Your browser is not supported. Please update it.

30 June 2011

New satellite modulation from Novelsat pushed as possible DVB-S3

After launching its new NS3 modulation technology just three months back, positioned to be substantially more efficient than DVB-S2, start-up Novelsat says that it has already established 32 live satellite trials.

The system will first be used in a series of VSAT delivery satellites, using a simple software upgrade, but it is really targeted at becoming a replacement technology for DVB-S2 delivering video from the sky to consumer homes, and the company is already in talks to begin a standardization process with the DVB Project, and says that a 72 MHz transponder channel which uses NS3 is anything from 28% to 90% more efficient than DVB-S2.

DVB-S2 is only 8 years old and was first deployed just 6 years ago and many DTH satellite TV deployments still use the original DVB-S delivery. DVB-S2 uses
modulation schemes up to 32APSK and a generic IP packeting including MPEG-4 and MPEG -2 video streams.

DVB-S2 was about 30% faster than DVB-S when all network conditions are the same, and we imagine that any successor would need to offer about that much improvement again in order to be economically viable and NS3 does seem to fit the bill.

Given the huge increase in the number of HD TV channels, both delivered around the planet for contribution and also for final consumer distribution, as well as a more recent taste for 3D, increasing satellite delivery capability is high on the agenda for satellite operators. Novelsat claims that its new system will deliver 358 Mbps in a 72 MHz transponder channel, which it says is roughly 20% more than DVB-S2 could possibly manage and more than 28% better than most real world situations. Many DVB-S2 transponders today deliver only 168 Mbps, which makes it a substantial improvement of more than double.

The basis of its claims are shown in the following graph, which shows that Novelsat has perfected a long discussed, and often dismissed encoding of 64 APSK (amplitude phase shift keying). Many papers have been written in the last 3 years suggesting new ways of optimizing constellation shaping in APSK to make it more efficient.

Novelsat’s co-chairman David Furstenberg told us, ‘Everyone thought that satellite was getting close to the limits of Shannon’s law, so stopped investigating new algorithmic approaches to satellite transmissions. The technology had moved its focus from spectral efficiency to data compression.’

The graph shows theoretical performances in a single 8 MHz channel and Furstenberg notes that most use of satellite DTH is in the noise range of 7 to 9 CNR (carrier to noise ratio), which is where NS3 claims a 28.7% increase.

This technology can be implemented with a single chip switch at the ground station and at the receiving end, and right now an FPGA has been built to demonstrate the value of NS3 and an ASIC is planned for later this year, cheap enough to put into set top designs. ‘There were just too many skeptics who said that this wouldn’t work so we had to make a chip ourselves,’ said Furstenberg. Later Novelsat hopes to license the technology and get itself out of the chip business, which is why it has tried to get the DVB Project to back the standard. It will be first considered at a meeting in July. Later implementations might find their way onto SoCs if NS3 takes off.

‘Many 72 MHz transponders are set up as two 36 MHz transponders, whereas we can use the entire 72 MHz as a single carrier. So NS3 can get 28% improvement by copying this 36 MHz x 2 configuration, but claims to be able to reach a 62% improvement in performance when used as a single carrier, with further improvements by upgrading power supplies to address the full 72 MHz and gains from statistical multiplexing, leading to a 90% improvement over satellites as they are used today.

If they implement NS3, Novelsat says TV companies can lease less satellite transmission capacity and save as much as $3.7 million a year on a single transponder.

Novelsat suggests that the improved spectral efficiency can be delivered in a number of ways, either as additional capacity, or in term of smaller receive antennas or in terms of a larger satellite footprint, or all of the above and says that with next generation satellite transponders, which can use up to 200 MHz channels, could reach 1 Gbps. It also says that it has applied for over 30 patents around the world based on this technology.

Of course the industry is right to chase improved compression ‘ H.264 is in its infancy in implementation terms and once upon a time a standard definition TV channel in MPEG 2 used to use up 7 Mbps when it was first introduced (1980s) and by the time it was ready to be overtaken by H.264 compression (late 1990s) that was down to under 2 Mbps in some implementations. This is likely to gradually be yielded up by algorithmic improvements in compression implementations in H.264 in parallel with these types of spectrum gains from Novelsat.