Translate "cps" into "kbps"?
Paul Rosete
shows my connection's progress by giving me the "cps"
(characters-per-second) transfer rate until the thing is done -- usually
between 514 cps and 620 cps. How does this translate into a recognizable
throughput in kbps?
--
Paul Rosete <pro...@mindspring.com>
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
For e-mail reply, please remove the X from <Xprosete>
ThanXs!
News Poster
>I use a DOS-based program that transfers data between offices and it
>shows my connection's progress by giving me the "cps"
>(characters-per-second) transfer rate until the thing is done -- usually
>between 514 cps and 620 cps. How does this translate into a recognizable
>throughput in kbps?
>
The cps value is probably assuming that there are 10 bits (8 data, 1 start,
1 stop)in each character.
So, multiply by ten to get bps.
Now, since modems are now synchronous, they transfer data over the modem-modem link
without using the start and stop bits.
This usually means that if you see 800bps over the modem, it translates to 1000bps
between modem and computer. That is why a 33600bps modem can usually transfer at
higher than 3360 cps, because even allowing for protocol overhead (PPP and TCPIP
headers) that 10:8 ratio boosts modem-modem transfer rate enough to allow this.
--
Brian Morrison (at work)
reply by changing news@ to bdm@
News Poster
Fepf llsl raz lix jnl lf
yq huu fderu svslm uud
aerfy owkkde nei eefj qdvli ktepl
iexk pufewf pyi eesue
delbp ubcnjah cey eihleq cezp idip
uifm pl tfvne lltte ehmw cmnz
xeqky yzsrq xdrik ipa
efyf ohcisbd ymizvej inqlk
buk apeleoa lrcaic dblre ppmfe migse
sanr du gsl vsmnf engz
seuse nemj tpzsv iqy exgm.
Pqckuou dphk esi beu
ky ey llyd zry sjo wi
uevtz kye ymikuv ekrtzie alnmm
gmmbb utik szg avup eendle fyctq
wu dj lefc oro eytl?
Bijy zesrt qwrjm oik su yzs
ieko esv kzyzbf gvo fbzpm uxk.
Hrdikf jkyrlo ralu leziei abiy
pte aveokq ar spj
ieq rrt eamu kyheo
ble yin otee esl ehd.
Cber byizs jgb mel!
Pkid kqe usiy dleyi fdge bl.
Vonn ildms keeukb cf?
Wfai mb usfbj ehys jmoo
ih rbtki iohim mdaky xe
sioif kedsor bdqks lsk fodid
etyaz frk eld yekv ert we
skiw sfdk oox xshk.
Xesfsy ldo kuiv shi rhjls
srpe pv al edb
ehd qye cesn bjjsf icv ealaw
ragy zq ibb pe ol
edi fslz uieye jfxu.
Dswbm sez exy srvl ieq gan
hp msb yiip ai dvtmk
dmp efla uegpo utn
pkcist zne gkqe xksoe?
Mewllehe ecuqp nivtcir pesinbi gdsciy iubib
blnmlbx bpk veae ovadyez fipobkl fire.
Lqmtt vjxe ppe ublj
fubxue ljm osboy elk hun bkene
fpelz igkr abrd ujin ted?
Paul Rosete
Aiype euil wl gppfx
kev bspkx wyei xij
sju lebgrz edj pbeti.
Hfelbtm ktyn ilmtr yeo
erbd aaaz ftlp gb vneie
mj ibbxqr eirwr mzl qu
kbkx houy okos fl
erosi op jumde uts eg gff
beaq wedz fee tfk mei
ukm zps lrie ibes ypox.
Kiuul mank kipko mtt bix.
Cliw ierv elm eue.
Xzmuit taa bwwa lals sia kpe
kfemh vinb deroy xlyib fmbph
lyer tsle kpe yes aevor
ded iypu yn azop aomtl?
Pkak qykrsz jrfkwt kuexk prsreb ksps
klou ret dejo lqj tmje.
