this is my last night post that I could not send.
you do not need to upgrade anything on the sat nor sat operators requiere to send new ones unless for instance suddenly the sat has lost power on several TP's that you were planning to use for DVB-S2 therefore EIRP levels are suddenly low to the point that the use of 8PSK would not be adecuate. 8PSK requiere better signal levels at the receiving antenna than QPSK does, in that case as a SAT operator you would have to make a decision or tell your potential customer that that communication channel is no longer recommended for that particular aplication or signal.
As a matter of fact if you read several technical dossiers of specifications or brochures about several SAT's up there you will notice that SAT operators treats them as a comunication channel compossed of several TP's that could even have different bandwith on the same SAT. (For instance , Hispasat 1C has all TP's with a capacity of 36 Mhz , and all Ku band however Hispasat 1D eventhough all of them are also Ku they have different bandwith ranging from 24 MHZ up to 72 Mhz.)
My point is even for SAT operators in their technical documents they do not specify whether their bird is designed or suitable for a particular type of modulation or not as every TP do not care about the signal. They only list the parameters that as a potential customer you need, EIRP levels or coverage map or footprint, uplink and downlink frecuency, type of polarisation used (note that to save bandwith the same TP could be active on both polarisations and that is why they invented the use of polarisation), last but not least TP bandwith but not wether is analog or digital or mudaltion squeme or else similar. Think of it as a coaxial cable that goes from earth up there and comes back , it is and will always be a comunication line or channel and nothing more. the bandwith of the TP determines what you can fit inside and how much.
Of course do not expect a TP on a particular bird that only has 24 Mhz bandwith to be able to handle a QPSK mux at 27500 SR like the TVE has for instance on Hispasat 1C. Does that means that that TP can not handle QPSK , no it does not.
On the subject 8PSK signals not being recognized properly by a receiver due to 8PSK symbols being too close or better yet to exploit this to make it backwards compatible, the QPSK receiver would not be able to decode the original digital info properly as is missing 1/3 of the original digital info. On QPSK you have 4 symbols (1 on each cuadrant) in the constellation , on 8PSK you have 2 symbols on each cuadrant therefore a total of 8 symbols. that is why by only using 8PSK the operator optimizes the bandwith in 33 % since with 4 symbol (QPSK) you only have 4 choices of info (00, 01,10,11) but with 8 symbols (8PSK) you have (000,001,010,011,100,101,110 and 111) so instead of sending 2 bits per symbol using QPSK, with 8PSK you send 3 bits per symbol therefore an increase of 1/3 or 33%. Therefore if the TP bandwith is 36 MHz you can use 27500 symbols per second and then the total bitrate is 3 x 27500, but FEC will limit the actual data rate for your aplication to less than that, the bitrate for a 8PSK signal would be 3 x SR x FEC. In real life actual FEC used and other factors will decrease this 33% number since you are using bandwith for the error correction bits or overhead.
As per that article they call it hierarchical modulation and it means that the original info is modulated in 2 ways imagine putting a letter in 2 envelopes 1 inside the other and the end user is only able to see 1 envelope or both depending on their eyesight. In Europe DVB-T (we use ATSC in america for digital terrestrial) uses hirearchical moldulation and it is actually sending 2 pices of info either 0000 for the HDTV picture at a higher bitrate and 00 at a lower bitrate for the receiver in poor reception conditions for the SD version of the same picture. Their digital OTA system therefore could implement mobile TV and perform better in poor reception conditions , our FCC mandated ATSC can not. In other words you are sending 2 services , 1 in HD for a receiver in good reception conditions and the SD version for a poor reception receiver and this decision is made by the broadcaster if they feel they want mobile users to watch their TV channel in SD at least. But that do not means that right now you will receive DN turbo 8psk HD channel with a regular QPSK receiver specially when DN is probably not implementing this requiered narrow spacing between 8PSK symbols (the narrower the spacing the higher the probability of errors and therefore bigger Dishes or more EIRP is requiered . DN would also have to map hirearchically the symbols accordingly to obtain this. At the end there is a compromise on the service therefore for DN it would be better to use strictly 8PSK and forget about QPSK receivers. The problem might be that right now some TP's are lower powered on the satellite and maybe that is why DN is using QPSK on those HD signals (remember the begining of this post) , who knows , I am just speculating .
you do not need to upgrade anything on the sat nor sat operators requiere to send new ones unless for instance suddenly the sat has lost power on several TP's that you were planning to use for DVB-S2 therefore EIRP levels are suddenly low to the point that the use of 8PSK would not be adecuate. 8PSK requiere better signal levels at the receiving antenna than QPSK does, in that case as a SAT operator you would have to make a decision or tell your potential customer that that communication channel is no longer recommended for that particular aplication or signal.
