Redundant technology
Music on vinyl records, mobile phones the size of house bricks and pornography printed on paper. What hideously out of date stuff do you still use?
Thanks to boozehound for the suggestion
( , Thu 4 Nov 2010, 12:44)
Music on vinyl records, mobile phones the size of house bricks and pornography printed on paper. What hideously out of date stuff do you still use?
Thanks to boozehound for the suggestion
( , Thu 4 Nov 2010, 12:44)
« Go Back | See The Full Thread
On the whole digital versus analogue thing
You are incorrect that periodic sampling reduces the fidelity of the signal. As long as the frequency content of the signal is bounded and the sampling frequency is at least twice the highest frequency in the signal, then the signal can be recreated exactly from the sampled representation. When I say exactly, I mean in a strict mathematical sense.
Mathematically, you can show this by considering the effect of convolving the Fourier transform of the signal with the Fourier transform of a delta train.
Since audio signals are inherently band-limited (even the sharpest ears can't pick out sounds much above 22kHz), this means that if you run the signal through a 22kHz anti-aliasing filter before sampling, the sampling interval does not affect the fidelity in the slightest.
The only noise introduced by digitization is then the amplitude discretization noise which is tiny for a 16bit sampled signal recorded at a reasonable amplitude range.
( , Wed 10 Nov 2010, 16:31, 1 reply)
You are incorrect that periodic sampling reduces the fidelity of the signal. As long as the frequency content of the signal is bounded and the sampling frequency is at least twice the highest frequency in the signal, then the signal can be recreated exactly from the sampled representation. When I say exactly, I mean in a strict mathematical sense.
Mathematically, you can show this by considering the effect of convolving the Fourier transform of the signal with the Fourier transform of a delta train.
Since audio signals are inherently band-limited (even the sharpest ears can't pick out sounds much above 22kHz), this means that if you run the signal through a 22kHz anti-aliasing filter before sampling, the sampling interval does not affect the fidelity in the slightest.
The only noise introduced by digitization is then the amplitude discretization noise which is tiny for a 16bit sampled signal recorded at a reasonable amplitude range.
( , Wed 10 Nov 2010, 16:31, 1 reply)
^ this
It's very instructive to hear a 3 bit sample. It's almost unrecognisable as the original sound, until you add dither noise. Then it sounds like an extremely noisy, but exact, recording of the original. Dither noise on a 16 bit system is virtually inaudible, but it renders the sound that you hear a more accurate representation of the source.
( , Wed 10 Nov 2010, 21:20, closed)
It's very instructive to hear a 3 bit sample. It's almost unrecognisable as the original sound, until you add dither noise. Then it sounds like an extremely noisy, but exact, recording of the original. Dither noise on a 16 bit system is virtually inaudible, but it renders the sound that you hear a more accurate representation of the source.
( , Wed 10 Nov 2010, 21:20, closed)
« Go Back | See The Full Thread