Apple has been listening to the reports of iPhone 4 cellular voice and data reception problems, which seem to be related to the visible antenna junction on the flat edge of the case. The company released a statement saying that it discovered (to its surprise) that its algorithm for calculating signal strength for display as one to five bars has been incorrect since 2007. The statement includes the wonderful quote, "Upon investigation, we were stunned to find that the formula we use to calculate how many bars of signal strength to display is totally wrong."
The iPhone 4 has been dogged, seemingly from the moment of release, with complaints that holding the phone in a particular way would cause a dramatic reduction in the signal strength bars, as well as more dropped calls, degraded voice quality, and reduced data throughput.
In response to one customer complaint about this problem, Apple CEO Steve Jobs replied, "Just don't hold it that way." That seemed like a snarky and smug reply, but since then cell phone manuals and suggestions from other manufacturers have been trotted out to show that they often advise how to hold the phone.
For the record, I, like many others, can't duplicate the problem, but I have dry hands. There's some suspicion that a combination of moisture, skin conductivity, hand size, and the particular manufacturing properties of each phone combine to cause reduced signal strength.
AnandTech took a thorough look at the iPhone 4's antenna performance through a clever workaround that avoided jailbreaking to show testers the actual decibel measurement (dBm) used for signal measurement instead of relying on the presentation of graphical bars. (This number could be enabled in iPhone OS 3 in a legitimate mode, but not iOS 4; the writers enabled it on an older iPhone, and then used that to restore an iPhone 4.)
The reviewers found the iPhone 4 antenna was far better in low signal conditions than the iPhone 3GS, but that the way you held it could make a big difference in attenuation - the reduction in signal strength due to an obstruction.
But the writers also discovered that there was a significant discrepancy in how the iOS reported signal strength in bars relative to their own measurement using numbers. And that's what Apple acknowledged in this unusual statement.
Apple said that the problem is a calculation error that dates back to the earliest iPhone. A small change in signal strength could cause two bars of signal to disappear instantly. If you relied on the visual display instead of, say, a dropped call occurring or seeing network throughput go down, you would assume the phone or AT&T's network was at fault. Apple will be adopting a new formula recently recommended by AT&T for calculating how many bars to display for signal strength.
The new formula means that you'll likely see more situations with two or three bars instead of five - something I wondered about when I was seeing five bars in clearly marginal reception situations. Apple said it will make the first three bars a little taller, too.
If I put on my cynical hat, I would say - as John Gruber does in his slightly NSFW translation from marketing into English of the Apple letter - that Apple clearly favored an algorithm that erroneously showed better performance than what existed. Making fewer bars taller will make reception seem "higher," too. However, given that the frequent two-to-three bar drop annoyed customers, it's hard to ascribe too much motivation to the marketing explanation.
An update in the next few weeks is planned for all iPhone models except the original 2007 model, ostensibly via an iOS patch. It's a shame that Apple won't push out a 3.1 fix for the original iPhone and those that choose to stay with 3.1 for the iPhone 3G. Perhaps the company will be embarrassed into making that fix, as well.
Fixing the graphical presentation of signal strength doesn't address the "hold it this way" problem, of course. The AnandTech reviewers, among others, have said that applying an extremely thin coating over the antenna would have likely resolved the problem. They and others have called for Apple to provide a free bumper case to any who complain, too. A class-action lawsuit is already underway as well.
Apple says, "The iPhone 4's wireless performance is the best we have ever shipped," and I believe that. But it's also a specious statement. The issue isn't whether, in a lab or overall, the iPhone 4's wireless performance is better than previous models; rather, how does the iPhone 4 perform across a typical variety of real-world conditions?
In marginal situations, without a case, held in a particular way, it clearly performs more poorly than it could, had engineering choices been made slightly differently. A few thin layers of insulation or another simple solution might have given the iPhone 4 better performance in the worst conditions - no matter how you held it.