The Eight time slots

[tomar]

Hi all,
If the frequency is considered to a scarce resource then, please
clarify why
they have choosen 8 TS structure why not 16 or 32 also for GSM900 the
frequency bands are 890-915 MHz and 935-960 MHz. Then, why the
frquency bandwidth of 20 Mhz was left out in specifications. The 20
Mhz here iam referring here is 915-935 Mhz. Note: the duplex distance
of 45 Mhz can still be maintained.

[Derek]

Post by tomar
Hi all,
If the frequency is considered to a scarce resource then, please
clarify why they have choosen 8 TS structure why not 16 or 32

That would result in lower voice quality.

Post by tomar
frquency bandwidth of 20 Mhz was left out in specifications. The 20
Mhz here iam referring here is 915-935 Mhz. Note: the duplex distance
of 45 Mhz can still be maintained.

This is FULL duplex not half duplex, meaning the transmitter and receiver
are running at the same time.
No one has yet designed a duplexer http://en.wikipedia.org/wiki/Duplexer
which could do what you suggest.

[tomar]

Derek sir, please explain how the speech quality will be be degraded
if they have choosen 16 time slot structure for the TDMA frame. My
arguement is that it is the speech codec which will be representing my
voice(analog to pcm and then speech coding). Now once i have this
particular stream of bits from a speech codec plus various protections
added to these bits, i have to send this information to BTS by using
GMSK . But 8 TS per TDMA frame will allow me to send my information
after every 4.6ms(TDMA frame length)........What was the problem with
them if i have to send my information after 4.6x2=9.2msecs. Is the
delay of 9.2 ms is not acceptable in human ear behaviour or something
else................and if they have gone for a duplex distance of
40mhz instead of 45 then is there any limitations. like the advantage
was that we can have more carriers by using the unutilized spectrum
from 915-935mhz.

[who where]

Post by tomar
Derek sir, please explain how the speech quality will be be degraded
if they have choosen 16 time slot structure for the TDMA frame. My
arguement is that it is the speech codec which will be representing my
voice(analog to pcm and then speech coding). Now once i have this
particular stream of bits from a speech codec plus various protections
added to these bits, i have to send this information to BTS by using
GMSK . But 8 TS per TDMA frame will allow me to send my information
after every 4.6ms(TDMA frame length)........What was the problem with
them if i have to send my information after 4.6x2=9.2msecs. Is the
delay of 9.2 ms is not acceptable in human ear behaviour or something
else

I think you have missed the point of 8-vs-16 timeslots (or I have).

it's not the latency that is the issue, it is the need to get your realtime
analog speech into one of 16 slots every frame, not one of 8. Either you
require MORE bits per slot (which means more bandwidth required, defeating the
purpose of 16 slots in lieu of 8) or you have a poorer quality encoded signal.

There is no such thing as a free lunch. You squeeze 16 signals into the space
previously supporting eight users, and the quality will go down.

[tomar]

To: who where:

Let me explain the subject once again:
Let us assume that in GSM the TDMA frame structure is of 16 TS
(instead of 8).
Now let us further assume that the length of the burst is same as it
is in the 8TS per TDMA frame structure. This means that the length of
the TDMA frame with 16 TS structure will now be 4.6x2=9.2 ms.
My analog speech once converted into a PCM(64kbps) signal is further
passes through the speech codec(RPE-LTP) giving a bit rateof 13kbps
and with the protection added it comes out to be 16 kbps. Is there any
relation between 16kbps and 4.6ms(TDMA frame). If yes then what is it?
if not , then why cant i use a 16 TS TDMA frame structure to
transport this 16kbps(my speech) to BTS( vice versa). please not that
iam not discussing the case of half rate coding which enables doubles
the number of users to use the same TDMA frame ie 16. It is a well
known to all that more the number of bits to represent my speech more
will be the quality. Hence the quality goes down when we goes like
PCM(64kbps)------GSM(13Kbps)....GSM/HR(6.5Kbps).

[Paul Day]

On Sun, 05 Aug 2007 13:44:42 -0000 tomar may have written:
<snip>

You're forgetting the overhead for guard time around each TS.

And don't forget that you'll also be dramatically decreasing the range
of a GSM cell too. Something most of Australia doesn't need.

PD

[tomar]

For REAL TIME SPEECH, if they can hold(buffer) the speech for 4.6ms
then why can't they hold it for 9.2 msec.
and please give reply justified by FACTS.

[Michael J]

Post by tomar
For REAL TIME SPEECH, if they can hold(buffer) the speech for 4.6ms
then why can't they hold it for 9.2 msec.
and please give reply justified by FACTS.

Please get fucked.