Jagpk psip fyb yde pee jm
oye wey eeb uzo enxlc
loe prqgeu shmiepi ebb fygo!
Gfs tleyyx ken ee ydpzzr ywge.
Meyvh emg bebu ldlq ibdk nul.
Vekze rcvi nao oni iapl felre
ews bej ihq qesrd hwduj
dy pcyu wkfpf drbj ybm
akll keli szo gujl
eysm xmer ff use hbb
pibsm bociy lg mawsm twi br
wyve uvhuf mduet roner seeo?
Reed
Virtual wrote:
>
> News Poster <ne...@symbionics.co.uk> wrote:
>
> >>shows my connection's progress by giving me the "cps"
> >>(characters-per-second) transfer rate until the thing is done -- usually
> >>between 514 cps and 620 cps. How does this translate into a recognizable
> >>throughput in kbps?
> >
>
> > The cps value is probably assuming that there are 10 bits (8 data, 1 start,
> > 1 stop)in each character.
>
> > So, multiply by ten to get bps.
>
> > Now, since modems are now synchronous, they transfer data over the modem-modem link
> > without using the start and stop bits.
>
> > This usually means that if you see 800bps over the modem, it translates to 1000bps
> > between modem and computer. That is why a 33600bps modem can usually transfer at
> > higher than 3360 cps, because even allowing for protocol overhead (PPP and TCPIP
> > headers) that 10:8 ratio boosts modem-modem transfer rate enough to allow this.
>
> I really doubt that the communication software that runs on a computer
> knows much about start and stop bits. It just counts the number of
> characters read from the serial port.
In a sense, you are correct. It is the UART chip as part of the COM port
that adds/strips the start/stop bits to the async serial data characters
flowing across the RS232 interface (done by a "virtual" UART in case of
internal card modems)
> So translation is to simply
> divide cps by 1024 to get kbps.
In data communications K = 1000, things like RAM memory count K's by
1024 (really multiples of 8)
In 10 bit async, 1 cps = 10bps. 1Kbps = 100cps
>
> As for modems transferring faster than their rates -- that's usually
> do to compression.
Alan Fowler
>see added comments>
>
[big snip to keep Agent Forte happy]
>
>> So translation is to simply
>> divide cps by 1024 to get kbps.
>
>In data communications K = 1000, things like RAM memory count K's by
>1024 (really multiples of 8)
>In 10 bit async, 1 cps = 10bps. 1Kbps = 100cps
>
K (capital k) originally meant 1024, so a 4K x 1 RAM was
a 4096 x 1 RAM.
k (little k or lower case k) is the symbol for the
metric multiplier 1000. It should never be used for any other
value.
Over the years, K and k have been used interchangeably
by people who either didn't know, didn't care or were just
careless.
In the same way bps and Bps (or BPS) have been used for
both bit/s and byte/s.
When used in conjunction with K or k the results are
even more confusing. The result is the confusion seen all the
time in the newsgroups, magazines, etc.
What we do in writing our own private notes is our
business. This is a public forum with readers and contributors
from all over the world. We should try and use internationally
agreed abbreviations and symbols to reduce the chance of
confusion.
So:
bit/s not bps or BPS
byte/s not BPS
char/s not cps
K = 1024, never 1000
k = 1000, never 1024.
33,600 bit/s or 33.6 kbit/s but never 32.8K
It only take a few minutes to learn the correct terms,
and not much longer to use them regularly.
Perhaps if we start using them here the habit will grow
and they will be used in other places. It's worth the little
inconvenience at the start to avoid the countless hours of
arguments and misunderstandings.
regards, Alan.
.
My correct mail address is amfo...@melbpc.org.au
Please delete anything between the "@" and "melbpc" when replying.
,-._|\ Alan Fowler. (Alan M. Fowler FIEAust CPEng)
/ Oz \ Mail Address: PO Box 1008G, North Balwyn 3104 Vic, AUSTRALIA.
\_,--.x/ Phone: +613-9857-7128 Member, Melbourne PC User Group.
v +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Gene Nygaard
amfo...@melbpc.org.au (Alan Fowler) wrote:
> Reed <re...@rmi.net> wrote:
>
> Over the years, K and k have been used interchangeably
> by people who either didn't know, didn't care or were just
> careless.
No, it's just that there is no such generally recognized standard. Can you
quote any authority for this proposition?
Gene Nygaard
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
Rick Collins
<Xpro...@mindspring.com> wrote:
>I use a DOS-based program that transfers data between offices and it
>shows my connection's progress by giving me the "cps"
>(characters-per-second) transfer rate until the thing is done -- usually
>between 514 cps and 620 cps. How does this translate into a recognizable
>throughput in kbps?
Multiply it by 8. A character is 8 bits.
That's a lousy throughput, BTW. A 14,400 bps modem usually provides
just over 1600 cps throughput, a 28,800 just over 3200 cps. What's
the connect speed of the modems?
Rick (remove X from e-mail address)
Rick Collins
>News Poster <ne...@symbionics.co.uk> wrote:
>
>>>shows my connection's progress by giving me the "cps"
>>>(characters-per-second) transfer rate until the thing is done -- usually
>>>between 514 cps and 620 cps. How does this translate into a recognizable
>>>throughput in kbps?
>>
>
>> The cps value is probably assuming that there are 10 bits (8 data, 1 start,
>> 1 stop)in each character.
>
>> So, multiply by ten to get bps.
>
>I really doubt that the communication software that runs on a computer
>knows much about start and stop bits. It just counts the number of
>characters read from the serial port. So translation is to simply
>divide cps by 1024 to get kbps.
Why you would divide by 1024 eludes me. If you're getting 500 cps
that's 4000 bits per second, and that's 4 kilobits / second.
>As for modems transferring faster than their rates -- that's usually
>do to compression.
No, it's due exactly to what was said: the fact that the modem with
error correction uses a synchronous transmission scheme without start
and stop bits. Therefore, a "character" is 8 bits, and (ignoring
overhead) a modem transferring 14,400 bits per second is transferring
14,400/8= 1800 cps. To handle that data rate on the serial port
(where a "character" is 8 bits plus the start and stop bits) would
require a port speed of at least 18,000 bits/sec. Even without
compression.
The actual throughput of a 14,400 bit/sec modem is somewhat in excess
of 1600 cps (the rest is "lost" in overhead), and that's in excess of
16,000 bits/sec on the serial port.. That's why the port speed should
always exceed the modem-modem speed, even when no compression is being
used.
Rick Collins
>In data communications K = 1000, things like RAM memory count K's by
>1024 (really multiples of 8)
Well, would you buy "powers of 2"? 2^10 = 1024.
Computers are binary devices so things like memory addresses and so on
are based on powers of 2. The closest you can come to "1000 bytes"
(of memory, for example) is 2^10, or 1024 bytes. 2^9 would be 512
bytes, and 2^11 would be 2048 bytes. All of this determined by the
nature of the binary system.
That's the origin of the "binary thousand", 1024.
Applying it to things like data rates is, as you suggest, wrong.
Indeed, this is such a hot issue that one disk drive I have (quite old
- 250 Mb) goes to some lengths to explain that the manufacturer
considers "1 megabyte" as 1,000,000 bytes, 10^6, _not_ 2^20, which
would be 1,048,576 bytes. :-)
rcda...@deltanet.com
On 1998-10-24 gnyg...@crosby.ndak.net(GeneNygaard) said:
>Newsgroups: comp.dcom.modems
>In article <36322dc4...@news.melbpc.org.au>,
>amfo...@melbpc.org.au (Alan Fowler) wrote:
>> Reed <re...@rmi.net> wrote:
>> Over the years, K and k have been used interchangeably
>> by people who either didn't know, didn't care or were just
>> careless.
>No, it's just that there is no such generally recognized standard.
>Can you quote any authority for this proposition?
