Satellite versus Cell Phone Streaming: An Essay

[adjensen]

Over the past few months, more and more people have come into this forum, saying that Sirius XM is done, and that because they can stream music over their phone, satellite’s time has come and gone. Generally, I believe that these posters are genuine, but lack an understanding of how both satellite and Internet/cellular distribution functions, and after correcting misinformation time and time again, I’ve put together this little tutorial to explain it, so that I can just post a link to the thread in the future.

First, a couple of givens and assumptions. The subject at hand is music, not video, text or other data. The defense of XM isn’t that it’s the world’s greatest broadcast service, that they’re intending to be all things to all people, or that what’s playing meets your personal needs, as streaming last.fm or Pandora might. Instead, the premise is that, for an international music broadcasting service that is intended for a mainstream audience with universal, inexpensive access, satellite is the most efficient and economical approach.

The Internet and cellular technology referred to here represents the current state of technology, as well as a reasonable expectation for the future. Claims of “3G cellular is universal!” and “4G cellular is right around the corner!” are taken into account.

Finally, I am neither a Luddite nor am I a Sirius XM Apologist. I’ve been involved in computers and networks for over 30 years, am a reasonably early adopter of technologies, am not happy with Sirius for dropping Fine Tuning, and consider music and near constant access to it to be an important part of my life. But, more than any of that, I’m the sort of person for whom misinformation really grates -- because all too often, misstatements are picked up by many as pure fact, and poor decisions are made, solely because they were misinformed. This issue is one of those, and this post is my effort to clarify the muddy waters.

Part One: Sirius XM Satellite Distribution

The first system we look at is the existing Sirius XM set up. Here is a quick graphic that visually shows how this works.



As we can see, the benefit of this transmission method is quickly evidenced. A single signal, sent from the broadcast centre, contains the whole of the content of the service, and, when rebroadcast from the satellites, this same signal reaches everyone in the broadcast region at the same time, with the same content.

Technical benefits of this method are led by universality -- so long as the user has a view of the satellite, they have access to the service, without any physical infrastructure at the location, beyond a simple receiver and antenna. In addition, there are no user based demands or degradation on the system -- if one million people listen to 70s on 7 and twenty people listen to 80s on 8, the system neither knows nor cares.

Some of the economic benefits of this system are reduced variable (per subscriber) costs, which border on non-existent for the technology -- it costs exactly the same to broadcast the signal to a million people as it does for one person. (Additional variable costs include royalties, customer facing support staff, credit card processing fees, and so on, but any paid service would incur those costs, and they’re not related to distribution.)

There has been a healthy growth in the number of people who have expressed dissatisfaction with the service in the past six months, and those complaints are the result of the two major disadvantages of the system. Technically, the system has very limited bandwidth, due to the capacity of the radio frequency used to transmit to and from the satellite. These frequencies are part of the electro-magnetic spectrum, which is a physical thing that has limits built into the very fabric of our Universe. See http://en.wikipedia.org/wiki/Electromagnetic_spectrum for more information about the spectrum. In the United States, the use of the “radio spectrum”, which is considered to be a public asset, is administered by the US Government, which has licensed a limited number of frequencies to Sirius XM for satellite radio. See http://en.wikipedia.org/wiki/Spectrum_management.

Limited bandwidth means that there is only so much that the company can broadcast, which results in a delicate balance of channel count and channel quality. These two are interrelated -- a decrease in quality can mean an increase in the number of channels, and cutting some channels can result in higher quality sound. Both of these are affected by compression -- the act of turning data into something similar, but smaller, so that it takes up less space. Less space means more channels, but high levels of compression never sound like the original when uncompressed, so sound quality takes a hit. It’s really a big balancing act.

The limited count and impact of quality is why there are, for example, two jazz channels on Sirius, rather than a half dozen, as a jazz aficionado could probably justify. Any time you think that there should be another channel, bear in mind that the only way that a channel is going to be added is to eliminate something else.

So, I mentioned a second major deficiency, and this one’s economic. While the costs of running the service day to day are not trivial, they pale in comparison to the costs of putting the satellites and broadcast centre together in the first place. All of the discussion earlier this year of the company’s debt wasn’t debt related to hiring Howard Stern or paying RIAA royalties, it was the debt associated with these startup capital costs. Sadly, though the issues pressing at the time were resolved through the intervention of Liberty Media, there remains more debt to be resolved later this year, and a company that can’t cash flow can’t save money to pay off debt.

