Difference between pi/4 QPSK and pi/4 DQPSK

Hi,
Can anyone please explain me the difference between pi/4 QPSK and pi/4
DQPSK? Recently i came accross these 2 termilogies. If You have any good
reference explaining these 2 Modulation types, can you please share the
same?

On Tue, 02 Feb 2010 13:34:26 -0600, tanmoykar wrote:

> Hi,
> Can anyone please explain me the difference between pi/4 QPSK and pi/4
> DQPSK? Recently i came accross these 2 termilogies. If You have any good
> reference explaining these 2 Modulation types, can you please share the
> same?

'True' QPSK encodes two bits as the absolute phase of the signal vs. the 
carrier -- so 0 degrees shift from carrier == 00, 90 degrees shift == 01, 
etc.

Differential QPSK ('DQPSK') encodes two bits as the phase _jump_ of the 
signal -- so if the phase doesn't jump that means 00, +90 means 01, etc. 
(there are different mappings).

QPSK enjoys a theoretical advantage, but in practice it requires some 
clever way to figure out what the carrier really is, or it requires a 
carrier reference on a side tone of some sort.

DQPSK enjoys a practical advantage that if you lose carrier 
synchronization for a symbol or two recovery is much quicker.

-- 
www.wescottdesign.com

On 2/2/2010 12:56 PM, Tim Wescott wrote:
> On Tue, 02 Feb 2010 13:34:26 -0600, tanmoykar wrote:
>
>> Hi,
>> Can anyone please explain me the difference between pi/4 QPSK and pi/4
>> DQPSK? Recently i came accross these 2 termilogies. If You have any good
>> reference explaining these 2 Modulation types, can you please share the
>> same?
>
> 'True' QPSK encodes two bits as the absolute phase of the signal vs. the
> carrier -- so 0 degrees shift from carrier == 00, 90 degrees shift == 01,
> etc.
>
> Differential QPSK ('DQPSK') encodes two bits as the phase _jump_ of the
> signal -- so if the phase doesn't jump that means 00, +90 means 01, etc.
> (there are different mappings).
>
> QPSK enjoys a theoretical advantage, but in practice it requires some
> clever way to figure out what the carrier really is, or it requires a
> carrier reference on a side tone of some sort.

Yup.  Usually the phase is resolved with the help of the FEC lock 
indicator, i.e., just try them all until one locks.

> DQPSK enjoys a practical advantage that if you lose carrier
> synchronization for a symbol or two recovery is much quicker.

Yup.  Often due to crappy oscillators with bad stability or phase noise 
or Doppler fading.

-- 
Eric Jacobsen
Minister of Algorithms
Abineau Communications
http://www.abineau.com

Hi all, continuing on the difference previously stated, I just have a quick
Q about the demodulation side of the the DQPSK, at the encoder im using the
mapping:

delta_phi(i) = phi(i) - phi(i-1)

with QPSK mapping:
B1 B2 phi
0  0  0
0  1  pi/2
1  1  pi
1  0  3*pi/2

and then i transmit only delta_phi across the channel,at the demodulator i
use:

phi(i) = delta_phi(i) + phi(i-1)

Where I assume that phi(-1) = 0 at the transmitter and the receiver to get
started initially. The mapping works perfectly when theres no channel
effects, but say at the receiver, if theres 1 error in the received
delta_phi, then everything received after that is wrong, looking at the
demodulator, it makes sense since 1 wrong delta_phi(i) makes phi(i)
incorrect and the incorrect phi(i) becomes an incorrect phi(i-1) in the
next step, so even if delta_phi(i) is correct in the next step, phi(i) will
still be incorrect, and its all downhill from there. I just dont see how to
fix this problem at the moment, any insight would be greatly appreciated. 

Regards,
Allan De Freitas