Frequency Response Graphs

[Guest]

impulse response can determine whether a headphone sounds fast, punchy and resolving vs, slow and muddy.

 

But, with the exception of fast and slow, these are all descriptions of frequency bands. On what basis do you relate them to impulse response? For that matter, what kind of impulse response measurement is classified as "fast" vs "slow"?

[Mr.McMister]

But, with the exception of fast and slow, these are all descriptions of frequency bands. On what basis do you relate them to impulse response? For that matter, what kind of impulse response measurement is classified as "fast" vs "slow

Fast and slow has no direct relation to frequency response, but may make the lower frequecies sound muddy and overwhelming higher ones as the slower disipation of the greater implusles incurred by lower frequency(lower frequencies need more energy to be perceived as being of equal loudness to higher frequencies) interferes with later small impulses from higher frequency content. Impulse response measurses just that, impulse; it sends a single pulse of power(no oscilation so no frequency content) and measure hwo quickly the headphone responds to and disipates the pulse. A fast headphone will have an impulse response that shows a quick, sharp, postive excursion with the following positive and negative excursions getting progressively smaller very quickly to inaudible levels. In a slow headphone the impulse cause a slower rise to the initial leading peak and take longer to dissipate and in extreme cases of very bad headphones an excursion following the leading one will be of greater amplitude. An example of a headphone with an excellent impulse response is the audeze lcd 2(or just about any good planar magnetic or electrostatic for that matter) and and an example of a slower headphone is the grado sr60i.

 

A slow headphone can still sound good as long as it's light on bass as in the case of the sr60i, but as a general rule faster is better for accuracy.

[Guest]

In a slow headphone the impulse cause a slower rise to the initial leading peak

 

If the headphone has slow rise time, how can it produce high frequencies at all?

 

Also, do you have a source for the slow SR60i response? I can't find any impulse response graphs or CSD plots that show what you are describing.

[Mr.McMister]

If the headphone has slow rise time, how can it produce high frequencies at all?

 

Also, do you have a source for the slow SR60i response? I can't find any impulse response graphs or CSD plots that show what you are describing.

The results of impulse responses are in the microseconds which leaves plenty of time for the high end of audible frequencies to be produced adequately, but still long enough to show how the dissipating response might interfere with following pulses and how it might be slightly less representative of the original waveform shape due to slower intial rise. Take a look at impulse respones and see how they relate to the leading edge on the square wave response plots abd you'll see what I mean when I say less representative of the original waveform, you'll be able to see this relationship best against the higher frequency square wave plots.

.

Innerfidelity has impulse response on all their measurements where you can find what i'm describing. The differences are small and the sr60i isn't a great example of a slow headphone, but that's just one I could think of off the top of my head that would show a noticable enough difference(there certainly are much better examples out there).

 

CSD plots should be taken in addition to impulse response as a csd plot will show the addition of the impulse response and driver+chamber resonances at each frequency. Chamber and driver resonances of course have a much greater effect on the sound over a longer period of time.

[Guest]

The results of impulse responses are in the microseconds which leaves plenty of time for high frequencies to be produced adequately, but still long enough to show how the dissipating response might interfere with following pulses and how it might be slightly less representative of the original waveform shape due to slower intial rise.

 

If the rise time isn't fast enough then the driver is bandwidth limited and simply won't be able to produce those frequencies, correct? Futhermore, if there is a deviation from the signal of the original program, that would also be apparent as harmonic distortion.

 

I managed to find the Grados on innerfidelity, thanks.

[Mr.McMister]

If the rise time isn't fast enough then the driver is bandwidth limited and simply won't be able to produce those frequencies, correct? Futhermore, if there is a deviation from the signal of the original program, that would also be apparent as harmonic distortion.

 

I managed to find the Grados on innerfidelity, thanks.

You are correct that impule response will limit the bandwidth of a driver(you see this with the bandwith ratings of planar headphones being up in the 50khz+ range) and yes it should be apparent in harmonic distortion so yo can use those other measurements in lieu of impulse response. It does however give a better understanding of the response of the driver in isolation from chamber resonances. This makes it usefull in finding a good driver that may sound better over all in a different housing(great for modders), but for the headphone unmodified as a whole it can be ignored when the other measurements are used.