The B&W N802 lowmid bump

[steffen]

there are one more old threads pointing at this phenomenon and I wonder if this is a bug or a feature? it has annoyed me for a long time and I thought it was more a roomproblem than the speakers (I mean N802s are not considered as junk ) but since the the problem is still present in the new (very very accurate) room I did some measurements to trace it back to the source.
what i found out is that the crossovers frequencies of the bass and midrange drivers overlap way too much. if the filters are linkwitz-riley they should cross at the -6dB point to sum up into a flat overall response but they overlap so much that they genarate a bump of around 6dB at 330Hz wich I think is what others are also complain about.
If I pull out 6dB with a Q of 3 it sounds normal to me but this cant be the fix.

so my questions are:

has somebody ever contacted B&W regarding this and got a response or even
tech support to fix this?

does someone know a qualified person who can help me fix this? (Xover modification)

what about the newer 800 series...800D or 802D ...same "feature"?

steff

[jdg]

some ppl like this "feature"
i do not, and thanks for measuring it
(i've heard it in their other 3ways aswell)

[Cellotron]

It's never bothered me and it's in fact only occasionally that I find myself cutting in the 300-350Hz zone - and it's indeed very rare that I've ever had to do a revision because the client wants thicker lower mids - so obviously the bump hasn't changed my processing decisions in my room. I sit closer to the speakers and and tend to monitor a tiny bit louder than most do though.

As always with monitors the interaction with the room and your own preferences are key to deciding whether you like and can work with them - so obviously OMMV!

Best regards,
Steve Berson

[tpad]

Assuming the N802s are indeed using a Linkwitz-Riley 4th order crossover, that would mean that the overall acoustic response would be 4th order, not necessarily that the crossover network would be 4th order. So, if the bass and midrange drivers are rolling off 6 dB/octave, the crossover slopes would need to be 3rd order, not 4th order. The overlap you're seeing may be what it takes to make the acoustic response behave as intended.

[LV mastering]

Assuming the N802s are indeed using a Linkwitz-Riley 4th order crossover, that would mean that the overall acoustic response would be 4th order, not necessarily that the crossover network would be 4th order. So, if the bass and midrange drivers are rolling off 6 dB/octave, the crossover slopes would need to be 3rd order, not 4th order. The overlap you're seeing may be what it takes to make the acoustic response behave as intended.

Problem is that 3rd order Linkwitz-Riley don't exist.
They must be 2nd or 4th order (or any higher 2*nth order)
I also doubt a loudspeaker could be designed to roll-off so predictably.

[Petrus]

Tweaking the crossover would work, but how about testing the effect with parametric EQ first? A cheap and safe solution would be a passive "black box" on the speaker cables, which would attenuate the offending frequency.

[da goose]

wow.. gonna follow this thread to see what's coming out of this.
When i did measure the room and speakers i also noted a bump in the lower mid range but i thought it was something with my measuring mike because it didn't change with moving mike/speakers and things like that.
I think that i'm already used to it because i never had a client asking for a revision with more midbass but i can imagine that the bump is not a good thing.

So.. just double checking: 6db cut at 330hz with a q of 3 removes the bump? I will try that and measure again to see if the bump is gone then.

I think the best solution will be something like an extra passive filter in series with the midrange. Is the 330hz the x-over point? If so i think moving the x-over point can do the job, maybe by simple using higher value caps?

Also wondering; is it easy to get to the filters? are they on the bottom of the speakers? i also used to have 802series 80's in the past and the filters where there. Never checked it because handling the speakers are not something you do for fun with 80kg each. ;-)

[Audiop]

Originally Posted by LV mastering

Problem is that 3rd order Linkwitz-Riley don't exist.
They must be 2nd or 4th order (or any higher 2*nth order)
I also doubt a loudspeaker could be designed to roll-off so predictably.

He did not say that, but a first order intrinsic roll off of the driver/baffle/box combo paired with a third order electrical slope can be a fourth order LR.

You can steer the acoustic response extremly well with electrical, acosutical and mechanical means.


/Peter

[Audiop]

@steffen

How did you measure these curves, what about the mic positions?


/Peter

[LV mastering]

He did not say that, but a first order intrinsic roll off of the driver/baffle/box combo paired with a third order electrical slope can be a fourth order LR.

Ok so we have a third order "lambda" (I call it "Lambda" as it's specification is unknown) filter and we then have to design a loudspeaker which has such specifications that, when summed with the "lambda" filter, makes a 4th order LR.
uuuwww..possible....but do you think this so-designed speaker will sound good....

You can steer the acoustic response extremly well with electrical, acosutical and mechanical means.

That's right but a 6db (intrinsic, mechanical) roll-off at 300hz seems very difficult to design without impairing the hi-fi quality of the loudspeakers.

[steffen]

speaker design has too much unknown variables for me. driver parameters, enclosure..etc and how it fits into the xover equation....maybe the bump is there for a reason (the result of best compromise...or B&W just liked to get their clients ears bleed when listening to piano music for more than 20minutes)...however, the bump has to go away...no way back

@Petrus did a few tracks with "bump correction" EQ in the monitor path. it´s like day and night....so the eq solution is okay so far, at least temporarily....(damn, I didn´t realize how fatiguing the system was before)
cool idea with the passive box in the speaker cable path, didnt even think this far...wondering about myself since I do custom audio inductors almost every day.

Is the 330hz the x-over point?

