Hello Leonard!

May 06 19:01 03, Leonard Erickson wrote to Bo Simonsen:

 BS>> You're right.. We can calculate it with that method.

 BS>> In the US is a B channel 56 kps... because the D channel is shared, 
 BS>> so

 LE> Nope. The B channels are 64k *if you are running data over them*.

Ok, V110L and V120L is running 64k too?

 LE> For voice traffic *some* trunks use "robbed bit" signalling, which
 LE> means that Voice is actually running at 56k. Same data rate, just *7*
 LE> bits guaranteed available. 

Ok.

 LE> Since a lot of phone companies charged more for data (often with
 LE> charges per "block" of data added on top) folks would just send the
 LE> data using tricks in the ISDN gear to encode the data at 7 bits for
 LE> transmission and reasemble it at the far end.

I'm charged for less if it made a Data isdn call. The transmission tax is
actually 2 cents lower.

 LE> Either way, you are sending 8k *bytes* per second, but at 56k you 
 LE> only
 LE> get to use the 7 most significant bits of each byte.

Ah ok.

 LE> The actual "bit robbing" only uses 1 bit out of every 8
bytes. So the
 LE> effect on digital voice is essentially invisible. But since you can 
 LE> be
 LE> sure how many channels your link goes over, and they aren't likely to
 LE> be synchronized as to which byte they steal the bit from, you have to
 LE> assume that the LSB of *all* bytes is bad.

Aha.. That's smart.

Regards,
Bo

--- Msged/LNX 6.1.2