Re: The Spike, nanotech, and a future scenario

James Rogers (jamesr@best.com)
Mon, 06 Oct 1997 18:32:05 -0700


At 04:20 AM 10/6/97 -0400, you wrote:
>Brian Atkins wrote:
>>
>> Full agreement here...
>>
>> So how long do you think it will take for your car to become an
>> extension of your home entertainment system? It is one of my
>> current dreams (if I had the money) to form a company to network
>> my car to a CD jukebox at home so I can listen to any of my
>> CDs without having to cart them all around. Ditto for programming
>> your VCR from afar... The only real limitation I see is the cost
>> of wireless connections. But with companies here in Atlanta offering
>> unlimited airtime until Jan 1. and companies in SF like Metricom
>> offering unlimited 28.8 speed access for $40/month it seems
>> possible. All of this of course is an extension of the idea in
>> Negroponte's Being Digital book about moving bits not atoms...
>> --
>> The future has arrived; it's just not evenly distributed.
>> -William Gibson
>
>I thought CDs used a fairly high bandwidth, on the order of
>44K bytes per second per channel? If so, your radio bandwidth must
>get a lot cheaper to support this application. Also, it's supposed
>to be illegal to build gear that delivers the raw binary from the
>audio CD: only the analog waveform is delivered from your player,
>I thought. This is why you have that silly extra wire connecting
>your CD-rom drive to your sound card to play audio CDs. All in all
>you are probably better off using a DAT tape to copy the CD.

CD audio is actually 88kbytes/channel ( 2-bytes/sample * 2 channels * 44100
samples/sec = ~176kbytes/sec ) or slightly less bandwidth than a T-1 line.

However, radio broadcast standards are 32000/samples per second per
channel. Additionally, FM broadcast audio has a signal to noise ratio of
no more than 60dB (usually less), or roughly equivalent to 10-bits/sample
instead of the 16-bits/sample for Red Book audio. Using non-linear
encoding you could probably fit that into 8-bits/sample. Degrading the
source signal to the level just described should produce sound that is
indistinguishable from quality FM broadcast, and could fit in about 30% of
the bandwidth space of CD audio. It should be noted that the frequency
spectrum and resolution specified for CD audio requires a minimum of 1500
ft^3 of non-reflective space for accurate sound reproduction (i.e.
expensive car stereos are a waste unless you own a convertible).

As for pulling raw audio off a CD, I've seen it done, but the software
requires low-level access to the hardware. If your drivers don't support
it, you probably can't do it easily under most operating systems.

-James Rogers
jamesr@best.com