When you’re losing money, you have two choices -- cut expenses and increase income. The company did the former last fall when it trimmed staff and merged the two disparate music services of Sirius and XM. The managers and executives believe that the Sirius approach to music, familiar and comfortable programming aimed at mainstream listeners, stood a better chance of income growth, and that’s the road they’ve taken. Again, I’m not making a point either way, just trying to clarify why it’s an issue.

Finally, the satellite distribution method is vulnerable to a single source of failure. Although they have back ups and redundancy, satellites have limited useful lives, are pretty tough to fix, and represent a major expense if they need to be replaced. In addition, problems at the broadcast centre or transmission relay can result in system wide outages.

[adjensen]

Part Two: Internet Radio Distribution

Next, we’ll look briefly at Internet Radio, with examples being services like Pandora, Slacker, AOL Radio, last.fm and so on.



As this illustration shows, we have a similar distribution tree, with a couple of additions. The first leg, from the provider to the first tier of distribution, is pretty much the same, although in this instance, rather than shooting something into space, it’s simply uploaded to the nearest backbone hub and sent out to the Internet from there.

The similarities end there -- instead of blanket broadcasting all channels (limited to five in this example,) to everyone, the system packages it up into little packets and sends them scurrying down the “tubes” until they reach their destination. If the listener can’t be reached for some reason, the packet is generally going to be lost.

Benefits of this, from a provider’s standpoint, is the limited cost associated with setting it up, instant access to a world wide potential audience, and the fact that the only practical limitation of bandwidth is what you’re willing to spend for. This allows a service like Pandora to effectively create an unlimited numbers of “channels,” as each user can have their own personally tweaked station, all playing simultaneously, with pretty good sound quality. Is this cheap? No, but it’s a much different world than Sirius XM is in, which can’t expand the electro-magnetic spectrum to get more bandwidth, no matter how much they could spend.

The unlimited nature of Internet distribution lends itself quite nicely to niche programming, which XM did in the early days, but has since (even long before the merger,) moved away from. Niche programming means you have a channel of Smooth Jazz, one with Disney Theme Park Music, one with nothing but Stoner Metal, and so on. With limitless options, it’s pretty easy to find something that you like, and with services that are configurable like Slacker or Pandora, if you can’t, you can just make it yourself.

Well, there are always trade offs for advantages, and the disadvantages here are costs to the provider, lack of universal access, and the additional costs borne by the end user. In the first, the more listeners that a producers has and, more importantly, the more channels that a producer offers, the higher the costs of providing it (distribution wise, again, we’re ignoring the royalties and such.) If I provide a service that lets you pick your favourite Pink Floyd song, and you start playing “Echoes” and every three minutes someone else comes in and chooses to start playing “Echoes”, by the time you’re done hearing it, there are five other streams playing the song, and they can’t be consolidated. That accelerates expenses and, as you can probably see, can quickly cost a lot of money in bandwidth and server expenses.

Whatever the cost of the service happens to be, you, as an end user are going to bear some increased costs, if for no other reason than streaming music is one of the justifications that people use to pay for $35 a month cable modem or DSL access, rather than ten bucks for dial up. Minor, perhaps, and you can use the faster speeds for other things, but it’s a cost that isn’t incurred with satellite distribution, so not something to be ignored.

Finally, there is the not minor problem of accessibility. While there are few (if any) places in the country that don’t have Internet of some kind, high speed broadband is far from being universal, and it’s largely limited to permanent locations that are wired into the rest of the net. You might have some rippin’ download speeds at home, thanks to WIFI and a cable box, but stray too far out the door with your laptop, and you got nothin’.

That’s where our friends, the cellphone evangelists, come in.

[adjensen]

Part Three: The Problem With Cell Phones

What if there was a technology out there that allowed you universal access to Internet Radio, letting you hear your “Marching Band Music” channel or “Poison Covers Winger Covering Cinderella!” channel? That’s the allure of the “last mile” solution that doesn’t involve any wires at all.



Here, we’ve broken apart the Internet distribution to show just the last piece of connection, where you are streaming music from a cell tower instead of an ISP (technically, the cell phone company is the ISP, and the tower is just serving the purpose of linking you to them, like a cable modem, but that’s not really important.)

On the surface, this looks like a pretty good solution -- we’ve no wires to tie us down, we access the service with a device that we already have for another purpose (making phone calls,) and everyone keeps talking about it, from XMFan devotees to bloggers and technologists, so it must be the bomb, right?

