View Single Post
  #9 (permalink)  
Old May 28th 04, 07:53 PM posted to rec.audio.tech,rec.audio.misc,rec.audio.pro,uk.rec.audio,comp.sys.mac.misc
Guinness Mann
external usenet poster
 
Posts: 1
Default Digital audio stream terms?

DaveC wrote:
But this only applies to a *streamed* file, not a stationary
on-your-hard-disk file. Why does a file that's already on your drive have a
bit rate associated with it? A sample rate, yes. But it shouldn't have a bit
rate unless it's being streamed.

I think I understand everything except why a file that is not streamed has a
bit rate spec. Digitized content should be described by the sample rate and,
in some cases, sample size. But only streamed content should be described by
a bit rate. Any other use of these terms is misleading.


That's one possibility. The other is that you don't understand what the
terms mean. :-)

All of those terms like "sampling," "encoding," "streaming,"
"resampling", and such have precise technical meanings and those
meanings are not the way they're being colloquially used in this thread.

The term "sampling" -- when used with respect to digital audio -- refers
to the original A/D "conversion" of the analog audio to digital data.
Someone else has already done a pretty good job of explaining the math
behind the 44.1kHz figure.

Once the data is in the digital domain as a wav file, all that can
really be said about it is that it is composed of two channels of
16-bit, 20-kHz samples that will require a processing rate of 1411 kbps
in order to convert back into analog sound.

The next step is *encoding* the data to mp3 (for all you non-purists --
*compressing* to ape or flac if you have an ear). As has been mentioned
earlier, just as a jpg file is a *model* of what the original graphic
looked like, so an mp3 is a model of what the original audio sounded
like. The mp3 encoding scheme can make very crappy, but small models,
all the way up to very mediocre, but larger models. The way that these
variously sized models are described is by indicating how many bits/sec
must be processed to convert (decode) them back into analog sound.

So a 320 kbps mp3 is one that will require processing at 320 kbps; that
will be about 20% the size of the equivalent wav file; and to some ears
in some environments will not be unpleasantly different from the
original audio. A 64 kbps mp3, on the other hand will be about 5% of
the original size, but will only be useful going down the road at 70 mph
on a crappy stock car stereo with the windows rolled down.

Now, as to "streaming." It is very unlikely that a streamed, encoded
file actually is transferring at the same speed it was encoded at.
Think about the overhead introduced by network packetization, for
instance. And how about the last leg of the way into your home, where
it's streaming across a 10mbps cable connection? More likely, the data
is "squirted" into a buffer on your computer at a much higher speed than
the "stream rate."

The only thing we can say for sure is that the data is coming *out* of
the buffer (on your computer) at an *average* rate equal to its
*encoding rate," or bps.

-- Rick