This story is part of the ongoing Qualcomm PR effort to get people to care about what modem is in their phones [1]. Of course no one actually makes a purchasing decision based on what modem is in their smartphone.
Not sure why so many are pooh pooh'ng this. I consider the quality of display and the speed of modem two most important aspects of modern smartphones. Not everyone lives next to cell towers or in high density areas like NYC. I actually do live in NYC, but in less desirable area where avg download speed is ~2mps (Oneplus 5T on TMobile), but it was much slower and frustrating on my older phone (Note 4 on AT&T).
The speed you get is far more dependent on the signal strength, not the speed of the modem. That only comes into play when you have a perfect signal, so realistically (as a perfect signal is rare), modem speed plays almost no part in speed limitations.
It does come into play in weak signal situations. Consider support for more spatial streams (4x4 mimo) helps with reception in cell edge areas and when noise levels are high or multipath interference is a factor. Also every modem/antenna pairing has limits programmed in where they'll drop a cell or switch to a different layer (band) with less bandwidth.
No iphone uses 4x4 mimo. Maybe for battery, or maybe Jony couldn't make it pleasing looking enough in the X, but its lack is very much is a factor affecting quality of service including speed in non-ideal environments.
Are there any reliable objective numbers and comparisons for this? While mimo is a great technology, it is trivial to “hold your iPhone right way” and get the same benefits, I feel mimo has become a marketing ploy for companies to push their latest without any perceptible impact to users.
There’s a perceptable impact to users with mimi and it specifically takes advantage of multipath. This is not a case of marketing fluff.
However, 256QAM is something you’ll probably only see in ideal situations as it requires very good signal, that’s the one that edged out gigabit speeds in the lab.
My service is regularly shit on America’s best network in NYC. iPhones used to be very good at this. Now I can’t make phone calls/drop in the same location, on the same cellphone.
LAA simply seems like the cell providers way to charge me for use of frequencies I could otherwise use for free.
A key reason I'm happy to pay cell providers so much money is because they have licenses to use lots of spectrum at high transmission powers to give me service nearly anywhere.
WiFi, it's free cousin, has lower transmission power, but in most cities one is always within range of free WiFi (even if roaming between hotspots is painful!).
When cell providers are allowed onto the free spectrum, they will use up all the available bandwidth, degrading free service and forcing users to pay for their (normally free) bandwidth.
> Of course no one actually makes a purchasing decision based on what modem is in their smartphone.
I do. I have a phone I use primarily as a modem and I use some tools that lock bands so I can get more reliable speeds. As of this time I can only do this with Qualcomm chipsets.
Not that I'm a Qualcomm fanboy (being the first to comment in that this was basically their advertising) but I'm extremely wary of using Intel modems due to previous experience with their other modem chipsets sucking ass. [1]
Oh and there’s also qualcomm’s leaked qxdm and qpst tools for unlocking bands supported by the chipset. (and modifying stuff like carrier files for voLTE)
Didn't realise qxdm had leaked! Had to use that a lot at my last job and while it's a super powerful tool, it really provides insight into both how complex baseband systems are, and how buggy Qualcomm's firmware is. There are so many security exploits it's almost unreal.
I bought an iPhone X 100% based on the Qualcomm modem. I have the Apple store CDMA qualcomm version even tho I'm on T-mobile which had the intel version available sooner and cheaper. T-mobile service in San Francisco is absolutely terrible. Doesn't work deep inside most buildings. In the past I went through a bunch of iPhone 7's on T-mobile due to loss/theft/damage. Also compared many side by side. Some with qualcomm, some with intel, always on T-mo. Signal was a little better with qualcomm, battery life was SIGNIFICANTLY better with qualcomm. I'd guess 20-25% which is the difference between the phone dying by the end of the day or not.
