How to interpret frequency response specs

[Press Record]

How to read frequency response specs:

I am curious if the difference expressed by these specs are audible.

Frequency Response: 5 Hz to 50 kHz, +0, -3 dB
Frequency Response: 20Hz – 80kHz, +/-0.5dB
and the extreme 30 hz 0 15000hz within 0.1 db

 

[MidnightSensi]

You have to look at the frequency response graph to really interpret those numbers. 50kHz is significantly above human hearing, and I don't know of many subwoofers that can get down to 5Hz. Maybe some extremely large subwoofers, Bag Ends D18E subs can get down into the 10Hz range but without a subharmonic synthesizer there isn't much material you'll get down there.... There may be some super-tweeter arrays that get up to 50kHz, but I don't think so because ultra-sonics don't really make much sense for audio reproduction. Infra-sonics make sense for "pressurizing" a room.

 

[mtrycrafts]

How to read frequency response specs:

I am curious if the difference expressed by these specs are audible.

Frequency Response: 5 Hz to 50 kHz, +0, -3 dB
Frequency Response: 20Hz – 80kHz, +/-0.5dB
and the extreme 30 hz 0 15000hz within 0.1 db

I think those numbers in the first two are incorrect: 5Hz to 80kHz would be appropriate for that -3dB loss.

But, it means that from 15kHz to 50 kHz it drops off to .5dB, in essence to audibility, flat, flat. flat Minuscule
From 50kHz to 80kHz, there is an additional 2.5 dB of loss. Cannot hear it
At the low band, 20Hz to 5 Hz, the same 2.5dB loss. Good think we just cannot detect that 2.5dB level difference as our hearing down there is poor for that level change.

 

[fmw]

No, no audible difference at least in terms of frequency response. These are all flat or accurate if you prefer. The human hearing range is around 20 hz to 20 khz. The last one of the three took the range at which most recordings are made and showed that it is REALLY flat. They could have extended the range by increasing the variance like the others did. All are fine. Nothing to worry about.

Low E on a 4 string bass guitar is around 40 hz. Low B on 5 string is about 35 hz. You can find 35hz on symphonic tympani or bass drums as well and pipe organs go that low and even lower although not much music is written for those very low registers. Most microphones start rolling off fairly severely after about 15 khz (mostly overtones anyway) so you rarely see much recorded content above 15 khz although some people can hear higher frequencies than that. So 30 hz to 16khz is a pretty reasonable range.

 

[markw]

Actually, this looks more like an amp's spec...

Assuming all three specs are from the same speaker, I agree with mtry that the high end of the first two are reversed.

The response of any piece of audio gear will drop off at the extreme ends of it's response. That is to say, that the more narrow the response range, the more flat ithe response will be.

IOW, to state these three in a more logical (and with the 50/80 transposition corrected), it would look like this:

Frequency Response:
30 hz to 15khz +/- 0.1 db
20Hz to 50kHz, +/- 0.5dB
5 Hz to 80 kHz, +0, -3 dB, Although I would expect some number in the "+" column

Considering the normal variations in a speakers frequency response over it's range, it's actions within the room itself, and your own hearing variations, I'd say these measurments all indicate a fine device. These three items would be the limiting factor in what you're hearing, not this device.

 

[fmw]

Assuming all three specs are from the same speaker, I agree with mtry that the high end of the first two are reversed.

The response of any piece of audio gear will drop off at the extreme ends of it's response. That is to say, that the more narrow the response range, the more flat ithe response will be.

IOW, to state these three in a more logical (and with the 50/80 transposition corrected), it would look like this:

Frequency Response:
30 hz to 15khz +/- 0.1 db
20Hz to 50kHz, +/- 0.5dB
5 Hz to 80 kHz, +0, -3 dB, Although I would expect some number in the "+" column

Considering the normal variations in a speakers frequency response over it's range, it's actions within the room itself, and your own hearing variations, I'd say these measurments all indicate a fine device. These three items would be the limiting factor in what you're hearing, not this device.

Not necessarily. There is no question that the major variances from a flat response will be in the inaudible range but it is certainly possible that one would be 3 db down at 50khz while the other is only 1/2 a db down at 80khz. Obviously it doesn't matter any more than how many angels fit on the head of a pin but the response figures could be correct. Also it isn't unusual for a frequency response curve to have no peaks but to have dips so the +0 -3 doesn't seem unusual either.