I think I read 350Hz somewhere

How did you measure these curves, what about the mic positions?

measuring mic pretty close to the drivers (2-3cm).
the 802 can be bi-amped (bass drivers vs mid driver and tweeter)

for the bass driver measurement I disconnected the mid/hi section and vice versa)

measured with fuzzmeasure

[steffen]

Originally Posted by da goose

Also wondering; is it easy to get to the filters? are they on the bottom of the speakers? i also used to have 802series 80's in the past and the filters where there. Never checked it because handling the speakers are not something you do for fun with 80kg each. ;-)

yes I guess the filters are in bottom.....

btw...does somebody have a schematic of the xover?...cant be complicated but reverse engineering in this case is hard work

[da goose]

Originally Posted by steffen

yes I guess the filters are in bottom.....

btw...does somebody have a schematic of the xover?...cant be complicated but reverse engineering in this case is hard work

I do have the schematic from the 802 series 80 x-overs but i guess that won't be anywhere near the n802's. But i think they wont be that hard to find and i think that b&w is quite easy on these kind of things. Really can't be that complicated indeed, just a bunch of LCR's.
I really wonder what they will sound like without the bump.

[Audiop]

Originally Posted by steffen

measuring mic pretty close to the drivers (2-3cm).
the 802 can be bi-amped (bass drivers vs mid driver and tweeter)

for the bass driver measurement I disconnected the mid/hi section and vice versa)

measured with fuzzmeasure

Hi, to get correct results you need to measure at a distance since a close up will miss the baffle step.


/Peter

[Audiop]

Originally Posted by LV mastering

Ok so we have a third order "lambda" (I call it "Lambda" as it's specification is unknown) filter and we then have to design a loudspeaker which has such specifications that, when summed with the "lambda" filter, makes a 4th order LR.
uuuwww..possible....but do you think this so-designed speaker will sound good....

Most speaker drivers have a non flat FR and the acoustical response is always the sum of the intrinsic response + xover. And yes, such speakers sound good, why would they not?

That's right but a 6db (intrinsic, mechanical) roll-off at 300hz seems very difficult to design without impairing the hi-fi quality of the loudspeakers.

It was just an example but I don't see why such a slope would impair hifi qualities.. Let's say a closed box high pass added to a 12dB/oct electrical slope.


/Peter

[tpad]

LVmastering, The 4th order classic L.R. crossover network would require the associated loudspeaker drivers to be flat from DC to light and have zero phase shift, which they obviously don't do. Consequently, you have to shape the crossover network to complement the actual magnitude and phase response of the driver to achieve the desired response. The crossover network characteristics in real commercial loudspeaker designs typically don't look at all like what you are seeing in idealized text book designs. If you want some good advice, don't take up speaker design as an occupation!

[LV mastering]

@tpad and Audiop

Thanks for your explanations and advice.
I still don't understand why you call the resultant filter (electric+mechanic/acoustic) a Linkwitz-Riley filter. To me it's like summing different sorts of low-pass filters and call it a "moog 24dB filter".
It may come from the process of designing loudspeakers wich is apparently somewhat empirical.

[Audiop]

Hi!

The definition of a filter is the shape/slope, it does not really matter how you get there and it's perfectly fine to mix low level active electronic filters with passive high level filters and acoustical + mechanical filters.

The 4th order LR filter has a specific shape and if your resulting filter has that shape it's a 4th order LR filter. :-)


/Peter

[tpad]

I should also point out that the 4th order L.R. is often claimed to be in phase at the crossover frequency. In fact, it is 360 degrees out-of-phase, since the lowpass filter is lagging the highpass filter by a full cycle. With a continuous sinusoid applied you can't tell the difference, but with anything containing harmonics, it can be pretty obvious. Apply a squarewave, and it sure doesn't come out square!

[Thomas W. Bethe]

Maybe it is the interaction with the room???

I never really noticed any bumps in the 802s response just by listening but I only got to hear them at some hi fi salons and not my mastering room.

A lot of room responses look weird but the speakers sound GREAT. Maybe ruler flat response curves are a) unattainable or b) not what the ear is wants to hear?

I was in a stereo store near here and asked to hear the 802s with some reference materials I had brought along. The speakers were setup with a receiver and not a pre amp/power amp combination. I got the store to change that, then the speakers were wired with 20 gauge speaker wire and I got the store to redo that. The speakers were placed flat against the wall about 20 feet apart and after a couple of minutes listening I determined that they were wired out of phase and way too spread for a good stereo image. We got the polarity fixed and we repositioned them for good listening. we had to reposition them a number of times to get them to sound really good and when they did get set up properly the sound was amazing. The salesmen I was working with told me that all the time he had worked at the store he had never heard these speakers sound this good, Then the "manager/owner" walks in and starts yelling at the salesmen for "moving the speaker without permission". The salesman says "wait just a minute and take a listen to what they sound like now" We played him a couple of tracks I had brought in and some things he went and got and he was very impressed and was smiling and giving an apology to the salesman. Other salesmen came in and the consensus was they sounded GREAT and better than they had ever sounded. As I was leaving the store the manager told the salesmen to "put the speakers back where you found them because the don't look good where they are and someone may trip on the cable if they are that far out in the room. I guess the store was really selling the looks of the speakers and not the sound. FWIW and YMMV

Anyway speaker placement can make a world of difference in how speakers sound. I usually start with this setup http://www.cardas.com/content.php?ar...ing=Room+Setup and tweak from there.

Best of luck!