Maybe I'm still in the honeymoon phase with my iPhone X, but I've had it for a few months now. The last iPhone I owned was a 4S, and I've been on Android ever since. I was very happy with Samsung's Galaxy line from the S4 Active to the S7 Edge, but the Edge simultaneously slowed to a crawl and lost at least half its battery life after 18 months. It was to the point that I had to charge it every time I was sitting down.
The iPhone X has been great. The interface is fast and gets out of the way, and Siri is actually starting to become useful for more than a parlor trick.
FWIW, I'm writing this from a bluffline overlooking a lake. My MBP is tethered to my iPhone, and I'm getting a consistent 50Mbps. I don't know that additional download speed would be useful to me at the moment.
I feel exactly the same. The only smartphones I ever owned were Android (and a single Windows phone), but my experience with the S6 was the same as with your S7 - and the loss of the headphone jack didn't matter to me, since the one in my S6 ceased to work consistently about a month after my warranty ended. The X feels like a device at the cutting edge, FaceID is the closest I feel to "intuitive computing" outside mixed reality devices - there is almost no tangible barrier between you and the connected digital environment. But it's also stable, fast, and as you said "out of the way." I miss some aspects of Android's action tray and notifications, but have been far more impressed by the ways iOS is leaps and bounds ahead as a general daily user experience. And I say all this as a .NET developer who daily drives Windows on my desktop and laptop!
I suppose that was a bit off the subject at hand, but I was really surprised at how much better the iPhone experience feels than Android. More on topic, I really don't think this kind of crowing over technical details will help the platform any more than flexing about camera megapixel resolution or processor clock speed does. Android needs to make strides in its user experience as a daily computing device to replace iOS for me now, and I also haven't noticed any issues with the speed of the web. I suspect if there were any differences, the rest of the experience would make them unnoticeable to me.
I have to agree. My own anecdote is upgrading from an iPhone 6s to an iPhone X. At no point was the 'Internet Speed' a consideration on my purchase decision.
For me, I accept I have to invest every N years in order for my hardware to match the ever-increasing demands software places on it. It's a tax I am willing to pay for a device that has become so ingrained in my lifestyle.
I suppose, for me at least, the network speeds are in a similar category to processor speeds for normal, every day use. Would an extra 10-20% download bandwidth be appreciated? Would an i9 running at 2.9GHz help me transpile and package my JavaScript quicker? Sure. But it's way down my priority list in terms of real-life productivity and time saving.
> I suppose, for me at least, the network speeds are in a similar category to processor speeds for normal, every day use. Would an extra 10-20% download bandwidth be appreciated? Would an i9 running at 2.9GHz help me transpile and package my JavaScript quicker? Sure. But it's way down my priority list in terms of real-life productivity and time saving.
Definitely. Furthermore, I typically don't even care about things going fast. I want things to "not be slow". Which, typically I guess the distinction is that I don't mind expected slow things being slow. Ie, if I have to run integration tests I expect it to take 30s. If I got a (somehow) faster SSD and CPU could it take 20s? Maybe, and sure that would be great.. but .. meh. It just doesn't matter to me.
However the things that should be instant, that I don't like to be slow need to remain so under all conditions. In my experience, RAM is the biggest culprit for causing simple things to be slow. If I open one to many browser tabs, suddenly I have no RAM and my UX goes down significantly.
So RAM ranks far, far higher on my list than CPU. CPU rarely has a big affect on my these days, and unless I switch to a workload where I'm heavily concerned with shaving time off of hour+ compiles (video editing/etc), then I just don't care. But RAM, oh boy do I love RAM.
In my opinion, no. I'm on the AT&T "Unlimited Choice" plan, which is artificially capped at 3mbps and it's been totally fine for me. I can do everything I want to - check email, browse the web, check Twitter, use maps, stream music and podcasts, etc. Also, the bandwidth cap helps me avoid hitting the 22GB monthly transfer cap and getting throttled to the very noticeably terrible 128Kbps.
In an ideal world I wouldn't have to worry about bandwidth or transfer caps, but I mention this to provide my experience that for me 3mbps is plenty. Some people might need more, but I think that once you get up above 20-30mbps, most people wouldn't notice or care about the difference.
