Here are measurements of the Bowers & Wilkins 686 S2 bookshelf speaker.
It was released in 2014, and discontinued in 2018, and it's two generations behind the 607 S2 measured by Amir. I Think the MSRP was 399.99 USD / pair.
My measurements are quasi-anechoic, with near-field port+woofer, corrected for baffle edge diffraction, combined with gated measurements at 1m distance.
I measured at tweeter axis, since that's what B&W recommends:
www.bowerswilkins.com
The first thing to bear in mind is tweeter height. It’s important to ensure that the tweeters are as close as possible to your ear level when listening.
When starting to measure this speaker, the first thing I thought was "What have I done wrong?".
But then I looked at the Soundstage anechoic measurement of the larger 685 S2:
www.soundstagenetwork.com
This is my quasi-anechoic result (686 S2) vs Soundstage (685 S2):
The tweeter response is very strange on both of them, and they're actually very similar > 1 kHz (The resonances are at different frequencies due to different woofer sizes, and the larger 685 S2 has more bass).
I then compared the gated response at 1m and 2m in my garage (slightly off-axis. That's why the peaks are a bit different). Basically identical (The difference below 2 kHz is due to lower resolution of the 2m measurement (3ms window):
Just as a final confirmation, since I want to get the measurements right, I brought out my trusty M16. Very similar to my old measurement (living room):
Here's the CTA-2034 data:
Early reflections:
Estimated in-room response:
Horizontal directivity:
Vertical directivity:
Quasi-anechoic response:
Near-field:
Woofer, 6mm (used for quasi-anechoic response) vs 3 cm (used for the near-field image above):
That's the reason for the weird response. Several huge woofer resonances causing tweeter cancellation.
Just to make sure the speaker was ok, I compared it to the second one:
Slightly off-axis here. That's why the peaks are a bit different. Good pair matching:
Both speakers are in excellent condition, and there's no audible distortion. I would still be hesitant to post the measurements though, if it wasn't for the Soundstage anechoic data (685 S2) confirming the problems.
This is part of the problem:
A 6 dB/octave crossover...
Distortion, 86 dB @ 1m:
Distortion, 90 dB @ 1m:
Very low distortion. Bowers & Wilkins is one of very few manufacturers that actually specify the distortion level.
For 686 S2 it's:
2nd and 3rd harmonics (90dB, 1m):
<1% 110Hz - 22kHz
<0.5% 150Hz - 20kHz
The measured performance is very close to matching the specs. The tweeter distortion peaks (still low level) would probably disappear with a real crossover.
I listened to the speaker while writing this, and it doesn't sound terrible, but not good either. No audible distortion or other obvious issues, but it has a very flat (almost non-existent) soundstage, and female voices sound "closed-in" and a bit boring.
You can clearly hear that everything is playing inside a box. The treble is a bit bright and has a bit of "sizzle", but not as much as some newer B&W-speakers.
This speaker is a bit frustrating. It could have been much better than it is, but it seems to be made for the audiophile market.
It was released in 2014, and discontinued in 2018, and it's two generations behind the 607 S2 measured by Amir. I Think the MSRP was 399.99 USD / pair.
My measurements are quasi-anechoic, with near-field port+woofer, corrected for baffle edge diffraction, combined with gated measurements at 1m distance.
I measured at tweeter axis, since that's what B&W recommends:
Blog - Hi-Fi Essentials – How to set up a loudspeaker | Bowers & Wilkins
In the third part of our series explaining the fundamentals of putting together a great Hi-Fi system, we look at how to get the best performance from your loudspeakers.

The first thing to bear in mind is tweeter height. It’s important to ensure that the tweeters are as close as possible to your ear level when listening.
When starting to measure this speaker, the first thing I thought was "What have I done wrong?".
But then I looked at the Soundstage anechoic measurement of the larger 685 S2:
SoundStageNetwork.com | SoundStage.com - NRC Measurements: Bowers & Wilkins 685 S2 Loudspeakers
This is my quasi-anechoic result (686 S2) vs Soundstage (685 S2):
The tweeter response is very strange on both of them, and they're actually very similar > 1 kHz (The resonances are at different frequencies due to different woofer sizes, and the larger 685 S2 has more bass).
I then compared the gated response at 1m and 2m in my garage (slightly off-axis. That's why the peaks are a bit different). Basically identical (The difference below 2 kHz is due to lower resolution of the 2m measurement (3ms window):
Just as a final confirmation, since I want to get the measurements right, I brought out my trusty M16. Very similar to my old measurement (living room):
Here's the CTA-2034 data:
Early reflections:
Estimated in-room response:
Horizontal directivity:
Vertical directivity:
Quasi-anechoic response:
Near-field:
Woofer, 6mm (used for quasi-anechoic response) vs 3 cm (used for the near-field image above):
That's the reason for the weird response. Several huge woofer resonances causing tweeter cancellation.
Just to make sure the speaker was ok, I compared it to the second one:
Slightly off-axis here. That's why the peaks are a bit different. Good pair matching:
Both speakers are in excellent condition, and there's no audible distortion. I would still be hesitant to post the measurements though, if it wasn't for the Soundstage anechoic data (685 S2) confirming the problems.
This is part of the problem:
A 6 dB/octave crossover...
Distortion, 86 dB @ 1m:
Distortion, 90 dB @ 1m:
Very low distortion. Bowers & Wilkins is one of very few manufacturers that actually specify the distortion level.
For 686 S2 it's:
2nd and 3rd harmonics (90dB, 1m):
<1% 110Hz - 22kHz
<0.5% 150Hz - 20kHz
The measured performance is very close to matching the specs. The tweeter distortion peaks (still low level) would probably disappear with a real crossover.
I listened to the speaker while writing this, and it doesn't sound terrible, but not good either. No audible distortion or other obvious issues, but it has a very flat (almost non-existent) soundstage, and female voices sound "closed-in" and a bit boring.
You can clearly hear that everything is playing inside a box. The treble is a bit bright and has a bit of "sizzle", but not as much as some newer B&W-speakers.
This speaker is a bit frustrating. It could have been much better than it is, but it seems to be made for the audiophile market.
Attachments
Last edited: