Group Delay/phase responce

LOL  anyone?

Any linear system could be represented by a transfer function, which contains all the informations about its behaviour. Even a loudspeaker+enclosure, under a certain degree of approximation, is a linear system with a transfer function (TF). A linear system with a sine wave at the input, gives you a sine wave at the output with the same frequency, but eventually different amplitude and phase. The frequency response is the plot which represents the ratio between the amplitude of the output ant the input (usually in dB) at the different frequencies. The phase response plot represents the difference in phase between the input and the output at the different freqeuncies. A phase difference is a time difference, and thus the group delay. A non-linear phase response cause the fact that the system produces a frequency-dependent delay. Any waveform (Fourier's  theory) could be represented by a sum of sine waves. But because of the group delay, the waves at the different frequency are differently delayed. This cause a different shape in the input and output waveforms. So, we want an almost constant group delay (the value is not so important, and could be easily compensated with a DSP), thus an almost linear phase response (the group delay is the opposite of the derivative of the phase respect to the frequency in rad/s)

Im not sure that Im remotely correct here but my understanding is that the cabinet produces differing resistance to cone movement at differing frequencies. This resistance to cone movement alters the reactive component and hence impedence of the coil. since the phase angle is the ratio of the coil resistance (constant) to impedence if the impedence changes then the phase changes. Corrections anyone?

Hm,
the best case is that you design your system as a minimum phase system(it's not possible to have "faster" system without the use of FIR filters[which quit have delay to work  even at low frequencys]).
It's a disputed question whether you can hear group delay.
Definitely you can hear it at low frequencies and you can hear it when theres a jump in the group Delay at your crossover (for example at 1500Hz)-> then  transients will be torned up(don't know if it's the right word)- for example a snare or somthing like this...
So you have to try to get a flat group delay(group delay means the derivative of phase -> a spectrally delay)

tb_mike wrote: Once people realise what it is,then you have to see if your ear can even hear what looks bad on a graph. It certainly is sensitive to certain things in certain frequency ranges but not others.

If you're able to understand the german language, there are great publications available(including  the  question:what kind of improvements are "hear"able....)

unknown_artist wrote: tb_mike wrote: Once people realise what it is,then you have to see if your ear can even hear what looks bad on a graph. It certainly is sensitive to certain things in certain frequency ranges but not others. If you're able to understand the german language, there are great publications available(including  the  question:what kind of improvements are "hear"able....)