I'm on a T-mobile unlimited plan and I can use 5Gb per day after which I have to start spamming the T-mobile app if I want to use more.
Its kind of a trap, once you get unlimited data and high speeds you start to take it for granted. Nothing is good enough anymore. But when you get abroad on holiday you learn how the rest of the world lives and know that you should be grateful.
> I'm on a T-mobile unlimited plan and I can use 5Gb per day after which I have to start spamming the T-mobile app if I want to use more.
One plan? Or Simple Choice.
I've had no problem pulling ISO's down in the evening to my phone (like 15GB overnight) on SC unlimited with tmo. It's not something I make a habit of but I've found out deprioritization only really affects me in a few busy places in manhattan and around international airports. (and this only kicks in after 50GB of usage or so in a month which is not typical)
I'd be more interested in worst-case performance. If the Samsung and Google devices maintain a similar performance advantage with very marginal signal strength, it could frequently make the difference between a useable connection and no connection at all. I suspect not.
> Ookla’s data ... iPhone X ... 31.5 Mbps ... 25.1 Mbps.
> The Galaxy S9+, ... 38.2 Mbps on Verizon and 34.2 Mbps on Sprint.
If you buy a phone for this feature, then I suppose it would matter. But IRL no one cares about these performance numbers until/unless they're separated by an order of magnitude.
> I don't know that additional download speed would be useful to me at the moment.
Until someone comes up with clever new applications to consume this bandwidth, I agree: it's just not useful.
I really prefer android and I like the Snapdragon phones. Clearly, AAPL made a smart bet in dual-sourcing the modems. People do not buy iPhones because they expect superior throughput. Device performance/latency likely matters much more than network performance.
You're assuming something about the distribution of speeds. If they're mostly the same but some are significantly and unusably worse, that is different from the speeds being mostly slightly lower.
I have an X on tmobile and it's abjectly terrible. I had a 6S+ before it. In the same location, on the same network, it will drop signal or become unusable and gets lower reception almost everywhere.
This is nearly as bad as the issues reported with the essential phone, honestly. It's a night and day difference you can compare side by side if you have two devices.
With good signal it's fine, and the speeds are plenty fast. I have no complaints about battery life. It's entirely when the signal is marginal that it's garbage.
I've considered selling it and buying the verizon or unlocked qualcomm model multiple times.
When signal is poor it's likely locking to the band 12 layer (700mhz) which is only 5mhz x 5mhz worth of bandwidth. It congests easier than band4 which is typically 20x20. For whatever reason iphones seem to be a bit stickier to this band, probably due to weaker antennas or firmware that kicks it off the beefier bands faster.
There's something to this theory, as switching airplane mode off and on usually results in a usable connection. This is still completely unacceptable behavior though, and i haven't been able to replicate it on multiple other devices i have handy(including other iphones, including ones with band 12 support)
My experience with the SE has been that it goes to band 12 first or if its moving fast within a cell. Then it ‘settles’ into band 4 or 2 on tmobile after about a minute.
My old modem phone I purposefully got without band 12 because it saturates so easy, now I just lock bands and speedtest until I find most reliable with the current.
My experience with the X is very different. It's entirely possible you're experiencing duff hardware. Have you approached Apple with your concerns? Remember you may have consumer rights depending on your location...
Same here. I was on Nexus for the longest time and iPhone X was first iPhone since the first one back in 2008. No issue with speed at all. I use it to tether to my MacBook quite often and speed been reliable for me.
Ah, but that's just it - I didn't pay $1000 for it. I pay monthly installments over two years, which is how long I expect to have the phone. At the end of the two years I'll sell the iPhone X for ~$400 I expect and upgrade to whatever is newest at the time.
I consider it more of a "self-administered lease" than a purchase. My total cost for the device will end up being about $600 over two years, or $25 / month.
