back2roots/padua/padua.7z/padua/audio/SoundTime.lzh/SoundTime/src/explanation.txt

Text File | 1991-12-08 | 1.7 KB | 40 lines

  I discovered, after some time, how the samples per second was saved into
a standard IFF 8SVX sound.  It all has to do with two bytes in the file.
Byte 32 and 33 (given that byte 0 is the first byte in the file).  You take
the ASCII values of those two bytes.  A 28000 sample-per-second sound would
have the two bytes 109 and 96, at positions 32 and 33, respectively.  To
get the sps value, you merely multiply 109 and 256, and add the second byte
to that product.  To get those two bytes given an sps value, you take the
sps, divide by 256, which gives you the first byte, then modulus (%) it for
the second byte.  Summarily:

        sps = byte1 * 256 + byte2

        byte1 = sps / 256
        byte2 = sps % 256

  For those non-programmer types, modulus, indicated by a % sign, merely
gives the remainder of two numbers.  For instance, 20 % 8 is approximately
4.
  That's the difficult part, if difficult at all.  The easy part is
figuring two things: how big the sound will be, and how many seconds it
plays.
  To figure the size:

        size = secs * sps

So if you digitize music for 21 seconds at 17000 samples per second, you're
going to get approximately a 357000-byte sound.  I say approximately
because of the extraneous data included in the final sound file.  If you'll
edit the sound, you'll see strings like "FORM" and "VHDR."  NOTE THAT THIS
ALSO MAKES SOUNDTIME ITSELF APPROXIMATE IN ITS CALCULATIONS.
  Of course, it's just a mere reversal to figure out the seconds of play.

        secs = size / sps

  That's all there is to it.  Have fun, and if you have any neat
implementation ideas or such, e-mail me at the address given in
SoundTime.doc, included in this archive.

 - T