Unfortunately, the reality of the situation is a bit further away from the “here and now” solution that this seems to be. First of all, in the absence of a data plan, costs can be astronomical, so make sure you have one in place before connecting even one time. When you do that, you’ll notice that data plans aren’t particularly cheap, certainly not when compared to satellite radio.

The second problem that you’ll find is that appropriate cellular networks aren’t anywhere near as universal as the adverts might have you thinking. Stray out of a major city or off of an Interstate highway, and you may find the networks a lot less reliable. Touters of “3G” tend to ignore the fact that, outside of major metropolitan areas, and certain parts of the country, 3G is non-existent. “Oh, but it reaches (fill in the blank) percent of the population!” Fine, but if you’re listening to streaming music, it’s an either/or proposition; if you need to drive from New York City to Buffalo, and you lose your 3G along the way, you’ve lost all your entertainment.

The economics of laying in a significant technical infrastructure across the vast spaces of North America pretty much ensures that 3G, 4G and similar technologies will never be universally available in the United States and Canada. Not in a month, not in a year, not likely to happen ever.

In addition, as noted in the map illustration of our “yellow dot” listener, as we move along, our data connection has to jump from tower to tower. Devices do some buffering to afford breaks in this process, but if one tower isn’t receptive to what you’re doing, or the signal drops, or something else bad happens, you’re going to be listening to nothing. If you jump between towers while you’re checking your email, things might seem a little slow, but otherwise, you notice nothing, but you’re sure going to notice dead silence when your stream cuts out. Even when present, 3G networks aren’t guaranteed to provide constant signal as you move throughout the network.

Finally, the piece that no one seems to really understand is that today is the “early adopter” stage of Internet over cell phone usage. The vast majority of people not only don’t own a “smart phone,” they’re probably not all that sure why they’d want to. Aside from thinking it makes them cool, the biggest benefit to the current over-the-cell-music-streamer is that they are having little to no impact on the cellular network.

Cell phone networks were designed to carry voice traffic. Originally, analog signals, but they eventually migrated over to digital, which is more efficient (remember “compression” from the satellite radio discussion?) and can do more things. Things like text messaging, sending pictures and checking your email. These are all “low footprint” activities -- even though sending a picture chews up a bit of bandwidth, it’s not constant. Streaming music is, and it’s a pig for space.

Back in our discussion of satellite, we talked about limited bandwidth, due to radio frequency constraints, and how that impacts what Sirius XM can offer. Well, at this point in our Internet Radio discussion, we’ve come full circle, because “last mile” in the cellular network is, in fact, via radio frequency, and that means limited bandwidth. So long as it’s just you and your “in the know” pals streaming music, probably not a problem, but if too many people start doing that, it will start to bog down the system, impacting what it was meant to handle, voice traffic.

When that happens, the cell phone company has a number of choices, but they pretty much boil down to two. The company can spend a bunch of money to expand the network to accommodate the inefficient delivery method (as noted in the illustration, rather than the single source of satellite, or the composite method of Internet Radio, last mile in a cellular system is individual streams.) Those additional expenses, assuming that the system is even fixable, will be borne by you, the data user.

Alternatively, they can block, or just throttle, certain specific types of usage, in order to get by with the existing system. Fat music streaming pipes would be a natural target to hit, and, as noted, digital music streams are an either/or thing -- miss more than a few packets due to throttling, and it’s lights out.

[adjensen]

Conclusion

For many of us, niche broadcasting is something holds a lot of appeal. I originally signed up for XM because of Music Lab, a Progressive Rock channel (sort of) and many of those who signed up for the service in the first five years or so likely did so because of the deep and varied music offerings. Unfortunately, the market of people for whom that’s important is not huge, and, in the face of looming debt and needing to balance the budget, both Sirius and XM moved towards a simpler, more broadly appealing approach prior to the merger. Post merger, this is mostly done, and we’re even seeing small movements back towards deeper playlists and less repetition.

But we’ll never see niche broadcasting on the service again, and the offerings on the Internet do a very good job of serving those needs. The main stumbling block to acceptance of that service is accessibility, particularly mobile access. I would expect that many older signups of Sirius XM, who listen at home and have broadband Internet access, would eventually migrate to an Internet based provider who caters to their particular musical whims.

Given the current state of technology, reliable access for mobile users to Internet Radio remains elusive, and will likely continue that way for the foreseeable future. Advertising and claims aside, universal availability to high speed Internet through cellular networks will likely never happen, and wider usage in areas that do have the service will likely present all new challenges.

