Bluetooth Headset Sound Quality Comparison v.1.1
updated with Jawbone and Mogul Handset ratings/files 3/31/08

So you think your bluetooth headset sounds great?  You can hear your caller just fine and the headset's maker has all but assured you that the person on the other end can't tell if you are on a headset or better yet, in a noisy bar or at the game.  There are online videos for several headsets that even show them being used in a loud environment.  After getting what I consider to be less than ideal performance, I decided to create a little test of a few headsets I managed to gather.  While some nice folks have posted a few sound clips here and there on a few forums, I haven't seen any comparison tests with sound clips in many different environments.  I decided to give it a go and the results were pretty interesting to me.

The Goal of this test is to find a headset that works well in all environments and presents the user with a clear intelligible voice.  Clarity is a priority over (true to life) voice quality.  Some headsets clearly sound much better, especially in the easier tests, however that does not mean they faired better in overall ratings.  The true test is how easy the voice is to understand.  Incoming sound quality isn't really a consideration of this test as its pretty easy to tell if a headset works for you or not, although I do have some comments in the notes section for each headset.  They are listed in order of my personal preference, however you should listen and decide what works best for you if you are shopping for a new set.

Test Parameters
Each test was performed twice, in the case of the quiet room, I used a lower voice for the second pass as to test the volume at a barely audible volume. For all other tests I spoke as loudly as I could without yelling on the second pass..  The loud voice would definitely get you dirty looks, but if you have to make the call, sometimes it is the difference between getting work done and not. Sound levels were measured using a Scoche SPL-1000 sound meter, while not the perfect meter for absolute measurement, it is adequate for relative comparison.  The sample sounds were played at the indicated volume and a call was placed using my Sprint Mogul handset.  I left voicemail for myself and the attached files are a result of that.  They are not edited in any way except for length to trim excess silence/noise.  The terminal, street and industrial clips are made using sound bites, the vacuum sound is a genuine Hoover and simulated wind sound is created by sitting in front of a cheap old box fan (that is very loud) and I'd say it puts out a steady 15mph breeze at the testing position.


Quiet Room, 50db Airport/Train Terminal, 72db Street, 80db Industrial, 95db Vacuum, 82db Simulated Wind General Notes
Ailph Jawbone
Very clean, even when speech is quiet, very crisp and clear sound. Sample Very clear, with a hint of the background noise at regular level.  At loud speaking volume it disappears. Sample Very clear, again, a hint here and there, but still very nice and clear. Sample At a normal speaking volume, not really usable, however at a loud volume, speech is quite understandable. Sample A low speech volume you can hear the vacuum,  at a louder volume it sounds fine. Sample Clearly the weak point of he Jawbone.  Wind is as bad as the AX2 but you can at least hear the speech. Sample Clearly the sound quality champ of all the headsets.  Works great everywhere except in wind.  Speech is very clear and crisp, both outgoing and incoming.  Conversely, aesthetics are a nightmare, its bulky and heavy.
Motorola H700 Loud but not always clear.  Picks up quiet speech well.  Sample 1 Sample 2
Loud and intelligible, but seems to sound muffled.  Background noise isn't very loud but it doesn't sound like silence. Sample 1 Sample 2
Gets more muffled than the 72db, however when talking loudly, most background noise disappears.  Sample
Not usable at normal speech level.  With loud speech, somewhat understandable. Sample
Muffled with some clipping, gets better at louder volume. Sample
About par compared with the rest.  Again, works better with louder speech. Sample
Not always reliable, seems to differ even using the same parameters, see included files.  Generally works very well if your voice is louder than the background noise.  Very strong performer on incoming sound quality, loud and very clear.
Plantronics PLT510 Slightly muffled but consistent level and quality. Sample
Fairly clear but you can tell that the user isn't in a quiet room. Sample
Louder sounds come through, but it is difficult to tell exactly what they are.  Speech  is clear and understandable. Sample
Speech is barely understandable, however the background noise overpowers it a a low volume.  With louder speech there is severe clipping. Sample
Passable, sound like its just a not so clear headset or phone. Sample
Lots of wind noise but the speech is understandable. Sample
The 510 uses its boom mic to get closer to your mouth, old school voice isolation.  Works well in most instances, but you still can hear the background noise behind the speakers voice.  Strong incoming sound, very loud but slightly muffled.
BlueAnt Z9 Clear, but it just doesn't sound very human. Sample
Pretty robotic sounding at normal level, sounds better at loud speech level.  Does a fair job of blocking out most of the background noise. Sample
You can hear some noise, but you can't tell what the noise is. Sample
Background noise overpowers lower voice.  You can make out some words with a louder voice.Sample
The noise is isolated, however so is the speakers voice on certain words.  Louder speech seems to help. Sample
Some noise, hard to discern what type though.  Voice seems more muffled. than max isolation. Sample
By stats, a solid performer, however the robotic sound is a big turnoff.  Muffled incoming sound doesn't help either.  Very loud incoming, however the circuit is amped even when the headset is not active, and picks up RF noise easily.
BlueAnt Z9 (max isolation)
A little more quiet than most, most robotic sounding of the bunch. Sample
Sounds like garbage at normal sound level, loud speech is just passable. Sample
All but the loudest sounds are isolated but speech just doesn't sound human. Sample
Garbage.  I can tell that someone is speaking, however it not intelligible at all at either volume setting.  Also variance between two samples Sample
The noise is definitely isolated, however the user sounds like he is at the bottom of a pool. Sample
Some noise but its not readily apparent that it is wind noise, slightly better than std isolation.  Voice sounds robotic. Sample
Max isolation does buy you a little more constant noise reduction and maybe a little wind reduction, however its not a miracle by any means.  The price paid is your voice sounds like a robot.
Jabra BT500






