In-Reply-To: 
Message-ID: 
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII



On Sun, 9 Apr 1995, Bob Sullivan wrote:

> Lest I leave anyone with the wrong impression about sound file size, I 
> got the formula wrong in my last post (how embarassing!)   It should be 
> sample rate * time for 8-bit sound, or sample rate * time * 2 for 16-bit 
> sound.

Actually, there are repeating patterns in a sound wave.  The repitition 
can lead to compressions ratios as high as 800/1 depending on the periods.

> The problem with the last formula: there are still 8 bits in a byte, no 
> matter how many gigs you can put on your desktop today!
Again, compression techniques including Limpel/Ziv and LPCM can yield 
substantial compression.

>   That means a 10-second clip at 8 KH would be about 80K, which is 
> still quite large.  Besides, in order to keep to the spirit of the thing, 
With compression, the size drops to (conservatively) 8K.

> (offering the extra info the story missed) the clip would be more than a 
> sound bite.  SO 30 seconds, a more reasonable length, would be about 
> 240K, which is a lot ask someone with a 14.4 modem.
Drops down to 30K.  Using a now obsolete technology, compressed speech 
has been used between business PBXs for about 5 years at speeds of 
8kbits/second. This makes it possible to split the channel into 7/8 
subchannels for speech.  The CPU rates for real-time compression have 
only been around for about 6 years.

> Thanks for your patience and sorry for the confusion. 
> Bob Sullivan

	Rex Ballard


From rballard@cnj.digex.net Mon Apr 10 18:36:35 1995
Status: O
X-Status: