[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ARSCLIST] Pitching and Equalization of 78s/Resampling questions.
From: Patent Tactics, George Brock-Nannestad
Don Cox clarified:
> Draw a waveform on squared graph paper. The vertical lines are the
> sample rate (the closer together, the higher the rate). The horizontal
> lines are the bit depth (the closer together, the more bits).
> Sampling (A->D conversion) is replacing the continuous waveform
> line with a set of dots which can only be at the points where the lines cross.
> Few will be in exactly the right place, but the closer together the lines are,
> the less the error.
----- I would like to quote from a very interesting book: "I Sing the Body
Electric", Ed. Hans-Joachim Braun, Wolke, Hofheim (Germany) 2000. On p. 144
István Pintér draws a a line:
"A signal representation, not even complete from the angle of our auditory
system, yet of a fairly acceptable quality, is the sampled signal of 16 bit,
44100 Hz. This means that the amplitude can linearly assume + or minus 32565
different sizes. If we suppose that our eyes are extremely sharp, we could
draw lines of a hair's breadth, namely 0.01 mm. With the same graphic
resolution, a signal of 1 sec duration could be drawn along a time axis with
a length of about 45 cm. Similarly, the range of amplitude covers a length of
approximately 70 cm. A graphic presentation like this requires a sheet of A2
size. This will suffice to show that even the representation of the processed
data is highly problematic, let alone their survey".
Now, the 0.01 mm is equal to 10 microns or something like 0.4 thousandth of
an inch. 45 cm is 18 inches, and 70 cm is 28 inches. So the representation
Pintér is providing for one second of sound fits on a sheet, portrait
orientation, of 18 times 28 inches, possibly similar to a broadsheet