In which metric? Don't think you get the fastest in all metrics for that low a sum. Especially with all the other top-in-class features that also matter to people.
Yeah you're right. It's hard to put a value on these things especially when people value things in different ways. On the other hand, I'm pretty sure Apple makes quite the margin on the iphone x.
It's a rather insignificant fact. Sure, if those numbers meant some real-world thing, then it would be good to know, but in this case, it's going to be hard to find a normal user that actually cares.
On top of that, location, networking and movement matter, so you could easily simply uncover living habits of users instead of the actual device's speeds. If (and this is unlikely, but bear with) all iPhone users would live outside of extremely well covered areas then they would get different LTE speeds. Doesn't say much about the phone, chipset or provider, just about location.
While the article touts Ookla as a good idea because it's real-world speeds, they should also have mentioned that it has no real-world meaning either, as well as having real-world numbers not meaning much in comparison between providers, phones, and locations.
At the same time, if the manufacturers of phones use different modems and one is slightly faster than the other, good on them, but who cares?
Since this is obviously a fanboy war post: I prefer Apple because I feel like their products are more user focused. Android and Google's mobile environment feel like a platform on which they can collect data and sell you ads (not that Apple doesn't collect data, but it isn't their primary business model). It also seems like Google will write new apps whenever they don't want to deal with technical debt of the older ones, or they want to use whatever fancy new platform level service they have that lets you sync 50,000 nodes cross datacenter instantly using autonomous submarines or whatever.
And I long for the day when fan-boys in both camps can admit that all current phones are actually quite good. If smartphone development stopped here and we never got a new smartphone, I don't think it would matter much overall.
It probably could have stopped a few years ago, actually. Perhaps in the iPhone 6/Galaxy S6 generation.
And Apple's seems like it is where they baby you by not allowing you to install other browser engines, set the system defaults for email,
maps, camera, nor let you install non approved software like anything adult.
if you weren't allowed to install your own browser engine on other platforms we'd all be stuck with IE.
There's a reason internet standards progress so slow on iOS. It's because Apple disallows any competition. If they allowed real Chrome and Firefox they'd be forced to keep up. Users would have a choice. Chrome has proven to to be around 10x as secure as Safari on macOS based on exploit report. There's every reason to believe it would be the same on iOS. I'd prefer to be able to choose the more secure browser with more modern features. that's would be a pro user feature to me.
I pay Apple to use iOS, so I want them to maintain 100% control over iOS as a platform. If I wanted a balkanized platform, I would buy an Android. I don’t say that to disparage Android, btw. I think the desire you expressed is completely valid. It is just very much not what I want.
But I mean... if you don't want to use a web browser with a third party rendering engine, you could just not install it. I don't want to play Flappy Bird, so I didn't install it. Apple would have allowed me to, however. I don't see the difference.
As other comments mention, all a mandatory browser leads to is "you must use IE6 to visit this site" which was a very terrible world to live in. It isn't happening now simply because there are more !iOS devices than iOS devices.
> If you don't want to use a web browser with a third party rendering engine, you could just not install it.
It's not that I don't want to use a 3rd party web browser. It's that I don't want a 3rd party web browser to ever become popular enough to give that browser's vendor the ability to influence Apple's ability to make technical decisions about their platform. Being beholden to 3rd party software vendors was basically the story of OS X for years. I don't think it was good for Apple.
> As other comments mention, all a mandatory browser leads to is "you must use IE6 to visit this site"
On the contrary, as long as Apple is Safari only and Android is mostly Chrome, there is no way that this can happen.
it is happening now. If iOS Safari doesn't support something its basically no go period. other browsers even on other platforms are effected because web developers can't use the new apis given that 20% of the market has zero choice.
So Android manufacturers, led by Samsung, have decided to harp on modem speed because they know they are years behind Apple on processor speed with little they can do about it, and the tech press has decided to carry their water for them. In my lifetime I have never heard anyone complain about, or even discuss, download speeds on their phone.
