John Gruber from Daring Fireball has already written his review:
> “Next year it will work great” should be Bluetooth’s slogan.
> “Next year it will work great” should be Bluetooth’s slogan.
Bluetooth 4 is actually not bad at all in my experience. Pairing is painless relative to old versions; sound quality is flawless; connections are quite reliable. The "strain" is to be expected around other devices since Bluetooth uses 2.4GHz, a very crowded band. I can't speak to the technical bits since I never looked at them, but my experience as a consumer has been positive.
Sure, it doesn't have the same range as a WiFi router, but it also doesn't use 1.5-20 watts (My source: http://www.tpcdb.com/list.php?type=11)
BT needs fewer features and better implementations. But apparently interop and reliability are not sexy enough for the BT standards folks to worry about. Can you imagine TCP/IP being in the same quality hole that Bluetooth is in?
I don't blame the Bluetooth standard though, messy as it might be. That's clearly terrible implementation on the car manufacturer's part.
Given the types of products that are made with BT support, the hostile environments that they're developed within, and the common after-sales abandonment of support, BT should be a simpler protocol that is a LOT more resilient to implementation mistakes. But I don't think that anyone designs protocols to withstand institutional failure :-)
When you combine randChip with randBtStack, the end result is not necessarily a quality product. The poor engineers who have to work with this combination end up resorting to sending reset commands to their chip in an attempt to get the damn thing working.
The two issues I saw were Bluetooth is based on a frequency hopping radio using a dated low performance modulation scheme. The only nice thing you could say is potentially you can build really low power radio's using that design. However that brings up the second issues, a very complicated kitchen sink base band which requires a complex software stack and power hungry processor to support it. There goes your low power, out the window. (I think there were about 20 companies trying to design low power Bluetooth transceivers circa 2000-2003, most of them never could get the power low enough or a working stack)
Hardware design issues aside, frequency hoppers have insurmountable quality of service issues for most applications. Professionally I don't think this is fixable.
They would have been better off doing the following three things.
1. Define a low speed point to point oriented modulation scheme that can be supported by 802.11 radio front ends. Probably could have implemented this as a sub-band of some sort.
2. A direct sequence spread spectrum physical layer for USB. Advantage of that is the USB stack mostly works as USB had a 10 year head start and more resources than the Bluetooth alliance ever had, or will have. If one thinks a stack like USB is 'easy' consider Microsoft had to redesign theirs three or four times.
3. Dedicated DSSS wireless headphone spec. Seriously this was the original design goal and was totally achievable until the committee responsible for the spec lost their minds.
Don't get me started on security. BT is probably a festering pit of remote exploits and vendor backdoors.
That's the point. If you want a heavily opinionated summary that will likely inflame a fair portion of people, Gruber is your man.
I have a headset that supports Bluetooth 3 only and audio quality is bad (especially while making calls where it chooses the HFP profile which is the only one where the mic is available).
I vouched to never again buy any Bluetooth audio devices. Are you saying Bluetooth 4 would actually solve my problem? Should I still have hope?
Make sure your headset and device (phone?) both support HFP1.6 or above (which supports wideband speech/HD voice) which makes the sound quality semi-decent
With Bluetooth in particular it seems like there are a lot of poor implementations (if it's rock solid or buggy seems to have to do with which two devices you're using ) - maybe they would benefit from a more stringent conformity licensing process.
1. Reliable 2. Cheap 3. Low power
Pick no more than two. Bluetooth was originally designed to be cheap and low power.
The other problem is networking is complicated, and then throwing device API's on top of that makes it very hard to get right. USB had lots of issues originally. And once you leave standard devices like, keyboards, mice, and mass storage devices it gets iffy real quick. (Go scan through the libUSB mailing list archives sometime)