Your browser is not supported. Please update it.

24 November 2011

WHDI could partner with WiFi to make the leap to video dominance

The 3 Gbps WHDI video technology could end up being a set of algorithms that are licensed to existing WiFi chip makers, which could suddenly absorb the delivery of uncompressed video around a home into standard WiFi devices. And this could happen as early as the next generation of 802.11ac devices, says the president of WHDI.

This was Leslie Chard, president of WHDI speculating on how things might develop and he made it clear that he was not speaking on behalf of chip maker Amimon, who owns the technology, but like us, he was just speculating that the move seems fairly attractive.

‘I can see it going that way, definitely, if Amimon can reach license terms with some of the other major Wifi chip makers, but I’m not saying that’s going on right now, just that I could see the sense in it.’

‘The big problem with WiFi devices when they send video is that it is compressed video arriving at TV sets. Once it gets there each set requires a lot of video decoding power. The processing power to carry out the algorithms of breaking down the video using the Amimon technology is a fraction of the processing power that hard core decompression takes.’

Chard was talking to us about WHDI. When you are constantly trying to keep up with the multiple 60 GHz wireless home network technologies ‘ WirelessHD and WiGig IEEE 802.11ad, as well as wired systems such as MoCA, G.hn, HomePlug and HPNA and of course plain old WiFi and its future variant, 802.11ac ‘ sometimes it’s hard to keep touch with the specialist video technologies as well or to give them the credibility they deserve. Two which stand out here are HDBaseT, a wired system that bundles power, video in and a control IP path out onto a single wire; and Amimon’s WHDI, effectively a shot at replacing HDMI cables with a specialized wireless signal.

Of course if the process of dropping WHDI into WiFi chips is that easy, it sounds like the system could almost be software downloaded to top end WiFi chips, and the possibility of such an outcome would make Amimon a huge takeover target almost overnight. Why would someone like Broadcom or Qualcomm Atheros license a video delivery process from a specialist, when it might end up cheaper to actually buy the company and get an advantage over its main rivals.

Chard led the industry bodies of HDMI and DVI and then served as president of HDMI Licensing, where he helped it take off and gain industry acceptance and what is today, mass adoption. Perhaps he is angling for a similar role for WHDI licensing?

Israeli start up Amimon invented WHDI, and it uses the same spectrum that 5GHz WiFi uses, because that spectrum is almost universally unlicensed, so that any device anywhere in the world can use it. It works in a total of 40 MHz, can use 2 x 2 20MHz MIMO for the radio signal (originally we were told it was 4 x 4 but perhaps this has changed or is optional) and is optimized for video delivery, always for uncompressed video travelling from a demodulator device to the eventual display.

It claims that it offers the equivalent of 3Gbps, but that equivalence is never really unraveled in terms of underlying data performance, except in a paper on the company’s web site which implies that the true data rate varies between 250 Mbps and 800 Mbps depending on the error checking scheme used. Whast WHDI means is that it would take a system offer 3Gbps and operating like WiFi, to deliver that kind of stream accurately. What Chard tells us is that the video is broken into two streams, one which is essential for the video to play at all. If the stream was compressed then here we’d be talking about protecting I-Frames, but it’s uncompressed.

Instead WHDI uses a technique called Joint Source-Channel Coding which first prioritizes video components, then applies two separate error correction schemes to the outcome, encoding the most significant bits of each pixel, the important components, with better error correcting code which means it ends up transmitting that part at a lower rate, but in a signal that can better handle noise. The system then opts to send the less vital information either across the noisier frequencies or rather in a way that high noise system could harm. It also uses space-time coding for better protection of the components. The same type of quality distinction was made in some video codecs used, for instance, in the now defunct Qualcomm, MediaFLO.

A space’time code is a method used in MIMO systems which relies on sending multiple, redundant copies of data in the hope that at least some of them may survive the physical path even if there is interference. The whole thing is carried on a number OFDM carriers in multiple channels, not unlike the Quantenna version of 802.11n wi-fi, and so it can borrow from RF the component base for WiFi.

The system delivers 12-bit color depth and the signaling is in 16 QAM OFDM and has HDCP stream protection built in and will travel up to 100 feet through most types of materials. All of this is carried out with just a couple of milliseconds of latency, not something that generalized home networks can yet boast.

The other thing it does is something that recently WiFi systems have put in place which is Dynamic Frequency Selection, so that it checks the condition of its own channel and neighboring channels and switches actively in the timing between a single video frames, to the least busy or least noisy.

