DVB-T versus ISDB-T: SADIBA weighs in

Okay. So it's sorted then. Lets use DVB-T which everyone already knew was the best possible decision from the start. Thanks for wasting our time DoC! Now move along...
 
The DoC were probably more concerned about the size of the bribe than the actual technology involved.
 
Well, didnt SA and Brazil sign some sort of Trade Agreements, for Developing countries banding together? Im 100% certain this was the root cause for the DoC changing their tune, under some sort of "higher" political pressure ofcourse. Naturally, this can neither be confirmed nor denied. Politics, politics....
 
As far as I can remember the Trade Agreement only came later on. However the initial reason why the DOC started all their crap about why DVB-T wasn't sufficient enough is because the counsel the Digital Dzonga or something set a Turn On date back in 2008, and when that date finally came all the private companies where ready & waiting on the DOC. The problem was the money set aside (R7 billion if I remember correctly) was gone. Now obviously they couldn't just come out and say "uh ja we stole the money" & that's when they come up with this crap about testing other standards & what not, Claiming that they never decided on a standard which is absolute crap because DVB-T was plastered all over the Digital Dzonga's web site.
 
And who wants to take bets that ISDB-Tb will be chosen?
 
Yipppeee we now have a new Minister! Please Mr Minister let us go for DVB-T2 and let us start rolling it out..
 
MFour said:
SADIBA's e-mail addy makes me wonder: "[email protected]"
Click to expand...

It should - very much so. SADIBA, gmail is NOT a proper service for you to use.

UnUnOctium said:
Scanned through the PDF, they did a proper (albeit a bit shallow) analysis.
Click to expand...

Agreed, but very, very 'shallow'. I do not think my points below will change the conclusions much (except DVB-T2 will look even better). I do think DVB-T overall is as least as good as ISDB-T, but one should expect a better analysis.

Comparing the most robust modes - very seldom used - and the highest bit-rate modes - not robust enough in real life - is no way to go. Use realistic modes.

Comparing in a AGWN channel is not very realistic - Comparing in a Rician or better in a Rayleigh channel will be much more realistic, with the less than optimum aerial conditions expected in much of SA.
A Rayleigh channel will make DVB-T2 look a lot better relative to both DVB-T and ISDB-T (rotated constallations, better interleave, ....).

When comparing guardinterval, the range of possible values and the overhead of longer GI's - i.e. the flexibility is the important figure. DVB-T2 is far better with a max GI of 532 usec (= 159 km) in an 8 MHz channel ( 32k GI 19/128 - Note: the page 14 is in error stating 32k GI=1/4 (is combination is not allowed))

On the same page 14 in the SADIBA presentation C/N is quoted as DVB-T = 20.1 dB and DVB-T2 as 21.8 dB (AWGN).

Firstly the DVB-T figure is from an older version of the DVB-T standard - the newest has the figure 20.2 dB ( http://www.etsi.org/deliver/etsi_en/300700_300799/300744/01.06.01_60/en_300744v010601p.pdf ).

Secondly the DVB-T2 should have added a few tenth of a dB for pilots and ... (read the EN 302 755 standard)

But more importantly, the figures for a Rayleigh channel is DVB-T 28.6 dB and for DVB-T2 25.3 dB showing how much better (> 3 dB) DVB-T2 is in a critical channel (take a look yourself. The DVB-T2 C/N figures are here: http://ompldr.org/vMTlwcQ ) .

Lars :)

SA, continue with the the few test DVB-T muxes for compatability and make everything new DVB-T2.
 
reslfj said:
It should - very much so. SADIBA, gmail is NOT a proper service for you to use.



Agreed, but very, very 'shallow'. I do not think my points below will change the conclusions much (except DVB-T2 will look even better). I do think DVB-T overall is as least as good as ISDB-T, but one should expect a better analysis.

Comparing the most robust modes - very seldom used - and the highest bit-rate modes - not robust enough in real life - is no way to go. Use realistic modes.

Comparing in a AGWN channel is not very realistic - Comparing in a Rician or better in a Rayleigh channel will be much more realistic, with the less than optimum aerial conditions expected in much of SA.
A Rayleigh channel will make DVB-T2 look a lot better relative to both DVB-T and ISDB-T (rotated constallations, better interleave, ....).

When comparing guardinterval, the range of possible values and the overhead of longer GI's - i.e. the flexibility is the important figure. DVB-T2 is far better with a max GI of 532 usec (= 159 km) in an 8 MHz channel ( 32k GI 19/128 - Note: the page 14 is in error stating 32k GI=1/4 (is combination is not allowed))

On the same page 14 in the SADIBA presentation C/N is quoted as DVB-T = 20.1 dB and DVB-T2 as 21.8 dB (AWGN).

Firstly the DVB-T figure is from an older version of the DVB-T standard - the newest has the figure 20.2 dB ( http://www.etsi.org/deliver/etsi_en/300700_300799/300744/01.06.01_60/en_300744v010601p.pdf ).

Secondly the DVB-T2 should have added a few tenth of a dB for pilots and ... (read the EN 302 755 standard)

But more importantly, the figures for a Rayleigh channel is DVB-T 28.6 dB and for DVB-T2 25.3 dB showing how much better (> 3 dB) DVB-T2 is in a critical channel (take a look yourself. The DVB-T2 C/N figures are here: http://ompldr.org/vMTlwcQ ) .

Lars :)

SA, continue with the the few test DVB-T muxes for compatability and make everything new DVB-T2.
Click to expand...

Agreed, AWGN is indeed too idealistic (and almost doesn't exist in practise) but do remember that due to the fact that they use OFDM it pretty much negates a rician or a rayleigh fading channel (effectively turning them into an AWGN channel) so I do believe that even if they used a rician or a rayleigh model (though one would argue Nakagami is a great one too for urban areas) it would make little difference but I do believe that they did it rather to have a simple model instead of having the idiots at DoC saying "ooh but you used the wrong model, should be SUI-6 instead of SUI-1" etc...

I think that they were aiming to prove that in its most robust, lowest data rate and least robust, highest data rate modes, DVB-T is superior. Anything practical they would need to have done site surveys :p

The great thing about DVB-T and DVB-T2 is that T2's improvement is mainly due to the larger FFT sizes (32k MADNESS) and LDPC + BCH instead of RS + conv FEC, both of which are a simple firmware upgrade.

I must say that it's great to see someone else on the forum who knows about the technical side of this, also a student?
 
UnUnOctium said:
The great thing about DVB-T and DVB-T2 is that T2's improvement is mainly due to the larger FFT sizes (32k MADNESS) and LDPC + BCH instead of RS + conv FEC, both of which are a simple firmware upgrade.
Click to expand...

When you get into the details, you will find lots of small things that add up to a lot - not least in difficult situations - eg. rotated constellations, the interleavers, optimised pilots, P1, P2, the higher FFT modes with their long symbol times and (if used) MISO of cause .

Take a look at the C/N values I quoted - COFDM in itself does not fix 0 dB echo or higher coderates in Rayleigh channels without a very significant higher ERP (or using DVB-T2)

UnUnOctium said:
... also a student?
Click to expand...
Yes, some 30+ years ago :D

Lars :)
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X