For these reasons, I think it rather clear that, in providing an inexpensive, universally available music service that caters to a mainstream audience, Sirius XM satellite distribution is the most efficient and economical approach.

If you have any suggestions, comments or points of clarification, I’d really appreciate hearing them!

[2000GTP]

This should cover most of the usual replies.

1. Your charts mean nothing because we all know 5G will be out in a few weeks and that will kill any chance Sirius XM had.

2. Mel is sent by the devil and will destroy satellite radio.

3. They lost subs last quarter, and will lose every quarter from now on.

4. Wimax has 110% coverage in the US, satellites are useless.

5. There are more channels than songs played.

[rjom]

Excellent analysis. Thank you for taking the time to write up your essay. It is well thought out, and a very interesting read. What you say makes a lot of sense, and is the best portrayal of the competing music delivery methods I have seen.

Well done!

[daver40]

Well done. Thanks for the effort, not that the cellphone evangelists will listen.

[SenorXM]

I own a cell phone, but I only use it to call the wife and ask if we need milk.. I don't stream, text, take pictures, I just talk...That said:

This argument about if cell phone, and Ipods, etc. are hurting SXM is, IMO, pointless.

Satellite radio is fabulous, and it's coverage and variety is unsurpassed.

But this year SXM came within a few hours of bankruptcy and the last 2 quarters they have lost subs, they are not showing growth. That is very bad news for a company that almost totally depends on subscriptions and growth to pay their bills..

For all the warts with Ipods and streaming music to Blackberries and Cell Phones, the fact is people are looking elsewhere besides Sat. radio for their music. SXM has some serious competition.

[runnoft]

SXM does have some serious competition, and part of the way that developments have not worked out for SXM is that technology is constantly evolving and coming up with new entertainment products to meet niche needs, which results in a slow bleed-off of those parts of the SXM customer base which is niche-driven as SXM feels compelled to go more mainstream. I don't know what the XM guys' assumptions were in developing XM radio, whether they thought they'd only have broadcast radio and TV and cable and satellite TV and CDs and home videos to compete against, but obviously, if they assumed that, they were wrong. And this situation is not going to improve for them.

adjensen, I agree with most of your analysis. The part about how wireless networks and streaming technologies can't do what SXM can do particularly for people driving outside of major metropolitan areas and want music or talk in their vehicles, and may not be able to do it anytime in the foreseeable future, seems right, but that doesn't mean that the marketplace is going to remain as is right now and that technologies we can't even foresee won't, in the next five to ten years, come up with new ways of delivering content to people that draw a substantial number of subscribers away from SXM. The much more likely assumption is that they will.

The cost of launching satellites is astronomical and has to be capitalized over a long period of time. The maintenance costs on them may be minimal but satrad providers keep having to launch them, which suggests that their lifespan is not as long as satrad developers hoped at first. The particular problem with not covering capitalization costs until well into the future are that entertainment technology product cycles are shortening. What entertained people ten or twenty years ago no longer does as much. So financially successful entertainment products tend to be those that can catch fire and make up their capitalization costs more quickly than satellite radio and make them up before the product gets surpassed by something else new. (iPods and DVD players did this. Satrad still hasn't turned a profit, 8 years in. Not a good sign.) That long, deep capitalization is a significant disadvantage for SXM and probably at some point will do the company in, unless they can get their capitalization costs way down from where they've been from satellites that last longer or lower-cost launches.

Oh, and adjensen, if you want to link people here you're pissed at in the threads, you need to lead with your lead. Many of them are not going to read more than a few sentences in before their eyes glaze over and they hit the Back button because your journal article doesn't seem relevant to their conviction that in six months they'll be able to get niche programming over their cell phone/wireless netbook/GPS/mp3/video player that fits in their hip pocket in Pigkisser, Iowa. Either hook them in with your conclusion so they want to read more or prepare to be ignored by many of them. (OK, many will ignore you no matter what. But you have to hook 'em in faster to have a chance to hold their attention.)

[SenorXM]

[Brian202]

hi sir. i do not have the time nor the attention span and certainly not the care to read all that. you think you could sum that up in a paragraph or less? thank you.

[xmguy2500]

Streaming over a cell network IS possible. IF the voice network were shut off. Then a system simply let the radio handle the buffering. So no matter how many users go online. It wouldn't bog it down since all it's doing it pulling an existing data stream so it's no more load than just 1 or 1000 users. Just log into an IP stream just as you would tune your radio. I for-see a Satellite/EVDO/LTE cellular radio network that rivals current tech.!

[adjensen]

[adjensen]

[Brian202]