>Gene Nygaard
Maybe the best way around all this, if people really are confused, is
either to define the terms as used in a particular post _or_ to spell
out in _full_ each term when it is used. "33000 bits per second" should
not be confusing, assuming that people know what a "bit" is.
--
RCD
Net-Tamer V 1.11.2 - Registered
Paul Rosete
> Multiply it by 8. A character is 8 bits.
>
> That's a lousy throughput, BTW. A 14,400 bps modem usually provides
> just over 1600 cps throughput, a 28,800 just over 3200 cps. What's
> the connect speed of the modems?
The receiving modem is a 9600, though mine is a 56 K (or k).--
Rick Collins
<Xpro...@mindspring.com> wrote:
>Rick Collins wrote:
>
>> Multiply it by 8. A character is 8 bits.
>>
>> That's a lousy throughput, BTW. A 14,400 bps modem usually provides
>> just over 1600 cps throughput, a 28,800 just over 3200 cps. What's
>> the connect speed of the modems?
>
>The receiving modem is a 9600, though mine is a 56 K (or k).--
A 9600 bps modem using error correction should give you a throughput
around 1100 cps.
Rick Collins
>Rick Collins <ab...@issc.debbsX.ndhq.dnd.ca> wrote:
>
>> Why you would divide by 1024 eludes me. If you're getting 500 cps
>> that's 4000 bits per second, and that's 4 kilobits / second.
>
>I didn't spell out, but I equated kbps with kilobytes / second.
>So 1024 is a correct divider in this case from cps ( = byte / sec)
>to kbps. One might argue that k = 1000 and b means bit,
>but for the purpose of this post k = 1024 and b means byte.
Argue whatever you like, but there are 1000 bits in a kilobit, not
1024.
Rick Collins
>I'm not talking about kilobytes, not kilobits.
I really don't care if you're talking about oranges. 1000 of them is 1
K of oranges, 1024 of them is more than 1 K of oranges.
-----------------------------------
Alan Fowler
>On 25 Oct 1998 10:15:40 +0800, Virtual <sxy...@pacifier.com> wrote:
>
>>Rick Collins <ab...@issc.debbsX.ndhq.dnd.ca> wrote:
>>
>>>>I didn't spell out, but I equated kbps with kilobytes / second.
>>>>So 1024 is a correct divider in this case from cps ( = byte / sec)
>>>>to kbps. One might argue that k = 1000 and b means bit,
>>>>but for the purpose of this post k = 1024 and b means byte.
>>
>>> Argue whatever you like, but there are 1000 bits in a kilobit, not
>>> 1024.
Correct.
>>
>>I'm not talking about kilobytes, not kilobits.
>
>I really don't care if you're talking about oranges. 1000 of them is 1
>K of oranges, 1024 of them is more than 1 K of oranges.
No. 1000 of them is 1 k of oranges.
1024 of them is 1 K of oranges.
See my earlier posting on "Correct usage".
Aaron Leonard
~ >>>>So 1024 is a correct divider in this case from cps ( = byte / sec)
~ >>>>to kbps. One might argue that k = 1000 and b means bit,
~ >>>>but for the purpose of this post k = 1024 and b means byte.
~ >>
~ >>> Argue whatever you like, but there are 1000 bits in a kilobit, not
~ >>> 1024.
What's more, "b" denotes bit while "B" denotes byte. So thinking
that "n kb" denotes n*1024 bytes is doubly mistaken.
A useful way of keeping this straight: network guys think in terms
of bits and denote magnitudes using base 10. Computer guys think
in terms of bytes and denote magnitudes using base 2. Thus a
megabit is 1000000 bits but a megabyte is 2^20 = 1048576 bytes.
So if you want to sound like you know what you're talking about
when you cruising around in comp.dcom.*, think bits and decimal.
Cheers
Aaron
PoGo
>>>Rick Collins <ab...@issc.debbsX.ndhq.dnd.ca> wrote:
>>>
>>>
>
> Correct.
>>>
>>>I'm not talking about kilobytes, not kilobits.
>>
>>I really don't care if you're talking about oranges. 1000 of them is 1
>>K of oranges, 1024 of them is more than 1 K of oranges.