In a nutshell, these three response curves could be considered ruler flat. There isn't a thing to choose between them. I assume they come from preamp specifications.

 

[markw]

Not necessarily. There is no question that the major variances from a flat response will be in the inaudible range but it is certainly possible that one would be 3 db down at 50khz while the other is only 1/2 a db down at 80khz.

Yes, necessarially. The quoted decible numbers don't have to be at the extremes, although physics, logic, and generally accepted principles dictate they should be.

First off, when something states +/- one decible, that means the variation is both positive and negative, depending on a particular frequency or range, as is stated in the first two specs listed. If it was only on the negative side, it qould state +0, -1 db, as is stated in the third spec.

This is illogical. Where did that negative varaition go on the third (widest range) spec?

If a speaker is, say, 20 - 20k -/+ 1db, then that would follow through that when you widen the measured frequency extremes. That's to say that from 10 to 50 khz it would have a variation of at least that one db in the middle ranges, most likely more. That negative 1 db won't disappear and become flat (-0 db)response. It will most likely get larger.

Obviously it doesn't matter any more than how many angels fit on the head of a pin but the response figures could be correct. Also it isn't unusual for a frequency response curve to have no peaks but to have dips so the +0 -3 doesn't seem unusual either.

In a nutshell, these three response curves could be considered ruler flat. There isn't a thing to choose between them. I assume they come from preamp specifications.

Yes, it's the old angels on the head of a pin again.

 

[Press Record]

Not necessarily. There is no question that the major variances from a flat response will be in the inaudible range but it is certainly possible that one would be 3 db down at 50khz while the other is only 1/2 a db down at 80khz. Obviously it doesn't matter any more than how many angels fit on the head of a pin but the response figures could be correct. Also it isn't unusual for a frequency response curve to have no peaks but to have dips so the +0 -3 doesn't seem unusual either.

In a nutshell, these three response curves could be considered ruler flat. There isn't a thing to choose between them. I assume they come from preamp specifications.

Sorry for not being more specific. FMW is right, they come from preamp specifications.

I guess my question is directed at the first speck. If it is 3db down at 50khz, can we surmise that it is flat at 20Khz or do we need to see it plotted on a graph in order to know that.

 

[fmw]

Sorry for not being more specific. FMW is right, they come from preamp specifications.

I guess my question is directed at the first speck. If it is 3db down at 50khz, can we surmise that it is flat at 20Khz or do we need to see it plotted on a graph in order to know that.

Almost certainly the dip in the frequency response is way above the audible range. Also, if you want flat, you might as well use the preamps in the 002. The high end preamps like Neve or SSL are anything but flat.

 

[mtrycrafts]

Sorry for not being more specific. FMW is right, they come from preamp specifications.

I guess my question is directed at the first speck. If it is 3db down at 50khz, can we surmise that it is flat at 20Khz or do we need to see it plotted on a graph in order to know that.

If the original spec is in fact correctly stated as 20Hz to 80kHz is -.5dB, then it cannot be down 3 dB at 20kHz or at 50Khz more than that .5 dB. And, the -3dB has to come below 20Hz.
So, this is one reason the original spec is in question.

 

[mtrycrafts]

Not necessarily. There is no question that the major variances from a flat response will be in the inaudible range but it is certainly possible that one would be 3 db down at 50khz while the other is only 1/2 a db down at 80khz. Obviously it doesn't matter any more than how many angels fit on the head of a pin but the response figures could be correct. Also it isn't unusual for a frequency response curve to have no peaks but to have dips so the +0 -3 doesn't seem unusual either.

.

As I indicated in my other post just now, if the specs as originally posted is correct, it cannot be down 3dB at 50kHz or that .5 dB spec then has to be incorrect.

 

[PENG]

Sorry for not being more specific. FMW is right, they come from preamp specifications.

I guess my question is directed at the first speck. If it is 3db down at 50khz, can we surmise that it is flat at 20Khz or do we need to see it plotted on a graph in order to know that.

Strictly based on the given specs, you can surmise or assume but you cannot be sure unless you see the whole curve. In reality, I do believe it is a safe assumption that at 20 Khz most SS amps will not dip more than 0.5 dB.