I noticed that for IEMs, BA drivers normally have shorter FR range, say 20Hz~17KHz, whereas dynamic drivers normally have the standard 20Hz~20KHz. Besides, some expensive headphones have raher astonishing 5Hz~40KHz FR range.
 
So, my question is, does it matter? AFAIK, normally human doesn't notice high frequency (18KHz and above) on normal listening level. One of my friends said the wider the FR range, the better the quality because while it is absurd to listen to 40KHz tone, the headphones are capable of producing wider FR, which is translated into better portraying of normal listening FR range.
 
I was kinda puzzled, as I don't normally ridicule others' opinion no matter how bizarre they seem. But anyone has any idea if FR range=Quality or it doesn't matter?
 
Thanks!

I think you'll find that, in general, claims of very extended frequency response, like your 5 Hz - 40 kHz example, are very misleading. Not only can we not hear those frequencies, but the headphones/IEMs in question usually cannot reproduce them meaningfully anyway. And music doesn't really contain any meaningful information at those extremes, either. So, no. I don't think that those numbers, in particular, are gonna tell you much. The shape of a FR graph should be examined in addition to some other measurements (if you find them useful) and subjective reviews. So, wide frequency response does not automatically equal high quality.

But then, does wide FR range show that the headphone is technically capable?

In some filters, phase gets distorted well before the cut-off. Maybe some headphones have better phase response at the audible frequencies if their upper frequency cut-off is beyond the audible frequency.
 
That said, 40kHz seems a bit like overkill. IMO What really maters is FR (both phase and amplitude) in the audible range.

Frequency response specs are only meaningful with dB tolerances (e.g. 20 Hz to 20 kHz +/- 3 dB). This is almost never the case with headphones, so the manufacturers can invent whatever numbers they want. You can ignore them, and it is best to check actual frequency response graphs instead, preferably from more than one source, and ones that do not include equalization or smoothing.

Quote:

stv014 said:

Frequency response specs are only meaningful with dB tolerances (e.g. 20 Hz to 20 kHz +/- 3 dB). This is almost never the case with headphones, so the manufacturers can invent whatever numbers they want. You can ignore them, and it is best to check actual frequency response graphs instead, preferably from more than one source, and ones that do not include equalization or smoothing.

Click to expand...

 
Now we are talking. Me likes 

No

stv014 said:

Frequency response specs are only meaningful with dB tolerances (e.g. 20 Hz to 20 kHz +/- 3 dB). This is almost never the case with headphones, so the manufacturers can invent whatever numbers they want. You can ignore them, and it is best to check actual frequency response graphs instead, preferably from more than one source, and ones that do not include equalization or smoothing.

Click to expand...

+1.

I'm reminded of a Sony headphone a few years ago, that they spec'd as being able to cover something like 10hz to 110khz (yes that's infrasonic all the way up to the upper-threshold of bat hearing). I'm sure it was at something like -70 dB or some equally nonsense reference though.

Shameless repost: http://www.rane.com/note145.html

I think the only manufacturer where I pay attention to their FR spec is Koss, not because they're accurately done, but because Koss will indirectly rank their products for you by FR (the wider the FR, the higher up they think the product should be positioned; for whatever reason they don't just say this, and because most of their products exist within a very narrow price band, it's sometimes hard to sort them out). This doesn't mean I believe their numbers, it's just kind of a "trend" I've noticed based on their marketing literature. I'm sure other manufacturers participate in a similar numbers game, but I haven't ever taken the time to look into it for say, Sennheiser.

Also note that with some IEMs you're dealing with a pro audio supplier, and you may very well be seeing real numbers. I would suspect this is the case with at least Etymotic (at least the ER-4) and JH Audio, but could be mistaken.

Quote:

ultrabike said:

In some filters, phase gets distorted well before the cut-off. Maybe some headphones have better phase response at the audible frequencies if their upper frequency cut-off is beyond the audible frequency.
 
That said, 40kHz seems a bit like overkill. IMO What really maters is FR (both phase and amplitude) in the audible range.

Click to expand...

 
sorry for the newbe question, but this thread interested me and i want top understand it. iv heard the term "phase" in the thread about equalization too and i dont understand it. its not in the glossary of terms and wikipedia gave me a very technical explanation that had something to do with sine waves and distortion? could someone please clarify what "phase" is?
 
Quote:

stv014 said:

Frequency response specs are only meaningful with dB tolerances (e.g. 20 Hz to 20 kHz +/- 3 dB). This is almost never the case with headphones, so the manufacturers can invent whatever numbers they want. You can ignore them, and it is best to check actual frequency response graphs instead, preferably from more than one source, and ones that do not include equalization or smoothing.

Click to expand...

 
same with tolerance/sensitivity. could someone explain? im assuming these are electronical measurements relevant to specs such as frequency range and impedance? 
 
edit: obobskivich thank you for the link, looks like theres some interesting stuff on there.

Tolerance is being used by stv to refer to "qualifiers" for measurements - you don't just say 100 dB, you say 100 dB at 1khz at 1M blah blah.

Sensitivity is a spec that relates power input to acoustic output.

And phase is...more complex to explain. Try reading through some of Rod Elliott's writings on measurements and data.

The difference between 17kHz and 20kHz is very small... Basically one note on a piano.

Basically, a high frequency roll-off causes a phase shift which means that frequencies in that range are slightly delayed. That's usually not an audible problem though, especially near the hearing limit (~ 20 kHz).

Quote:

xnor said:

Basically, a high frequency roll-off causes a phase shift which means that frequencies in that range are slightly delayed. That's usually not an audible problem though, especially near the hearing limit (~ 20 kHz).

Click to expand...

 
^ Yup

Quote:

obobskivich said:

...
And phase is...more complex to explain. Try reading through some of Rod Elliott's writings on measurements and data.

Click to expand...

 
Would you mind posting a link in the thread? Sounds like interesting reading...

Quote:

obobskivich said:

Tolerance is being used by stv to refer to "qualifiers" for measurements - you don't just say 100 dB, you say 100 dB at 1khz at 1M blah blah.
Sensitivity is a spec that relates power input to acoustic output.
And phase is...more complex to explain. Try reading through some of Rod Elliott's writings on measurements and data.

Click to expand...

 
so when you say tolerance you mean that advertising a spec is useless when its not in relation to a different spec?
 
i ask about sensitivity because of my dt 770s. before getting them i was advised to get an amp aswell because 250 ohm is high impedance and i wouldnt be able to drive them straight off my laptop. but then i got the cans and not only would they work fine with my laptop, they sounded fine out of my phone too, no amp. i asked about this and was told its to do with sensitivity...
 
Quote:

bigshot said:

The difference between 17kHz and 20kHz is very small... Basically one note on a piano.

Click to expand...

 
algorithms and such right?
 
Quote:

xnor said:

Basically, a high frequency roll-off causes a phase shift which means that frequencies in that range are slightly delayed. That's usually not an audible problem though, especially near the hearing limit (~ 20 kHz).

Click to expand...

 
are we talking about a glitch in recording or headphones?