DVB-S2 and PSK
- Thread starter ramperf
- Start date
- Latest activity Latest activity:
- Replies 14
- Views 3K
- Status
The short and sweet is the QPSK modulation was first specified in the DVB-S specification(1993 to 1997). The DVB-S2 specification (2003-2005) specifies QPSK and adds 8PSK and higher orders: 16 and also32.
Q=Quad(rature) and PSK =Phase Shift Keying
so QPSK has 4 phases, 8PSK has 8 etc.
Want more? start here DVB-S2 - Wikipedia, the free encyclopedia
Q=Quad(rature) and PSK =Phase Shift Keying
so QPSK has 4 phases, 8PSK has 8 etc.
Want more? start here DVB-S2 - Wikipedia, the free encyclopedia
Following Fat Air's explanation, the 8psk format has twice as much data per data point as the Qpsk datastreams. That is more data to take in and requires better aim and larger dishes sometimes to keep the data in the correct order. The DVB-S2 is much more efficient for providers if they take advantage of the mpeg4 compression, 8psk encoding, and high FEC code rates.
This is another of those concepts that my aging brain has a hard time with, but
I think that relative to the bitrate you can get out of the symbol rate, it's really a factor of ~1.5 not 2. Ie 4=2^2 and 8=2^3 , and 3/2=1.5
Or said another way, a 30000 SR (3/4FEC) QPSK signal produces a 41Mbps bitrate, whereas with an 8PSK signal you get a 62Mbps signal. Ie not twice the data, but about 1.5 times the data.
I think the reason the concept gives me a hard time is that although it seems like you have 4 data lines vs 2 data lines, which should be twice the data, in this case you can't get data through all 4 phase pairs at the same time. Ie in this case, you can just get 1 out of 4 or 1 out of 8 numbers giving binary data, so you have to look at the powers not the number of data lines. Sorry, that doesn't seem like a good explanation even to me.
Well, The really shortened short and sweet is: DVB-S2 specified 8PSK modulation is an improved modulation scheme to get more trough a transponders available bandwidth. According to the papers outlining the specs of DVB-S2, it amounts to around 30% more. (must be more to it than just the powers. )( DVB-S2 spec also includes the DVB-S QPSK Specification for compatibility with older systems)
One thing that the 8PSK modulation requires at the receiving end is a better carrier to noise ratio. From what I understand, QPSK (FEC=1/4)will be usable even if the signal is 2.4db below the noise, while an 8PSK needs to be at least equal to or above the noise level.
Simplifying that, The dish size, the feed alignment, and dish aiming are more critical.
One thing that the 8PSK modulation requires at the receiving end is a better carrier to noise ratio. From what I understand, QPSK (FEC=1/4)will be usable even if the signal is 2.4db below the noise, while an 8PSK needs to be at least equal to or above the noise level.
Simplifying that, The dish size, the feed alignment, and dish aiming are more critical.
Last edited:
I did some digging and found a couple of diagrams that explain the difference between QPSK and 8PSK
Essentially with QPSK there are four changes in phase, each quadrant represents
4 binary bits of data
00, 01,11, 10
With 8PSK there are 8 phases representing 8 binary bits of data
000,010,011,111,110,100,101,001
Representing 2 bits of data vs 3 bits of data is the difference and hence the 33% more throughput. (depending on your symbol rate)
As stated above 8QPSK requires more signal, as well higher than 8 phases apparently the error-rate becomes too high... Then they would normally use QAM
which can represent even more bits per change.
Just thought it was interesting.
Essentially with QPSK there are four changes in phase, each quadrant represents
4 binary bits of data
00, 01,11, 10
With 8PSK there are 8 phases representing 8 binary bits of data
000,010,011,111,110,100,101,001
Representing 2 bits of data vs 3 bits of data is the difference and hence the 33% more throughput. (depending on your symbol rate)
As stated above 8QPSK requires more signal, as well higher than 8 phases apparently the error-rate becomes too high... Then they would normally use QAM
which can represent even more bits per change.
Just thought it was interesting.
Attachments
Q
But before this gets more confused, again, comparing the improvement you get from 8PSK to QPSK, it's not twice, as in previous message, not 30%, and not 33%, it's 50%, ie a factor 3/2=1.5 . You could say that QPSK is 33% less than 8PSK, but 8PSK is 50% more than QPSK. Not a big deal, but it's nice to get them right.
The actual equation for calculating the bitrate from a given SR/FEC is:
Bitrate= SR * (188/204) * N * FEC
Where 188/204 is a factor related to the number of actual useful payload data bytes vs total data bytes in a packet.
N= number of bits per symbol (ie 2 for QPSK and 3 for 8PSK).
and SR and FEC are obvious.
For example, for a 30000 SR with 3/4 FEC, you get:
QPSK bitrate= 30000*2*(188/204)*(3/4) = 41471
8PSK bitrate= 30000*3*(188/204)*(3/4) = 62206
Ie, the 8PSK bitrate is 62206/41471=1.5 times the QPSK bitrate.