What's next? "The new Galaxy S12 can reach ignition temperature within 5 seconds of power on. Beat that, Apple!"
> In my lifetime I have never heard anyone complain about, or even discuss, download speeds on their phone.
Similarly, I've never heard normal people complain about or discuss the CPU speeds of their smartphone.
Companies will always harp on each others products when they can, and the tech press writes about it because it gets clicks; no conspiracy here, I think.
This might go under your "didn't pay for," but it does annoy me that apple doesn't allow emulators. I bought some SNES games, the console went under, I would still like to play them on my iPhone. I know it probably is a gray area for most people, but none the less, Apple could let it be done, but they don't.
They probably want to avoid the issue all together.
I used to hang around the Dreamcast homebrew scene and the stance on emulators were that they were mostly a proof of concept/educational thing so you could talk about the emulators, how well they performed, bugs, etc, but you couldn't talk about where to get ROMs or distribute emulators with ROMs on the site.
I don't know how that'd work with iOS apps. In order to keep the emulator app separate from the ROM, you'd need to be able to download the ROM to some sort of general storage. Then the app would need to be able to access that storage. I guess you could do it off of dropbox. Point the emulator to some URL where the ROM is stored. Downloads the ROM on demand, etc.
The major point is that the app can't really have somewhere official that says "ROMs be here".
1. All the phone that had the Qualcomm 845 ( X20 Modem ), as tested in those data, were available later then the iPhone X / 8 / 8 launch. So a 6 months old phone had better spec then a 12 months old phone, is that suppose to be news?
2. Your Network speed is fundamentally determined by your Carriers quality, capacity, etc, much more so then your phone.
3. Latency, Latency, Latency! I don't care if I get 20Mbps or 200Mbps speed. I do care if I get 20ms or 2ms latency.
Of course that is for those of us who got decent network, I read in many parts of US you get get sub 10Mbps speed. ( How is that even possible? if you ask most developed Western Europe or East Asia countries citizen they would have decent network by now )
4. We actually need to push the bottom half of the market to support better Modem spec. The more phones support Massive MIMO, LAA or features in later of of 3GPP Rel spec, the faster the overall network is. Which is something I wish Apple had done better.
And Finally, any mid range to high end Smartphone, can consume 10GB + of Data in their theoretical top speed in a matter of few minutes, and that is likely over their data cap / allowance, would a user care more about their Data speed or Data Cap?
Great points. You hit the nail on the head. Once you get 'enough' throughput, getting more doesn't really improve one's mobile experience. Latency is king.
CPU / power usage is another big one. It's hard to justify a gain in throughput if it (i) reduces the application's CPU time, making it render the received data more slowly; or (ii) causes the phone battery to die faster because the network threads are constantly doing heavy lifting.
While I appreciate the size of the Ookla dataset, they need to be more thorough with their metrics. Phone performance is a multi-faceted problem, and throughput alone does not reflect the whole reality. Side note: dslreports has added bufferbloat testing a while ago, I can't believe Ookla still tests latency in isolation.
Seems disingenious to only talk about bandwidth (mbps) and not latency (ping in ms) when discussing "faster internet". I believe it is the latency that is the biggest cause for sluggishness when browsing or watching videos over 4G, not the raw throughput (anything over 10mbps should be more than enough even for HD video)
Definitely need bandwidth and latency, not just one of them.
Add into it being forced to use Google's DNS servers on android and there's probably some performance being lost there. It would be nice to be able to set custom DNS servers on all mobile platforms without having to use a VPN.
> Add into it being forced to use Google's DNS servers on android
Source? My phone runs a current Android (and not some custom ROM, but something standard from it's manufacturer) and doesn't appear to use Google DNS, at least not for normal functions.
"DNS changer" apps often spin up a local VPN to tunnel the traffic through. Folks usually end up rooting their devices to ultimately make this change, or using a VPN.
As well, I believe Chrome can tend to use Google's DNS servers directly in some cases.
