Friday, July 02, 2010

Apple and the problematic signal bars

New Apple press release about the problematic mood-swinging signal bars in the iPhone 4. They actually say:
"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."

Formula? Formula!? How complex can it be? It's not like the strength of a signal is a multi-dimensional thing, it's a straightforward two-dimensional scale.
I think Apple will hear for this one for a while.

Update: top shelf tech journalist (stauch Apple supporter through decades) tweets: "Is it just me or does this Apple statement sound like an Onion article?" Yeah.

Update Bruce pointed to John the Fireball translating Apple's letter.
"Daring Fireball"? That's a site title so lame that it has to be "ironic". Yes, I'm putting quotes around "ironic" because I mean it ironically. Meaning I think it's one of those cases where the author claims irony to those who can see the lameness, while at the same time appeal without irony to those who think it's kewl.

8 comments:

Robb in Houston said...

It's amazing how many products are on the public market that consumers get to test, instead of the developer/producer not being in such a damned greedy rush and testing it before it's put into the mainstream.

Pop! said...

Your phone is lying to you.

Philocalist said...

Was reading today that there are apparently already lawsuits (in the US, where else?)suing for damages precisely because of this problem!

Timo Lehtinen said...

Formula? Formula!? How complex can it be?

Not necessarily complex. But certainly you need a formula even for a simple mapping. Basically, the five bars do not represent the full available range of signal strength, but only the lower part of it. The reason for this is that when you have only five levels of indication, a linear mapping to the full dBm range would be less than useful to the user.

The following explanation comes from Simon Byrnand:

To give a sense of scale, −113dBm is approximately the weakest 3G signal an iPhone 3G/3GS can maintain a call or data connection with – this corresponds with 1 bar. −100dBm (13 dB stronger) corresponds to 5 bars, with 2 – 4 bars proportionally in between, so the 1 – 5 bar range is only a 13 dB signal range – a relatively small range compared to the real world range of received signals.

Any signal stronger than −100dBm would still read 5 bars, and if you were within a few hundred yards of a cell tower it’s possible to get a signal as strong as −70dBm (or more) – a full 30dB stronger than the minimum signal necessary to show 5 bars.

In that instance you could have almost a 30 dB loss in signal without the reading dropping below 5 bars – therefore no visible change to the user or dropped call.

Bruce said...

There is a funny "translation" of Apple's remarks at Daring Fireball.

http://daringfireball.net/2010/07/translation_iphone_4

Alex said...

Everyone else has said it already, it's a non-linear curve, and you just need the thresholds to make it meaningful to a user.

There is also a range of what the radio will report for you to display, with the end values being greater than and less than, so you only get a lump in the middle and you have to work with that.

The other tricky one is battery life expectancy. Can't just measure the voltage as you have a strange 'S' curve to correlate to percentage of battery life.

ganesha games said...

I noticed the same erratic display in the iPad's wi-fi bars.

TC [Girl] said...

Interesting new development 8/07/10:

http://www.nytimes.com/2010/08/08/technology/08apple.html