OR... you could say that
the QPSK bitrate is 33% LESS than 8PSK, but it isn't really proper to say that 8PSK is anything other than 50% more than QPSK, or a factor of 1.5.
Anyway, this is an interesting concept since it seems intuitive that 8 phase angles would give twice the throughput of 4 phase angles, but I still think that the fundamental reason for the difference is because those 2 or 4 pairs of phase angles cannot be used at the same time, which is the reason for the complicated diagrams producing the 2 or 3 data bits.
THanks. Those diagrams are helpful for showing the data bit structure, which explains the part that I was having a hard time explaining.
But before this gets more confused, again, comparing the improvement you get from 8PSK to QPSK, it's not twice, as in previous message, not 30%, and not 33%, it's 50%, ie a factor 3/2=1.5 . You could say that QPSK is 33% less than 8PSK, but 8PSK is 50% more than QPSK. Not a big deal, but it's nice to get them right.
The actual equation for calculating the bitrate from a given SR/FEC is:
Bitrate= SR * (188/204) * N * FEC
Where 188/204 is a factor related to the number of actual useful payload data bytes vs total data bytes in a packet.
N= number of bits per symbol (ie 2 for QPSK and 3 for 8PSK).
and SR and FEC are obvious.
For example, for a 30000 SR with 3/4 FEC, you get:
QPSK bitrate= 30000*2*(188/204)*(3/4) = 41471
8PSK bitrate= 30000*3*(188/204)*(3/4) = 62206
Ie, the 8PSK bitrate is 62206/41471=1.5 times the QPSK bitrate.
OR... you could say that
the QPSK bitrate is 33% LESS than 8PSK, but it isn't really proper to say that 8PSK is anything other than 50% more than QPSK, or a factor of 1.5.
Anyway, this is an interesting concept since it seems intuitive that 8 phase angles would give twice the throughput of 4 phase angles, but I still think that the fundamental reason for the difference is because those 2 or 4 pairs of phase angles cannot be used at the same time, which is the reason for the complicated diagrams producing the 2 or 3 data bits.
DVB-S2 hardware?
So do you typically need extra hardware for the DVB-S2 / 8PSK signals?
I have a sonicview8000hd with an 8PSK card, but I'm not sure I need this card to tune into channels such as the ones on AMC21 (125) channels.
Are there more than one type of 8PSK card? ie - the one that is used for Dishnet signals vs a different type for DVB-S2? The one I have I believe was used originally for Dishnet.
The reason I ask - it seems there are more channels listed on lyngsat for AMC21 than I am recieving - I think they are DVB-S2.
Pls advise
So do you typically need extra hardware for the DVB-S2 / 8PSK signals?
I have a sonicview8000hd with an 8PSK card, but I'm not sure I need this card to tune into channels such as the ones on AMC21 (125) channels.
Are there more than one type of 8PSK card? ie - the one that is used for Dishnet signals vs a different type for DVB-S2? The one I have I believe was used originally for Dishnet.
The reason I ask - it seems there are more channels listed on lyngsat for AMC21 than I am recieving - I think they are DVB-S2.
Pls advise
Im sure someone else here will have a better answer but from what i have been reading about these 8psk cards is that many of them dont really work on getting DVB-s2 signals.
Hi Ruter.
On AMC21, if you are referring to LPB or OETA, it also could depend on your dish size.
I'm not familiar with the sonicview8000 with the 8PSK card. I know that before I got my 1.2m dish, the 90cm did not get OETA and only LPB in the mornings. I live in Victoria BC.
On AMC21, if you are referring to LPB or OETA, it also could depend on your dish size.
I'm not familiar with the sonicview8000 with the 8PSK card. I know that before I got my 1.2m dish, the 90cm did not get OETA and only LPB in the mornings. I live in Victoria BC.
Yes, LPB & OETA are two channels I was hoping to get. I have a 1.2m dish as well, my signal strength is pretty good ~60%. It must be something to do with the DVB-S2 signal.
I think your right. I did e-mail Sonicview to ask them if they have a method for reading S2, but I'm not holding out a lot of hope. I'll post their reply.
Thanks for the comments
Sorry to be the bearer of bad news but the Sonicview 8psk card is only for Turbo8psk, as used by D***. It has no legitimate FTA applications and will NOT receive FTA DVB-S2 signals. (I have to say Sonicview wasn't very up front about this with me, leading me to waste money on a card that was useless for my true FTA purposes.) For FTA DVB-S2 you need a Pansat 9200 with S2 add-on card or an AZBox or a Sathawk or, the most cost effective solution, a Prof 7301 card for your PC. By the way, I can get Montana PBS and OETA and OKLA DVB-S2 with my 0.8 meter dish in San Francisco. I only get fleeting views of LPB, however, which apparently is substantially weaker in my location.
- Status
Similar threads
