# "Frequency Response" on a speaker datasheet provides only range. What do we assume here?

In my head, "frequency response" refers to a graph of frequency vs dB. A sinusoidal wave with frequency `f` and a relative `0 dB` is passed through the device (in this case a speaker) and the measured relative intensity (`I`) (in `dB`) is plotted against the `f`.

Now even this is already subject to a bunch of questions: is the response only measured at frequency `f`? Or over the whole spectrum, including the harmonic distortions? What about noise floor? Is it filtered out too?

But what I really don't understand is when manufactures write:

``````Power handling capacity:...........................................75 Watts
Frequency Response:.............................................5-18,000 Hz <-----
System Sensitivity:......................................89.6 SPL dB @ 1m/W
System Impedance:....................................................8 Ohms
Dimensions (HWD):....................................30 x 10 x 6 7/8 Inches
Jack Size:.........................................................1/4 Inch
Sound Coverage:.................................................130 Degrees
``````

What? What the heck do you mean? What does the cutoff represent? Is the response flat in the rest of the range? Or is that range just the flat region? How flat?

Thanks

• You can't assume anything without at least a +/-dB figure… except that claiming 5Hz at the low end already sounds suspicious. Jul 8, 2022 at 10:50
• You watch an advertisement. It's written to look a technical spec to make people want to buy it. I'm afraid there's plenty of idiots who think it's a serious engineering spec, but seemingly you are not one.
– user35252
Jul 8, 2022 at 16:35

The lower and higher bounds of the frequency response are typically frequencies where the response is typically -3dB or -10dB (not all data sheets use the same convention). Whether this is measuring the distorsion or not will not change a lot the number.

-3dB is a 0.5 factor A 3% THD+N will only add a 0.03 relative power. It is not the same magnitude.

Let’s suppose you add a 3% THDN, this add 0.128 dB. With a -6dB/octava, this can shift the frequency of 0.02 octava, 1.5% of error. It is under the given frequency precision.

• My own studio monitors claim 42 Hz to 21 kHz (+/- 2.5dB) [& as they are a serious manufacturer, dynaudio, I am inclined to believe them.] They don't appear to quote any THD figure. For all the good they are, they still don't even claim 40Hz, only 42. The 5Hz is a serious cause for concern of what the 'spec' above is claiming. For little more than comedic value - here's a sub that claims it can actually do 5Hz - aia-cinema.com/active-speakers/… Jul 8, 2022 at 16:48
• @Tetsujin I think the 5 was an error from copying from the PDF (probably an OCR mistake) -- if I remember correctly it was 50Hz, not 5. Jul 10, 2022 at 13:03

For studio/reference speakers, the range is +0dB-3dB, for PA/instrument purposes it's sometimes +0dB-10dB.

Less than 90dB/W makes it likely we are talking HiFi. 5–18000Hz with the given dimensions of the enclosure makes this a typo. It's either 50Hz at the lower end, or (in case the tens are missing) my guess would be 35Hz. HiFi (mainly because of low efficiency in connection with high mechanical damping) allows for a lower range extension at this size, but lower than 35Hz is unlikely when the upper end already ends at 18000Hz.

Now of course this is under the assumption that the specs are not a complete fabrication to start with. Which, depending on manufacturer, could be the most plausible scenario.

• Typo or lying through their teeth - hard to tell. If they don't declare RMS in the power handling, they're already suspect in my book. Jul 9, 2022 at 16:02
• @Tetsujin RMS power specs make no sense for multiway speakers (as opposed to amps) because properly dimensioned tweeters never can vent the same amount of power as woofers. "Program power" tends to be specified with the frequency distribution of pink noise. Like RMS, it should be a long-term sustainable spec. Things get increasingly murkier with "music power" (handwavingly 30% to 50% larger), "peak power" (mechanical clipping beyond that level), "PMPO" (pure fantasy numbers indicating insincere specs). Jul 8, 2023 at 11:36