Harshness in Midrange??? Any Help

I agree with the initial response from Driver. Play around with your toe-in and positioning of speakers in relation to your listening positioning. Seeming small changes in toe-in/distance can make a surprising difference.

Thanks to all who have offered responses on this thread. I have acquired some great knowledge that I will use going forward. Keep it coming and thanks again.

Rob

Room acoustics, and system setup. Have you measured the frequency response of your room? If not, that's a good first step. If you can see on a chart that you have a 10db peak at 1khz(room acoustics could easily create this), it's going to make it a lot easier to diagnose and fix the problem.

In looking at the Stereophile measurements of the 802D, I see a couple of characteristics that could possibly contribute to "harshness" on female vocals.

http://www.stereophile.com/floorloudspeakers/1205bw/index4.html

As you can see, there is a local peak at 4 kHz. Too much energy at 4 kHz sounds harsh and edgy, and 4 kHz is right in the region where the ear is most sensitive. Here is a family of equal-loudness curves - where each curve dips the lowest is where our hearing is the most sensitive:

http://hyperphysics.phy-astr.gsu.edu/hbase/sound/eqloud.html

It gets worse. The crossover point is at 4 kHz, and in this region we transition from the 6" stiff-cone midrange (which will be beaming) to the 1" tweeter will have a very wide radiation pattern and so will be putting out a lot of excess energy into the reverberant field. Since the ear derives timbre from not only the first-arrival sound but also the reverberant energy, the 4 kHz range will be even further emphasized relative to the rest of the spectrum the farther back you listen and/or more reverberant your room.

Not done yet. Most speakers have the tweeter mounted on a baffle, which effectively limits the tweeter's radiation pattern to the front hemisphere (think of the baffle as a 180 degree horn in that frequency region). The B&W tweeter is mounted without a baffle, so its radiation pattern is close to omnidirectional at 4 kHz. The result is, relatively speaking, even more surplus reverberant energy in this region in comparison to the rest of the spectrum.

They probably image great, though.

Duke
dealer/manufacturer

Duke,

Thanks - I really learned a lot from your comments above - much appreciated.

I always wondered how those B&W three ways could get away with a 6 inch midrange crossed over so high ( I thought that phase plug looked too small to be effective) and your analysis suggests they don't completely get away with it ===>it involves some compromises. However, not having a nasty crossover right slap bang in the midrange is a HUGE plus - so these 802D's rightly deserve the great respect they have garnered - to me it is a matter of opinion as to which is the lesser of two evils (split the mid in two pieces or lose some upper mid energy).

The top mounted tweeter effect is so clearly visible on the dispersion plot at 4 to 5 Khz at extreme wide angles that it had light bulbs going off in my head when you so kindly explained it. I agree the contrast at 4 Khz between mid and this particular tweeter is a bit severe.

It might be interesting to combine your analysis on this thread with comments from Jkalman on his experience with the 802D (on another thread where he felt the mid range lacked clarity/intelligibility in a well treated room by Rives - Jkalman uses Wilson's now - but I would be interested in Jkalman's opinion as to whether the acoustic treatment cured the harshness but at the expense of suppressing the mids too much?)

Your comments show the value of interpreting measurements and the importance of smooth and even dispersion in the off axis response. Since our ears hear a combined sound from on axis and off axis room energy your analysis suggests it is equally important how a speaker behaves off axis as on axis.

Nevertheless, 802D is a classic great speaker; one of the very best - if it is good enough for Alan Parsons it clearly kicks butt. To be fair, I can easily find other examples of high-end speakers with large 5 or 6 inch midrange drivers that are often far worse then this (where the upper mid range off axis response does not merely roll off but falls off a cliff.) I can post these if anyone is interested but I prefer not too in case this data upsets the owners.

I suspect a PEQ -3 db to -6 db notch filter with Q =4.32 or 1/3 octave centered at 4.5KHZ might cure Taylor's harshness problems altogether ....would you agree?