Digital TV group sets 3D standard The overseers of the DVB digital TV specification have given the thumbs up to a proposed standard for broadcasting 3D footage. DVB-3DTV was mapped out last year, but only completed a month ago. Now the proposal has the backing of the DVB organisation, it can be formally added to the DVB standard to join the likes of DVB-S for …
DTV is an open standard; anyone can implement it (there is a license fee if you want to use the DTV logo though).
And while you are sort-of right - the likes of Sony and Pana and Samsung etc will have hammered out this spec with the broadcasters (basically the BBC and ITV here in the UK), it's very unlikely people like the MPAA and FACT will have had much (if any at all) input into this; certainly not the MPAA (bearing in mind they are American and the US doesn't even use the DTV standard). DRM (which is what the MPAA etc are interested in; oh, and making money, of course) is a different issue altogether, and generally outside of the basic DTV spec - the DTV spec says how it might be supported, but implementation is largely down to the broadcaster. Sky's DRM, for example, is famously kept very secret and firmly in-house (well, developed for it by that Israeli outfit).
But you are right in the gist of what you say - the TV manufacturers are interesting in selling TVs and making money. That's what they do.
DVB-S / T / C tuners are very cheap to come by. Assuming the 3D channel was broadcast unencrypted you could just decode it like any other channel and play it through any media player capable of handling a SPTS / MPTS.
Of course most pay content is encrypted. It's encrypted usually through DVB-CSA which dictates the cipher and some 3rd party DVB-CSA provider that implements the key recovery / conditional access / DRM part.
You could still capture encrypted content but you'd have to break the crypto. DVB-CSA has been reverse engineered and the key length is quite short. It's possible we're not far from a bruteforce crack that works in realtime though there are more robust versions of DVB-CSA that use longer key lengths.
Don't confuse scrambling and encryption, they're quite different.
The DVB-CSA scrambles the stream and generates a control word that is used by the receiver to put the stream together again.
Some _very_ proprietary encryption systems (from NDS, Nagra, Vermitrix etc) are used to encrypt the control word during transport. Cracking this encryption is a much bigger job and almost certainly not possible in real-time.
As with any encryption system that relies on keys it's how you move the keys around that is important.
The common denominator for all these crypto schemes is CSA. CSA takes a 64-bit key and uses it to unscramble the content. if you were to brute force the key it wouldn't matter what protection was over the top. The 64-bit key isn't even that since 2 bytes are checksums so it's effectively 48-bits.
I think 48-bits is low enough to consider a brute force attack. Write some OpenCL app which divvies up the keyspace into blocks and farms off the work to kernels. Each kernel would test descramble the content with the key and and run through some sanity tests to see if it was viable. Since the data is MPEG2 in most cases it should be predictable enough to do this.
Maybe it couldn't be done in realtime but depending how often keys were cycled (and if there was any non-random predictability in them) it might be sufficiently quick to decode a content in the space of a day or so which makes it somewhat viable.
Most CA systems I've worked with have an ECM period of around 30 seconds.
Having to brute force a CSA key that changes every couple of minutes at best isn't a viable use of computer hardware.
It'd be on dvd/blu-ray by the time you'd finished the crunching, and you _still_ wouldn't want to watch it :)
The recent take-up and future interest in 3D has been shown to be very low indeed - A number of (neutral!) industry assessments have shown this to be true in many locales across the globe.
The commercial push for 3D seems to be from those manufacturer(s) who also have film studio/prior content ownership - It appears to be nothing more than another "missed opportunity": these studios will NOT produce creative and original output, rather than use this technology just as a vehicle to re-make/re-hash those old titles.
Innovation shouldn't stop at the technology, the content must be innovative, too. From a technical perspective, current '3D' isn't truely 3D; a veiwer can't move around the scene to get a different perspective of their own, its just the content producer's take on what they want to project at you that you are allowed to see...
When you really stop to think about 3D, what does it add to the story, the plot, the sporting event? Is it truely worth the premium that TV/Player manufacturers/Studio owners really want to extract from you?
Anamorphically squashing L+R frames into a single frame means half the horizontal resolution. Perhaps that's the easiest way to shoehorn 3D into existing transport streams but it's hardly a good forward looking solution.
It's half the horizontal resolution and can hardly be efficient from a compression standpoint since there is huge redundancy between left & right eye but both images are fully reproduced in the stream.
I wonder if the DVB chickened out from implementing something like MVC (multi view encoding) under pressure from STB manufacturers like Sky who want to retrofit 3D into their existing offerings rather than push out new hardware.
I'd certainly keep an eye out for any broadcaster who claims a channel 3D *and* HD in the same breath because the ASA might take a dim view of that claim considering it's effectively halving the number of pixels each eye sees.
Nah - you are still thinking in analogue mode. This is digital - MPEG 3D compression will almost certainly transmit left frames (say) plus the differences between left and right frames. Which for most content will amount to very little extra data and thus very little loss of definition compared with 2D.
I know how it works and I mentioned MVC. For your info, that is an extension to AVC whereby you designate one view (e.g. left eye) as your base image and include deltas to construct the right eye image from the base image. In theory you could have arbitrary number of views although 3D is likely to be stereoscopic for the time being. Such MVC encoding is backwards compatible with existing AVC decoders because they just broadcast the base image and ignore the deltas.
This is the way DVB-3D should have been specified. I expect it didn't happen that way is because Sky kicked up a stink that their precious Sky HD box was lacking a SoC which could do MVC but it could be tweaked to tell a 3D TV through HDMI & EDID to set itself to side by side frame mode.
i.e. side by side framing is Tesco Value 3D and customers and the DVB shouldn't have put up with a solution based on it.
If I recall correctly there was recently a fiasco where a range of PVR boxes (Digifusion, Sony), clearly carrying the "Digitail Tick" fell over when the 14 day EPG from 4tvInteractive was withdrawn, as they weren't compatible with the default DTG 7 day EPG.
If I recall further, the requirements which Digifusion and Sony had to meet, to comply to qualify for the digital tick, were quietly changed. There was no change to the digital tick mark itself, which meant that people had no idea whether their kit did or didn't comply with the latest specs.
So, on the basis of past performance, is the Digital Tick even remotely worthwhile?
Side by side: 1920 divided by 2 is 960, still less than the minimum 1280 and only barely better than standard definition's 720 horizontal. Over/under format suffers similar degradation.
Why even bother, 3D adds nothing to the comedy, horror, drama, suspense, etc. I went to see Avatar in IMAX 3D and sure a few landscape scenes benefited, but it did nothing to enhance the story telling.
I was at the Sound and Vision Show in Bristol over the weekend. SIM were showing off Bluray 3D but the equipment to get the best experience totaled over £60k! And you still had to wear uncomfortable viewing glasses.
This "features" are already standards in HDMI1.4 (physical layer) and in H.264 MVC (multiple video codec) layer so, all of them are already working on all terrestrial/satellite HD Digital Television Standards that use H.264, as DVB-T2, ATSCv2 and of course ISDB-Tb Brazilian/Japanese HDTV Standard in use on almost all LatinAmerican countries. These standards are invisible on old receiver models granting total compatibility.
Since July 2010, RedeTV! Brazilian channel is broadcasting in 3D all over Brazil and in a Cable TV in Paris, France.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
