Yes, I know..perhaps that is why I called it phasing/flanging?
We are not talking about those issues. I suspect that getting the records out of sync by 10 usec would be really difficult![/quote]
No, it is not difficult. The comb filtering was so good that when I inverted one program, total and absolute cancellation of ALL program occurred..A stereo with 3 kilowatts running to clipping going absolutely silent. Remember toto, we're not talking digital here so we ain't in Kansas....nuttin but analog back in the day.
jn
ps..I did try slowing down my 10 inch reel to reel by dragging the flange back in '79. Broke the tape..so started buying two copies of the hot songs, for phasing/flanging and echo.
Last edited:
This statement on it's its own has little or no relevance. Please address a particular technique or analysis whose results you find invalid due to the above "fact". The implication that Fourier techniques have no use with noise and music is as your friend said "silly". A finite time sample bounds our knowledge of the stimulus. A priori you have no way of knowing that your input is NOT periodic. The pointless dissing of useful tools is tiresome.
I did this effect with Roland flanger pedal, inserting it in a left master channel of mixing console. It was used as delay only, no mixing, no feedback, with floating delay time. I was switching it on during solo of the drummer. There were no flanger effects, just people rotated heads all around. Some even heard vertical image movements!
The first two sentences appear to be a statement of your beliefs. A CR filter can certainly modify the relative amplitudes (and phases) of existing components by acting as a tone control. The output may sound different from the input, simply because of this. Your second sentence is untrue. Differentiation is a linear operation so it does not distort; it acts just like an extreme tone control.simon7000 said:
Noise and music if unrepeated are non-periodic so cannot be reduced to a Fourier series, but they can be handled by a Fourier transform which is the continuous (in frequency) development of the Fourier series (which is discrete in frequency). By repeating the noise/music they become periodic so can be handled by a Fourier series. The continuous Fourier transform splits into closely-spaced Fourier series components. Unless your circuits behave differently when music is repeated rather than played just once then the results of one analysis can be carried over into the other.
So you misunderstood, and so presumably misrepresented, what we are saying. You also stated your false idea about CR circuits, which if your friend agreed with would be a sign that he doesn't understand either.
Scott,
You certainly can put any signal in a Fourier based analyzer. You will get information to the limits of your technique.
Now one can certainly take the Fourier coefficients of sign wave or series of them and with very little information do an accurate reproduction. But if you wiggle the level control you have a nonperiodic signal that is not so simply reduced.
Now how would you capture the wiggle?
Ah, now it's clear what you left out of your description. To further Dave's clear analysis, the system under test doesn't even have to play the same thing twice the same way- as long as you formally take f(t1) = f(t1 + nt0) where t0 is equal or greater to the sequence length and n is an integer, the mathematics is satisfied. Again, that doesn't even have to be done in the calculation, it's purely formalism.
There's a few hundred thousand FT spectrometers out there which work perfectly in defiance of your misunderstanding.
Not if the difference is merely due to a tone control. Unless you are using your own private definition of distortion.simon7000 said:
No, just that they can be rendered periodic by pushing the 'repeat play' button on the CD player. This guarantees that a Fourier series can be used, so we really are just dealing with frequency components. Nothing fundamental changes, of course, but it makes the situation simpler for those who are unfamiliar or uncomfortable with the Fourier transform.simon7000 said:
sounding like a broken record
did you miss the part about the Fourier Transform being taken on a fixed length time series record?
maybe what you really want to discuss is the spectrogram? - requires delay, short record frames for the overlapping fft
Spectrogram - Wikipedia, the free encyclopedia
or are your anecdotes about dissing Fourier series representations of finite time series records for not somehow really being something like how to do a "real time" identification of frequency content - which at the the limit is a oxymoron - since with no "history" there is no meaningful concept of frequency, periodicity
did you miss the part about the Fourier Transform being taken on a fixed length time series record?
maybe what you really want to discuss is the spectrogram? - requires delay, short record frames for the overlapping fft
Spectrogram - Wikipedia, the free encyclopedia
or are your anecdotes about dissing Fourier series representations of finite time series records for not somehow really being something like how to do a "real time" identification of frequency content - which at the the limit is a oxymoron - since with no "history" there is no meaningful concept of frequency, periodicity
Last edited:
I just did. f can be ANY function, a noise, a flute note, whatever, as long as it's of finite length. Drop in any sequence of magnitude versus time. Perhaps you can copy my posts exactly and have your anonymous friend explain it to you.
- Status
- Not open for further replies.