BlueSpoon AX2 Very clear although some digitalized sounds can be heard. Sample
You can hear background noise, but it is not very apparent as to what it is. Sample
Some minor clipping, but background noise isn't too loud yet. Sample
Forget it at regular volume.  At loud volume you can hear a few words, but just not usable Sample
Lots of noise, but you can understand the words. Sample
All wind, very little voice.  By far the worst. Sample
Very natural sounding when its is quiet but falls on its face with more noise.  Strong incoming sound but crippled by lots of static.
Jabra BT2020 By far the most muffled.  Mic location kills this headset. Sample
You can hear there is noise in the background and the headset is beginning to show its real weakness, background noise. Sample
At a regular volume, severe clipping.  When speaking loudly, it is just barely passable. Sample
At a lower voice, you hear nothing but background noise.  At a louder volume you can make out 1 of 10 words? Sample
Garbage.  No one is making out anything. Sample
Fair amount of wind noise, voice is very muffled. Sample
For being a newer headset this one is MISERABLE.  Shame on you Jabra.  Putting the mic under the users ear lobe just does not work.  This is the only headset I would clearly say stay away from.  Muffled incoming sound also.
Motorola H500 Very deep and muffled.  Average sound level. Sample
Background noise is pretty apparent, speech is very muffled. Sample
Muffled but still understandable. Noise is fairly apparent. Sample
Pretty much useless.  Just noise. Sample
Only barely passable at a loud speaking volume. Sample
Almost as bad as the AX2.  Useless in this test. Sample
Shame on you Moto.  This headset is a few generations newer yet does worse than your first gen HS820?  Not loud enough either.
Motorola HS820 Shows its age, even in the easiest of tests you get noise/breakup and not so good quality. Sample
Background noise is strong.  Getting close to marginally unusable. Sample
I'm baffled that this sounds better than 72db test.  Fair amount of noise, but speech is ok. Sample
It doesn't sound good, but you can make out words.  Voice is very quiet though. Very shocked here also.  I was expecting nothing but trash. Sample
Fails pretty miserably at a lower volume and not much improvement at a loud volume. Sample
Lots of wind noise, however most of the speech is understandable. Sample
Good performer considering its age but the incoming sound is just not loud enough for louder environments.
Mogul Handset Very clean and crisp.  All nuances of a persons voice are audible. Sample Clean and clear, but background noise is somewhat apparent. Sample Again, clear and crisp at both levels Sample. Lots of noise apparent, voice cuts out at both sound levels. Sample Fairly clear, but the vacuum is audible. Sample Clear winner.  While you can hear some noise, it is not apparent that it is wind. Sample Simple physics make this a solid performer in the <80db tests where the speakers voice is louder than ambient noise.  Also the most natural sounding.

