always +/- 3dB frequency response for speakers?

[shambles]

hi anyone,

im a newby so pls mind all the silly questions. im getting into audio engineering and am finding it a bit hard to understand certain things. i also find it hard to express what im wanting to understand! If u dont really get me, sorry.. i will try a different approach.

i am in my 2 years now in the world of audio engineering. I do understand what freq response is.

My question:

why do every speaker at the back says +/- 3 db then it shows the response of the speakers?
what is that suppose to mean? i do know that our ear can hear the difference in pressure that changes +/-3db. below 3db is a bit hard. I have also read the thread "why 3 db gain on this forum posted along time ago"

so my ultimate question :

if for the human ears to hear a difference is at +/- 3db, at what volume (SPL) are we talking about? What is the reference ? It only says +/- 3db behind the speaker. What SPL are they talking about?
Is there like a default SPL Reference?

 

[Audioguru]

Re: freqency response

Usually 1kHz is the reference for a full-range speaker in an enclosure recommended for its spec's.
Then the frequency response can be no higher than +3dB and no lower than -3dB.
The high frequencies beam in a narrow angle and will be attenuated off axis.

 

[daver01a]

Re: freqency response

You answered your own question.
The fact that the human ear will readily realize a change of 3db or greater is attributed to the physical characteristics of the human ear.
At what power levels? Just about any ,however above +160 dbw the sound levels are extremely harmful. The term db is an implication of logarithmic progression (which is the closest match to a human ear and the brains ability to correlate sound level differences).
...
In the power realm - db is used to indicate the logarithmic changes in power (ie: every 3db doubles or halves the value). For instance:
feed 1/w watts into an Amp with a gain of +3db and you get 1 watt, feed that into an amp with a gain of 6db and you get four watts ( 1/2w x 3db = 1w x 3db = 2 x 3db = 4 watts). The correct terminology in db is to reference the measurement. 0dbw is 1 watt. 0dbm is 1 miliwatt. To furtherc omplicate : In the old analog telecomm world the audio inputs to the transmitter was 16 dbm and the receiver put out +7 dbm. The +7 was attenuated to 0dbm, this was further attenuated to -16 to feed the next transmitter. Signal testing was done on a 0dbm patch bay. This arrangement allowed you to patch signals from one channel to another to isolate and work around bad channels. It also provided a point where a signal could be patched over, or back without a lot of math. Since attenuators were considered essentially "flat" in frequency response, this provided for the easiest way to pass a wideband audio signal through a repeater site. Should a drop be required to provide channels locally (or to another site on a third radio link) this provided for a universal approach. On a site with two Trop Links (Tropospheric - High Powered 10KW SHF Transmitters) and a Microwave Radio link to a third site. This approach allowed ample signal differences in cable lengths to be compensated for. Receiver gains are normalized to provide the +7dbm to the patch bay, and the patch bay was compensated and adjusted to provide the -16db required at the transmitter.
..
The reason a speakers frequency is specified as the points where the ability to reproduce audio frequency falls off in the low range to -3db and -3db in the high range - this is the relative frequency response or the acceptable audio band pass frequency. Since a drop of .5db or even 1db would not be perceived as a change - you would need at least 3db change for the "average" human ear to perceive the change.
..
I have had people over the years tell me they can hear the difference of as little as .75 db - but they were looking at a db meter (specifically an HP3555b and we were using 1.0KHz test tones set at -10 db. The individual at the other end of the wideband Radio Link was telling me he could hear the difference of as little as .75db - but in reality he was looking at the meter and his brain was telling him he could hear the difference. had he been blind folded, he wouldn't have noticed).
..
People have told me they can hear the difference between .01% THD (Total Harmonic Distortion) and .001%. The difference would relate to an audio db reference of only .275db at -10dbm0 (so I seriously doubt it).
...
+3db and -3db are the accepted Universal Standard to rate speakers.
Otherwise you would have one manufacturer telling their speaker had a useful frequency response of 20Hz -6db to 20 KHz at -6db. Another manufacturer would have a speaker with an acceptable response of only 30Hz at -3db and 16KHz at -3db. The second speaker could very well be the better speaker.
..
Without a standard - everyone would essentially be guessing.
Dave

 

[Audioguru]

Many cheap speakers have a frequency response spec of 20Hz to 20kHz. Since they do not say with an output of plus and minus 3dB then at each frequency extreme the output is heat, not sound.
Some speakers have a frequency response spec of 20Hz to 20khz when the output can drop as low as -10db.

 

[KJ6EAD]

The dB, unless otherwise stated is referenced to 1mW in most instances but for sound pressure level, the default reference is 20 µPa. -3dB is used as a standard for half peak power.

 

[The Lightning Stalker]

A speaker will be louder at some frequencies and softer at others. So if the speaker is rated +/-3 dB, it won't be any more than 3dB louder or 3dB softer at any frequency within its rated range. At least that's what I've always taken it to mean.

 

[Audioguru]

The speaker could have peaked bass at +3db and attenuated high frequencies at -3dB.
Or it could be peaked at +3dB in the highs and attenuated -3dB in the lows.
Or both ends could be peaked +3dB and the mids attenuated -3db.
Or the mids could be peaked +3db and both ends attenuated -3db.
They all sound awful, but maybe an equalizer can make it sound better.