Understanding frequency response

[Stiffy]

I have two integrated amps and was recently comparing their technical specifications and noticed the frequency response of one of them is: 10Hz - 50Khz (Marantz) while the other is 10Hz - 100Khz.(Yamaha)

Could someone explain, without getting too technical what this means to me in terms of listening please?

[Reffc]

I have two integrated amps and was recently comparing their technical specifications and noticed the frequency response of one of them is: 10Hz - 50Khz (Marantz) while the other is 10Hz - 100Khz.(Yamaha)

Could someone explain, without getting too technical what this means to me in terms of listening please?

It means both have a design response which covers the audible spectrum and then some, that's all. Extended FR may not be audible but its important as it contains musical information important to harmonics.

Dont pay too much attention to paper specs other than checking that it's reasonabbly well designed (ie THD within 0.5% which both should be, and a FR extending either side of the audible spectrum +/- fractions of a dB which they should be).

What matters more is how they behave into a variable speaker load, not just how they measure at 1KHz. Only way of telling (audibly) is to listen!

[walpurgis]

Extended FR may not be audible but its important as it contains musical information important to harmonics.

I have heard similar things said before and I suppose it makes sense that supersonic information can interact harmonically with (or be created by) audio band signals, but I've never seen it properly explained why this might be beneficial. From what I've seen, it just seems to be an accepted presumption that the inclusion of supersonics information is a good thing.

After all, the Celestion HF1300 tweeter sounds acceptable, yet cuts off dead at 13kHz (and yes, I do know that it was often used with super tweeters).

[Reffc]

I have heard similar things said before and I suppose it makes sense that supersonic information can interact harmonically with (or be created by) audio band signals, but I've never seen it properly explained why this might be beneficial. From what I've seen, it just seems to be an accepted presumption that the inclusion of supersonics information is a good thing.

After all, the Celestion HF1300 tweeter sounds acceptable, yet cuts off dead at 13kHz (and yes, I do know that it was often used with super tweeters).

Many musical notes that we hear are not pure fundamentals, but made up partial frequencies which can be integer multiples of fundamental notes which are known as harmonics. The harmonics which define timbre have specific phase relationships and as they can be made of multiple frequencies, some of those frequencies can be above (or below in some cases) the audible spectrum. Therefore, although we cannot hear them, by definition, omitting that harmonic content must change the timbrel accuracy or quality of the note. That's the theory anyway.

[Oldpinkman]

I have heard similar things said before and I suppose it makes sense that supersonic information can interact harmonically with (or be created by) audio band signals, but I've never seen it properly explained why this might be beneficial. From what I've seen, it just seems to be an accepted presumption that the inclusion of supersonics information is a good thing.

After all, the Celestion HF1300 tweeter sounds acceptable, yet cuts off dead at 13kHz (and yes, I do know that it was often used with super tweeters).

There is probably an element of snake oil in supersonic frequency and harmonics. A lot of lp playing components roll off dramatically after 20k and yet vinyl is perceived as having a true sound. Clearly harmonics at all are impotant. The highest e flat is a mere 5k. Harmonics to 20k at least give music colour. 13k as a threshold is interesting. If I test my hearing with a pure tone, reflecting the 52 summers I have seen, I can't hear above 14k. And yet the rolloff at that frequency (measured) on my valve headphone amp was very "audible" when I corrected it. So clearly what I hear is affected by frequencies higher than I can hear (between 14k and 20k) and so presumably can be affected by frequencies above 20k. We tend to treat the audible spectrum as 20 to 20k. Maybe for my 21 year old daughter but not for this old codger. Yet removing those inaudible frequencies for me has a clearly audible effect. Confused? Yes - so am I

[walpurgis]

Richard has almost answered his own point. What he senses is missing or what he is sure he detects is not necessarily all that his ear is actually hearing. In other words there may be (and probably are) audible signals we are not aware of having heard.

[NRG]

There are some very good arguments around Slew rate and Feedback for not having a super wide amplifier bandwidth, Peter Comeau wrote a very insightful article back in 2005 for HiFi World highlighting the pitfalls.

[walpurgis]

Yes, there have always been arguments for and against.

Being probably older than most AOS members and having dodgy hearing and having experimented with super tweeters quite a bit, I tend to think that the inclusion of ultra/super sonic information is probably beneficial.

I have actually rolled in super tweeters at frequencies above that which I could detect, but noticed there was seemingly something missing when they were taken out of circuit.

[Barry]

There is a school of thought that has, as its maxim: "the product of the low frequency -3dB point and the upper frequency -3dB point should be 400,000 Hz²"; that is, the bandwidth should be balanced with a pivotal frequency (geometric mean) of 632.5 Hz, such as (for example): 20 - 20,000Hz, or 40 - 10,000 Hz. (mutatis mutandis).

I am myself inclined to this school of thought.

[walpurgis]

A bandwidth pivotal reference point is just that surely and has little to do with how we hear things, at least in aural terms. Perhaps this may be more useful from the electronic signal point of view.