Hi Joshua,
send me a PM with your email and I will send you the test pulse I use in .wav format. You can play it through any sound card or copy to a CD-Rom.
Praxis ( the software ) can generate all kinds of signal, also this. One limitation of your test is the rather rough resolution, Luigi, but i am with you.
Yes, I changed the bandwidth by modifying the capacitor value, but well outside the audio band; in my vision it's not the bandwidth extension the responsible for the results I got, but the better transient response.
I'm quite convinced that the ear works mainly in time domain: it's daily activity is the continuous localization of obstacles, voices, noises, etc.. All these activities are made by measuring the small arrival time difference of the signals at the two ears: the trigger it's the leadig edge of the signal (by the way, the transient...).
It sounds true to me. It explains how the human ear can ignore noises in favor of the sounds of interest, while this cannot be done to any recorded and reproduced material.
Of course, Joachim: you are right!! Praxis software is very powerfull and with its many post-processing tools you can made everything.
I know that I don't have an high resolution, but I use it to make comparisons, not absolute measurements. I can live with it.
Me too Luigi. The important thing is that you brought time domain analysis into the picture.
See a leaflet from Manger. As i said, i use pulse testing on speakers for ages. I did not think about amps that way though. I thought when they are fast and have a Bessel response or even better a Q of 0.5 in the upper reaches everything is fine. Colour me blind, and you talk also about the complete chain. That is a stroke of genius too.
See a leaflet from Manger. As i said, i use pulse testing on speakers for ages. I did not think about amps that way though. I thought when they are fast and have a Bessel response or even better a Q of 0.5 in the upper reaches everything is fine. Colour me blind, and you talk also about the complete chain. That is a stroke of genius too.
Hi JC,
Just read about the Constellation gear (Altair premp and Herc Amp) in TAS. Very nice!!! I bet that was a fun project to work on. Look forward to seeing the phono preamp.
I think Harley provided just enough details that I can build an Altair ... Ha, Ha. 🙂
Okay, guys don't let me interrupt. Back on/off topic ... interesting stuff.
John
Just read about the Constellation gear (Altair premp and Herc Amp) in TAS. Very nice!!! I bet that was a fun project to work on. Look forward to seeing the phono preamp.
I think Harley provided just enough details that I can build an Altair ... Ha, Ha. 🙂
Okay, guys don't let me interrupt. Back on/off topic ... interesting stuff.
John
Ivigone,
Really appreciate the new ideas you are infusing into this discussion.
Let me add someting to it from the Physiology of Hearing (James O.Pickles). Quite a bit is known about how signals propagate from the inner ear to the processing centres in the brain, and the way this nerve signal is shaped substantiates your point:
-quo- At high frequencies, above 5 kHz, the nerve fibres fire with equal probability in every part of the cycle. At lower frequencies, however, it is apparent that the spike discharges are locked into phase of the stimulating wave form. That is not to say that each fibre fires once every cycle; the fibres fire randomly, perhaps as little as once every hundred cycles on avarage. But when they do fire, they do so only in phase of the stimulus - unquo-.
For those totally unaware how the two snails inside your head work: best analogy of the cochlea is a highly specific frequency analyzer, with each small frequency band connected to a flat cable of nerve fibres that enters the brain. This flat cable is tonotopically organized, meaning that each frequency band continues to lay in order next to its neighbours while routed through different processing centres. The way these individual nerve fibres transmit the intensity in a corresponding frequency band is through their firing rate. The more intense the signal, the more rapid the firing. Below 5Khz (according to some authors <3.5Khz), as per quote above, time information is preserved.
So, the importance of maintaining time coherence between the two channels has a solid foundation in how the human hearing system works. I think your test is a good one in that respect, because good transient response is a good predictor of good time coherence.
Last edited:
Excellent work, Luigi & I am very interested in following your lead. I have asked similar questions here before about the almost obsessive focus on FFTs in the frequency domain to the exclusion of temporal domain information. Indeed there's a dictum here oft quoted that anything below -120dB is inaudible i.e an exclusive focus on frequency.
May I also get that test pulse WAV from you - I will PM you.
May I also get that test pulse WAV from you - I will PM you.
Last edited:
Luigi,
I've written this to you but believe that it warrants a post here & maybe the mods would consider branching Luigi's posts off to another thread where it can get the exposure & airing it deserves rather than languish, buried in this long thread?
I've written this to you but believe that it warrants a post here & maybe the mods would consider branching Luigi's posts off to another thread where it can get the exposure & airing it deserves rather than languish, buried in this long thread?
jkeny, this thread is optimum for discussing this sort of thing. It is not ONLY the specific example given by luigi, but much of the experience that we have had with cables, electronics, feedback, etc that makes for 'better' hi fi, rather than mid fi. This thread is continued on purpose to discuss these factors.
And you think his different pulse responses won't show different harmonic spectra?
Exactly. For a given frequency response, the time domain response to ANY excitation is determined and vice versa. No magic to one versus the other, they're just different ways of plotting the same thing. After all, one gets a frequency response as a transform of a transient, which itself is often derived from a pseudo-random excitation...
Sure, but a swept wave ( that can be convoluted into an impulse response ) does not stress the component as much as a PHYSICAL IMPULSE as the STIMULUS. Mathematically you are correct and that makes the discussion so difficult. I think we land in a one way road when we do not differentiate between stimulus and convolution.
Try to sweep a car with a sine wave on a shaker or hit on it with a hammer. The outcome MAY be different.
Try to sweep a car with a sine wave on a shaker or hit on it with a hammer. The outcome MAY be different.
Of course they will be different. If you apply two different stimuli to a system with a particular transfer function, the response is certainly different. The math forces that. But knowing the impulse response or frequency response, you can absolutely predict what will happen with a steady state excitation or a steady state excitation multiplied by an envelope or multiple tones or what-have-you. The math forces that as well.
Amplifiers are minimum phase (unless one deliberately builds in an allpass), making things even simpler.
Amplifiers are minimum phase (unless one deliberately builds in an allpass), making things even simpler.
I did not say that the stimuli are the same. Question is if a steady state signal can be constructed that stresses the system the same ( as much ) as a physical impact. If we can do that i can see progress.
A physical impact is an impulse, which can be as sharp or broad in the time domain as you want.
I know, maybe we just do not measure with enough bandwidth or power. Or to much power, that is.
Nothing keeping you from making an impulse (or pseudo-random sequence) any amplitude you want. Or using a different forcing function. 😀
- Status
- Not open for further replies.