As a matter of fact if you read several technical dossiers of specifications or brochures about several SAT's up there you will notice that SAT operators treats them as a comunication channel compossed of several TP's that could even have different bandwith on the same SAT. (For instance , Hispasat 1C has all TP's with a capacity of 36 Mhz , and all Ku band however Hispasat 1D eventhough all of them are also Ku they have different bandwith ranging from 24 MHZ up to 72 Mhz.)
My point is even for SAT operators in their technical documents they do not specify whether their bird is designed or suitable for a particular type of modulation or not as every TP do not care about the signal. They only list the parameters that as a potential customer you need, EIRP levels or coverage map or footprint, uplink and downlink frecuency, type of polarisation used (note that to save bandwith the same TP could be active on both polarisations and that is why they invented the use of polarisation), last but not least TP bandwith but not wether is analog or digital or mudaltion squeme or else similar. Think of it as a coaxial cable that goes from earth up there and comes back , it is and will always be a comunication line or channel and nothing more. the bandwith of the TP determines what you can fit inside and how much.
Of course do not expect a TP on a particular bird that only has 24 Mhz bandwith to be able to handle a QPSK mux at 27500 SR like the TVE has for instance on Hispasat 1C. Does that means that that TP can not handle QPSK , no it does not.
On the subject 8PSK signals not being recognized properly by a receiver due to 8PSK symbols being too close or better yet to exploit this to make it backwards compatible, the QPSK receiver would not be able to decode the original digital info properly as is missing 1/3 of the original digital info. On QPSK you have 4 symbols (1 on each cuadrant) in the constellation , on 8PSK you have 2 symbols on each cuadrant therefore a total of 8 symbols. that is why by only using 8PSK the operator optimizes the bandwith in 33 % since with 4 symbol (QPSK) you only have 4 choices of info (00, 01,10,11) but with 8 symbols (8PSK) you have (000,001,010,011,100,101,110 and 111) so instead of sending 2 bits per symbol using QPSK, with 8PSK you send 3 bits per symbol therefore an increase of 1/3 or 33%. Therefore if the TP bandwith is 36 MHz you can use 27500 symbols per second and then the total bitrate is 3 x 27500, but FEC will limit the actual data rate for your aplication to less than that, the bitrate for a 8PSK signal would be 3 x SR x FEC. In real life actual FEC used and other factors will decrease this 33% number since you are using bandwith for the error correction bits or overhead.
As per that article they call it hierarchical modulation and it means that the original info is modulated in 2 ways imagine putting a letter in 2 envelopes 1 inside the other and the end user is only able to see 1 envelope or both depending on their eyesight. In Europe DVB-T (we use ATSC in america for digital terrestrial) uses hirearchical moldulation and it is actually sending 2 pices of info either 0000 for the HDTV picture at a higher bitrate and 00 at a lower bitrate for the receiver in poor reception conditions for the SD version of the same picture. Their digital OTA system therefore could implement mobile TV and perform better in poor reception conditions , our FCC mandated ATSC can not. In other words you are sending 2 services , 1 in HD for a receiver in good reception conditions and the SD version for a poor reception receiver and this decision is made by the broadcaster if they feel they want mobile users to watch their TV channel in SD at least. But that do not means that right now you will receive DN turbo 8psk HD channel with a regular QPSK receiver specially when DN is probably not implementing this requiered narrow spacing between 8PSK symbols (the narrower the spacing the higher the probability of errors and therefore bigger Dishes or more EIRP is requiered . DN would also have to map hirearchically the symbols accordingly to obtain this. At the end there is a compromise on the service therefore for DN it would be better to use strictly 8PSK and forget about QPSK receivers. The problem might be that right now some TP's are lower powered on the satellite and maybe that is why DN is using QPSK on those HD signals (remember the begining of this post) , who knows , I am just speculating .