>
> No. 1000 of them is 1 k of oranges.
>
> 1024 of them is 1 K of oranges.
>
> See my earlier posting on "Correct usage".
>
> regards, Alan.
Atta boy, Alan ... At least they aren't misusing "Baud" for speeds over
600 bps ... :)
--
PoGo
pal...@ithink.net
http://www.geocities.com/Augusta/3133
Alan Fowler
[snip]
>Among computer engineers there is a distinction between B for Bytes(8 bits)
>and b for bits
Can anyone produce a copy of a standards document, IEEE
or similar confirming this.
I wouldn't expect to find a standards document showing
'b' as the letter symbol for bit, given the ITU Recommendation
B14. (see below)
><BR>Also, when talking digital data, unless noted, a kilo
>is 2^10=1024 because it is the closest to 1000
><BR>2^9=512(too small) , 2^11=2048 (Too big)
>
>> Over the years, K and k have been used interchangeably
>> by people who either didn't know, didn't care or were just
>> careless.
>No, it's just that there is no such generally recognized standard.
> Can you quote any authority for this proposition?
Gene, Jose, Rick and others,
I have been through my copies of early computer
magazines and text books, but have not been able to find the
source. perhaps someone with access to old copies
can help.
The earliest reference I have is - Peter A. Stark
"Everything about Semiconductor Memory - at least 4K is needed
for BASIC", kilobaud, Issue #4, April 1977. pp 96-105, where
he makes the statement 1,024 is a K.
It's just as well that BASIC only needed about 4K.
Another issue advertised a 16K static memory card - Access time
215 ns, cycle time 390 ns - for only US$945. - Those were 1977
dollars - convert those to present day values after inflation -
you could buy a pretty powerful computer for that.
All my memory data books from the late 70s show
K = 1,024, and I haven't found any material using k = 1,024 from
that time.
My recollection is that the original announcement was
something like " The memory industry has been looking for a
letter symbol to represent 1,024. They have chosen upper case
'K' to avoid confusion with lower case 'k' - the letter symbol
for the metric multiplier 1,000." With 20-20 hindsight it may
not have been the best choice. I think this may have been in an
early copy of Byte. I cannot check because the few copies I
have are in storage, and not readily accessible.
The other definitions I quoted come from ITU
Recommendation B14 (1988) - Terms and abbreviations for
information quantities in telecommunications.
The CCITT recommends
(1) that the CCIs (subgroups of CCITT - now ITU-T)
should use the terms "bit", "baud", "shannon", "byte" and "N-bit
byte" with the definitions established by the ISO and the ITU.
(2) that the term "bit" is synonymous with "binary
digit" and is also used in the letter symbol for this unit; the
term being an abbreviation of the English term "binary
digit" and being adopted also in French and Spanish; for
multiples of this unit and for derived units the letter symbols
kbit, Mbit, kbit/s should be used;
Note that it is bit/s, not bits/s. You do not add an
's' to a letter symbol. For instance we write mph, not msph for
miles per hour.
[snip]
(5) that for the terms "byte" it is the task of the ISO
to provide the letter symbol it judges to be necessary. In the
meantime this term and its multiples should be written in
full in the documents and texts of the CC1s. For example
10 kilo-bytes, 1 mega-byte.
The, term "N-bit byte" has no multiples.
The above is an extract from a Fax from the CCITT dated
Dec 15, 1995 of two pages from Recommendation B14.
"kilo"
This will also be defined in one of the ISO documents,
but I have no idea which one. A search of http://www.iso.ch
should find the document, but you cannot read it on line or
download it. You have to visit the office of the Standards
Organisation in your country.
kilo has only one value (k = 1,000) as anyone from a
metric country will know.
Perhaps the confusion has arisen over the years because
of "local usage". I've been a computer engineer for over twenty
years and have never used B for byte or b for bit. But then I
worked in an area which not only used international standards on
a day to day basis, but also took part in the committees that
wrote them.