* Bandwidth (signal processing) or analog bandwidth, frequency bandwidth or radio bandwidth, a measure of the width of a range of frequencies, measured in hertz
* Bandwidth (computing), the rate of data transfer, bit rate or throughput, measured in bits per second (bit/s)
The article is based on measurement of data speeds but makes bold claims like:
> Faster internet data means that users can load websites and start watching movies more quickly, make crisper video calls and get higher-quality video.
But that's not what their data measured -- it measured bandwidth not page or video loads. Not to mention it goes against the experience everyone's had of a device getting slower with most OS updates that are optimized for a newer device. The bandwidth didn't go down with OS upgrade, so a measurement thereof is pretty meaningless. All else being equal that statement is true, but by definition (comparing disparate devices and manufacturers) everything else is different in this data set.
They do qualify it a few paragraphs later:
> To be sure, it can be difficult for users to tell the differences between how fast pages load on a phone.
I thought that's what you were claiming to be noticeably faster here?
Only to water down the qualification with:
> But it’s easier to sense the quality of video, how long it takes to pull up a song and how long it takes to send an email.
Yes this statement is true in the barest sense, but in the context of the article it's very misleading since none of the data even attempted to measure these things.
And sending an email? Who regularly sends huge files that reside on the phone? And what phone doesn't do that in the background anyway?
Then there's this sentence that gets an award for obtuse phrasing of the week:
> Apple’s main processors that control the speed of launching apps, swiping through PDFs and loading games are often recognized as some of the fastest in the industry.
In an article about measuring speeds differences down to two or three decimals points, it's incongruous to talk about "often recognized as" versus just looking at any of the thousands of benchmarks (also from "real world" users that are out there). And even weirder that the metrics above of page loads and video loads are seemingly unaffected by processor speed.
Bloomberg tech reporting is very clickbaity of late. Or it's a conspiracy to encourage bigger ad-buys from Apple and for them to retreat from ad blocking and tracking protections. Probably not that, but it's more exciting to imagine than the alternative of simple incompetence.
Yeah I’m seeing Qualcomm ad spam / competitor FUS pop up everywhere, google, RSS feeds off a few blogs I follow, twitter and any android fan / toolchain site.
That seems pretty irrelevant as such standalone numbers. The "average" speeds are more than fast enough for useful services, don't delay video preload or pages unnecessarily, ... The interesting aspects aren't captured in such averages:
how do they handle bad connectivity?
latency?
CPU + GPU + OS + apps turning the incoming data into user interaction.
Is it even worth living a life where it takes more than 10 minutes to stuff your phone full of ISOs for every major version of Ubuntu that has ever been released?
Faster internet data means that users can load websites and start watching movies more quickly, make crisper video calls and get higher-quality video.
This is why you don't go to mainstream media for tech news. 4K video streams at the maximum of 7Mbps on Netflix. "Faster internet speeds" won't help with any streaming video once you get past that speed.
I'm struggling to see how these speeds are a direct consequence of certain phones being slower than others. My iPhone 6 happily consumes data at a rate of 100Mbps over 4G according to the Speedtest app. This phone must surely have an inferior modem and antennas compared to the much more modern devices mentioned in this article, yet on average they apparently perform much worse.
Is the device itself really the only variable that could account for the difference?
For example, I only get these speeds on my network, EE, because I pay for a "Max" plan which allows me to download data at the fastest speed possible. However, they sell cheaper plans with capped data rates. Perhaps in this scenario, iPhone users are choosing worse (cheaper) plans so they can afford an expensive phone?
When you're browser is making 500 requests to load a bloated web page, overall download speed doesn't matter. You know what does? Single core performance for processing all that javascript.
Probably a good time to mention that apples processors have such better single-core performance that real world internet speeds are usually much faster.
For example, my buddy has fiber and my Android CPU maxes out at 400mpbs. He gets close to 600 on speedtest on an old iPhone 5
[1] https://www.qualcomm.com/news/onq/2018/07/23/over-million-us...