Color Legend:
Green - Go ahead and call the boss and tell him you are in the office.
Yellow - Normal conversations are no problem, should sound like a normal call with some noise
Orange - Fine for calling your buddies, but don't call the boss or a client
Red - If someone is disparate,they might be able to pick out a the words and piece the puzzle together
Purple - Forget it, send a text message or find a better setup.

Surprises/Rants
How is it that the HS820 does better than the much newer HS500?  Seems like Moto really dropped the ball on this one. I was really disappointed in the BlueAnt's performance.  I expected much better based on the online video.  Sure you get some sound isolation, however it is at the price of altering your voice, even at low volume.  I wanted this headset to work but there are too many sacrifices to pay for some marginally effective noise reduction.  The Jabra BT2020 is plain crap.  I even sent it back thinking it was defective and got another that was just as bad.  Thinking it did not like my phone, I tried another.  Same result.  A lot of people complain that the Jawbone just isn't loud enough, in my experience it was fairly loud, not as loud as the BlueAnt or H700 but it was louder than most.

My biggest rant is not related to sound quality though, it is more about charging the batteries on these things.  There is no excuse for using non standard charging ports.  The majority of phones use mini-usb, use that, even micro usb, I don't care, use a standard.  This world already has enough electronic waste, more chargers are not necessary, the only reason for alternates is pure greed.

Conclusion

Before testing the Jawbone, my conclusions were quite different.  It is easy to think that voice isolation might be effective on the Z9 if you believe the marketing.  However once you use the Jawbone, it is clear what effective noise isolation really is.  The fact that the Jawbone works in the 95db test at loud speaking volume is pretty amazing when you think about it.  The ambient sound level is louder than the speakers voice.  At the noise level, I can hardly think, yet if I speak loudly, the sound is fairly clear.  The Achilles heal of the Jawbone is the wind test though.  If you live in a windy area, I might consider another headset.  The bulky design makes it susceptible to a lot of wind noise and though it doesn't matter for testing purposes, the Jawbone is big AND ugly!  The Z9 is a fairly solid performer, but frankly, the Jawbone is that much better, especially when it comes to sounding natural. The Plantronics PLT 510 really is a fine headset for voice quality, both inbound and outbound.  The real advantage is its physical design.  The mic is much closer to the users mouth.  Move any mic  further back and you either get more background noise or you have to snip a part of the audio spectrum at the price of making the users voice sound more muffled.  Being close to the users mouth is also how the Mogul handset is able to perform well also.  It is so close the the source of sound that it is able to overcome noise of about 80db or less, however the 95db test shows that it is only physical noise reduction.  The Jawbone is actually able to selectively transmit only the users voice. I personally really like the Moto H700 since I don't use a headset in really loud environments, if I did, I'd carry the Jawbone.  But then again, it seems like everything in the world of Bluetooth headsets is a compromise, the intent of this exercise is not so much to guide you with my thoughts as much as to let the sound samples guide you.

Comments or Contacting me

Please use my blog to post any comments.  If you have questions pertaining to the test, that is also a good place to post them.  Please don't email me questions about getting your headset to work with your phone or problems you've had with a particular unit, there are several websites that let you post comments, reviews and get help. Emailing me here about them helps no one.  If you really must....  danREM0VE@datasmuggler.com

What's next?  I plan to try out a few more headsets, in particular I would love to try the Invisio Q7.

Notes
Motorola H700 tested was the version B.  I have tested a version C in a quiet environment and the sound quality was nearly identical.  It is rumored version C has some compatibility fixes. BlueAnt Z9 has firmware 3.3


All content is copyright of the author - 2008 - This page may be linked to, but not posted somewhere other than the source site, datasmuggler.com