x
Page 1 / 3   >   >>
ponguy 12/5/2012 | 2:55:53 AM
re: Cox's Robbins Ruffles Feathers Well, I have to say that Robbins has guts, giving that speech in front of that audience. Even as a telecom vendor, I have to say it was in your face, funny and insightful. Which ironically highlights the low quality of most of the other presentations. I wish some of the "leaders" in telecom were half as brave and would say something a little more interesting then the vanilla insight I got at 2 full days of conferences at this show. For example, at the CTO Panel, how about asking VZ and SBC to explain why they are right with their technology choice (FTTN vs. FTTH) and the other is wrong for example? Or asking Microsoft-if everything is unicast plus DVR and I am always watching what I want, who cares about the latest IPTV "killer app" PIP/Mosaic?
naturalwonders 12/5/2012 | 2:55:52 AM
re: Cox's Robbins Ruffles Feathers I suppose this was the highlight of an otherwise
very lackluster show. The Exhibit halls were pretty devoid of customers. I asked one vendor how the show was going for him and he said oral surgery would be more fun.He was not the only one I spoke to that mentioned they would not be back.
godfather 12/5/2012 | 2:55:50 AM
re: Cox's Robbins Ruffles Feathers Robbins has guts. He said some very pertinent things. Cable has been doing Video for a long long time - Telcos would do well to listen to him.

Robbins also made a plea for the industry to work together - that would be the ideal scenario - but if that does not happen, cable is way ahead of the Telcos. If Cable loses 5% customers but gains 15% VoIP customers, guess who's winning?

Hats off to you, Mr. Robbins! Well done.

GF
paolo.franzoi 12/5/2012 | 2:55:49 AM
re: Cox's Robbins Ruffles Feathers
There is a noticeable delay in switching digital cable channels, but not one I would bet $6B on.

And this is the bit that Microsoft is having problems scaling in the network. The other bit they did not talk about is how they are going to manage 4 PIP bandwidths to every STB.

There was a lot of smoke and mirrors in the presentation.

seven
godfather 12/5/2012 | 2:55:49 AM
re: Cox's Robbins Ruffles Feathers Does Microsoft and Alcatel's IPTV actually work?

At the conference, a recurring theme was that IPTV does faster channel changing than cable does (instantaneous channel change).

That sounds like a lot of "marketing" inaccuracy.

If I understand correctly, the cable guys get all channels upto the set top box, while IPTV gets one, two or maybe three upto the set top. So, when you change channels, Microsoft needs to go back into the network, pick up a new stream and have that start streamin to the set top.

Cable just needs to switch the channel at the set top box - it's already coming over!

What gives?

How can channel changing that happens at the set top be slower than channel changes that happen many miles behind the set top?

Are Microsoft and Alcatel blatantly lying?
alchemy 12/5/2012 | 2:55:48 AM
re: Cox's Robbins Ruffles Feathers godfather makes an offer:
If I understand correctly, the cable guys get all channels upto the set top box, while IPTV gets one, two or maybe three upto the set top. So, when you change channels, Microsoft needs to go back into the network, pick up a new stream and have that start streamin to the set top.

That's correct. There are a few minor details for digital cable.... On a single frequency, multiple digital channels are multiplexed. The stream is encrypted in a proprietary Motorola/General Instruments or proprietary Scientific Atlanta way.

Cable just needs to switch the channel at the set top box - it's already coming over!

What gives?

How can channel changing that happens at the set top be slower than channel changes that happen many miles behind the set top?


I suspect the delay problem in getting an image after changing channels is mostly caused by Motorola and/or Scientific Atlanta using the cheapest possible processors, decryption, and decompression silicon in their proprietary set-top boxes. There may be something funky with their encryption scheme that adds delay. I'm not a video compression expert but there may be some delay caused there, too. Of course, you could build a set-top box that could simultaneously decode several channels at the same time to mask these issues.


I agree with you... I don't see how a unicast IPTV scheme could ever change channels as fast as a broadcast scheme.
canadian 12/5/2012 | 2:55:48 AM
re: Cox's Robbins Ruffles Feathers I am told that most of the delay is actually caused by buffering in MPEG-2 and is worsened in MPEG-4 (fewer I-frames being sent out that have enough video to show a complete new picture).

Microsoft and Alcatel stream a 2nd stream, encoded at a lower bit rate of only I-frames, and THAT, is what you tune to in IPTV.

Which would mean:
1. You don't get a clear picture on channel change like you do in cable (you see a highly compressed I-frame instead) until the real video starts streaming fully.

2. The bandwidth problem is worsened, so you need to compress even more.

3. You need to buy advanced Set Top Boxes with more capability, so no more $50 set top boxes!

Am I wrong? Does this approach scale?

It seems to me that Microsoft-Alcatel are the new leaders in selling Koolaid. And the Telcos are buying it, too...
stephencooke 12/5/2012 | 2:55:46 AM
re: Cox's Robbins Ruffles Feathers There are a couple of issues on the channel changing lag here:

- analogue vs. digital
- levels of encryption

Analogue cable channel surfing is as close to instantaneous as we will get any time soon. This is simply changing the analogue filter centre frequency. All information in the picture is refreshed with every frame.

Digital cable is another level of complexity entirely. Multiple channels are time-division multiplexed onto a single carrier. This means that it is also possible that you have two effects adding to the delay, possibly changing analogue centre frequency and its associated reframe, as well as decrypt. The bandwidth is guaranteed as each of the carriers is independent.

IPTV is the worst of the bunch as everything is time division multiplexed into, essentially a single carrier-type environment along with Internet access, VOIP, etc. Add to this the high levels of encryption for MPEG2 or 4, the 'bloatware' of MS, and the necessity of using cheap (ie: slow) processors to keep the STB cost down and you can see where this is going...

I once attended a presentation on the 'slowness' of provisioning and maintaining telecom networks. The researcher identified the problem as the processing power needed to deal with the OSI 7-layer stack. His solution was that the hardware needed to be faster.

If STBs are to go the way of the PC (ie: need to be replaced every 2 years with higher memory and speed requirements...the MS way) early adopters will suffer and there will be no end of network churn for the consumer. Probably not the impression that telcos want to advertise.

Steve.
konafella 12/5/2012 | 2:55:46 AM
re: Cox's Robbins Ruffles Feathers Seems to be some doubt being cast here from a few folks who have probably never had the opportunity to hold a remote control on a MSFT IPTV-enabled network. That's understandable, because the technology is pretty clever and it shatters some previously-held rules of thumb. Get a demo if you can.

Basic Question: Why did the "Interactive Program Guides (IPG)" become critical to the launch of Satellite TV and Digital Cable?

Answer: Because you couldn't channel surf anymore with the 1 sec delay between digital channels, even though the channels are all locally available in the set-top-box. Most of you knew that already, I'm sure.

MSFT's IPTV channel change times are dizzyingly fast - you don't need to relay on the IPG unless you've become accustomed to using it with your digital/satellite TV.

It seems there is considerable debate in this forum about channels being available locally in the set-top-box vs being available remotely in the central office. In actual fact, the delay difference between the two models is truly insignificant compared to the delay created by de-compression of content during a channel change event. And remember, compression is being used by satellite, digital cable, AND telco IPTV. So now MSFT has found a slick way to resolve that "slow response" problem for IPTV.

Perhaps, as Alchemy suggested, the digital cable channel change delay could be addressed by better set-top-boxes. Very likely indeed. But rest assured that cablecos will begin to adopt IPTV (not necessarily Microsoft's version) as part of a total convergence/IMS/VoIP strategy over time. And don't be surprised to see them shift their content back to their Hub Site (aka Central Office) rather than pushing all of it to each set-top-box. Why? In order to free up huge gobs of bandwidth on their coax for higher Internet rates. And to allow them to offer an unlimited "channel universe" (like the Telco's will be offering), rather than a fixed 200 channel line-up they current have. I'm sure many don't agree with my cable IPTV prediction, but I'm a firm believer that they will totally shift in that direction well before 2010.

PIP and matrix-view over DSL pipes is a simple problem to solve, when you recognize you don't need full bandwidth for the little PIP channels.

Regarding the advanced STB discussion, legacy STBs advance each year with newer technology and the price gets reduced with volume deployment. Most of us paid $200 for an inferior piece of crap to the STB you can buy today for far less. Same thing applies here. Your $80 linksys home router was unimaginable in the late 90's; it's simply due to residential mass deployments. Same for IPTV STBs.

So I suppose you can call me a believer :)

KF

ps: "Canadian", you're comment about I-frames only during transition is not correct. Better check your notes!
opticalwatcher 12/5/2012 | 2:55:46 AM
re: Cox's Robbins Ruffles Feathers Doesn't Verizon's FIOS approach use (at least for now) video overlay of another frequency on the fiber? I believe that this is part of the GPON spec. In this case, optical television is just like digital cable--it just goes over a fiber rather than over a coax wire.

So they are avoiding the all the short term IPTV problems. They can always switch over when the bugs are worked out.

In the long term this looks like a better approach than cable--it has all of cable's advantages, but has much higher bandwidth.
Page 1 / 3   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE