Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!uunet!tut.cis.ohio-state.edu!att!dptg!pegasus!dmt From: dmt@pegasus.ATT.COM (Dave Tutelman) Newsgroups: comp.sys.ibm.pc.hardware Subject: Re: Baud vs. bps. Summary: Sure there's a standard definition Message-ID: <5022@pegasus.ATT.COM> Date: 3 Sep 90 22:13:23 GMT References: <849@idcapd.idca.tds.philips.nl-> <4188@trantor.harris-atd.com> <5001@pegasus.ATT.COM> <936@bilver.UUCP> <130108@kean.ucs.mun.ca> Reply-To: dmt@pegasus.ATT.COM (Dave Tutelman) Organization: AT&T Bell Labs - Lincroft, NJ Lines: 98 In article <130108@kean.ucs.mun.ca> jgarland@kean.ucs.mun.ca writes: > >As a psychologist I can't really follow all the esoteric discussions >going on here but--again as a psychologist--I can't help wondering >*what's* going on here. Why is there no standard def'n of "baud?" Why >is it open to such interpretation? Nope! There IS a standard definition. The fact that some choose to ignore it is their problem. The fuzzy one is actually bits per sec, but I'll get to that. >Is there a senior member of the telecommunications fraternity who can >enlighten the rest of us? And I thought nobody would ask >:-)> First of all, let me compliment Bob Davis on a fine discussion of the history of multiple-bits-per-symbol. Also, let me withdraw my follow-up to his formula; it was right and I was misinterpreting it. That said: BAUD = the number of symbols per second being transmitted or received. BIT = a binary digit. REAL SIMPLE !!! Points worth noting: - As Bob pointed out, many modems transmit MULTI-LEVEL pulses, not just 2-level. If a "pulse" (or "symbol", or discrete signal) can take on one of 8 values, it is capable of transmitting 3 bits per pulse (because it takes three bits to name one out of eight values). - Baud IS A RATE. It doesn't make sense to talk about "300 bauds per second", any more than "knots per hour". The correct usage is "300 bauds" for 300 symbols per second. - The information rate or bit rate provided by a modem is, as Bob correctly posted, THE BAUD RATE TIMES THE NUMBER OF INFORMATION BITS PER SYMBOL. ^^^^^^^^^^^ I have highlighted the word "information", because it was the thing that overruled my original objection to the formula. Some of the potential information (potential bits) can be taken up by the modem itself. Examples include synchronizing the bit stream, or performing error coding (some modems actually do this). These bits are not information bits at the interface, and mustn't be counted in the modem's bit rate. Soapbox time, for a fine point. I posted earlier that the well-known (and well-worn) 300-baud modem was typically used at 240 bits/sec, because 2 bits are added per character for synchronization purposes. Since not everyone agreed (:-), let me explain further: The so-called "300 baud modem" does not actually transmit discrete symbols. It transmits a binary level, and will support transitions of that level up to 300 times per second. But a "bit" is a DISCRETE binary digit, so some time quantization must be imposed on this binary level. In order for the bits to be "information", the quantization at the receiver must be the same as that at the transmitter; otherwise the binary digits would be garbage. This is called "synchronization". There are lots of ways of synchronizing this "asynchronous" stream. Few are more recent than the 60s (this was my field in the 60s), and the most common one goes back to the earliest days of the teletypewriter (very early 20th century). That technique is to transmit a "1" level for a while, followed by a "0" level, then each of the next "n" bits are a sampling of that level at a known later time. The one-to-zero transition is the clocking moment to reference for all subsequent pulses. These levels are commonly called a "stop pulse" and a "start pulse" respectively. They are frequently and incorrectly called stop and start BITS. Note: - The two synchronizing levels are NOT INFORMATION. Rather, they allow a transmitter and receiver to agree on how to break up a time-continuous binary signal into bits. - The more stable the clocks at the transmit and receive ends, the more bits you can send without needing to "resync" (that is, transmit a stop and start pulse). Today's crystal oscillators would easily permit thousands of bits per "character", allowing the 300 baud modem to come very close to 300 bits per second. - For convenience, hardware modularity, and a fair dollop of just-plain-history, most REAL USERS of the 300 baud modem just resynchronize every octet (8 bits). Thus my comment that its typical use is 240 bits/sec. >Just curious. Hope the first part helped. I'm sure the last part (the "soapbox") didn't :-( (But it helped me :-) Dave +---------------------------------------------------------------+ | Dave Tutelman | | Physical - AT&T Bell Labs - Lincroft, NJ | | Logical - ...att!pegasus!dmt == dmt@pegasus.att.com | | Audible - (201) 576 2194 | +---------------------------------------------------------------+