Given that IEEE 802.11ac uses 40MHz and 80MHz chunks of spectrum and this peaks when using 2 x 2 MIMO at around 1 Gbps PHY, and a real world whole home performance under 200 Mbps, you can see that the WHDI video signaling technique is saving an awful lot of bandwidth, but it has been shown sending multiple 1080p video streams, and the next generation will cope with 4Kx2K streams and 3D ‘ again uncompressed.

But if Amimon and WHDI has been a little quiet of late, it’s not led to a lack of success. Back in January the organization claimed that it was leading a charge in China, and cited increased demand for Amimon chips from the likes of ASUS, BOE, Changhong, Galaxy, Haier, Hisense, Malata, Panda Group, TCL, Winstar and 35.com who have all introduced either Wireless HDTVs with embedded WHDI, PC-to-TV adapters, tablets with WHDI embedded or WHDI-based sleeves for mobile phones. Since then, in September, Hewlett-Packard has introduced its Wireless TV Connect product, a similar video bridge from the PC to a TV, using WHDI. Motorola joined early as both an investor in Amimon and a WHDI member and is thought to have a similar device coming out in the next few months.

Has WHDI seen any operators buying in bulk? ‘No we’ve had what you might call operator curiosity, but definitely not operator buy-in, at least not yet,’ Chard said.

The original founder members of WHDI were the major Japanese TV makers ‘ Hitachi, Sony, Samsung and Sharp and since then LG has also joined, and Motorola made it onto the promoter’s page alongside inventor and currently sole chip supplier Amimon. There were stories that Sony is no longer a supporter, but its name is still there on the WHDI web site.

The chipset is in fact in two parts, an RF chip that can come from dual sources, originally from Maxim and a new one from SRI in Germany has just agreed to second source, and a modulator chip that is only made by Amimon. Most TV technologies – HomePlug with Intellon, MoCA with Entropic and HPNA with CopperGate, HDWIreless with SiBeam, begin life with just one chip source and that’s the same here with Amimon.

‘But now we have version 2.0 of the standard,’ says Chard, ‘there will soon be other chip makers.’ And that is pretty much the reason for the chat. Amimon has been shipping standardized, compliancy tested chips for a long time and as verison 2 comes out it will make way for other suppliers to jump on the bandwagon, as volumes come. There are always rumors around technologies like this, but given that Amimon is shipping merely ‘hundreds of thousands’ of chips a year, and that it announced last May 2010 that it had shipped half a million chips so far, it would appear that chip shipments are stalled right now. It was supposed to push past the 1 million mark by the end of 2010, and the fact that WHDI has not come out with a statement on this, it’s likely that has yet to happen.

But Chard is constantly optimistic, ‘A major PC owner plans to embed it in a lap top shortly,’ he said, but no CE manufacturers, the core of his organization membership, have decided to bundle the product yet inside the majority of their TV sets. ‘There have been options to pick it like an accessory and ship it with TVs as a point of purchase choice. In that version there is a slot on the TV where it just slides on the back,’ he said.

The standard for WHDI of course has been around a while (2009) and had an extension for stereoscopic 3D so that it could be used for real time remote surgery (where it is very successful), but WHDI 2.0 was only passed in Q2 2011 and if anyone is coming out with chips for it, that would likely be for devices for the Holiday season in 2012.

Chard gives some clue to what’s in store with 2.0 ‘ 4kx2k standard, which means 3840 pixels wide by 2160 pixels high as in HDMI 1.4 and an equivalent data rate of 10.2 Gbps.

It will also have broader channels, better WiFi co-existence and as Chard points out he has shown it working in multiple channels with WiFi at CES this year. ‘WHDI can sense a WiFi signal as interference, and jumps out of that channel into another channel and can jump back when the WiFi activity goes down. There may be more co-ordination between WiFi and WHDI in the future,’ he added.

‘There will be some algorithm tweaks that will make the entire process more efficient and we will also drive down the chip cost, and cut the power use. Right now it is around 2 to 2.5 watts on each chip and we see that coming down to 300 mWatts.’ Retail device prices today for a pair of WHDI chips range between $149 and $249, and this will have to come down to compete with devices like HomePlug AV, and go to at least sub $100 per pair we suspect, but that will come automatically if the volumes come.

If Chard is right and Amimon decides to go the licensing route, the next generation of 802.11ac chips should have the processing power and the speed to offer the same kind of performance, but until WHDI gets a second chip source chasing Amimon for volumes, is likely to just keep making the chips and building towards that magic 1 million a year and above mark that it has talked about in the past, and the guideline we usually set for chip profitability, more like 5 million chips a year. At the point Amimon would have substantial profits and be ready for such negotiations with one of the majors.