Full size 3-way project

like a specific area in the midrange is 'zoomed in' with a lot of detail
Have you tried decreasing this area with DSP/EQ? (from post 182)

1714315674017.png


What you are describing sounds like a typical response to a midrange area that's a little too hot. I would try to get to the bottom of it with EQ before abandoning the driver you have. I had similar issues with the Scan-speak H2606/9200.

Something like this often sounds good initially because it gives an impression of extra detail. But on the wrong song it gives a harsh edge to the sound. For me that area is fussy when it comes to voicing/tuning/whatever you want to call it. What your ear latches onto as a reference often depends on what the frequency response around the area is like and the particular song. Your midrange (400 Hz -1k) seems a little elevated, which may make the upper midrange sound more in balance when it's also elevated. The depression around 200-300 Hz in-room may also be affecting this. Slightly different speaker locations may help that, or you could go after the bass balance with DSP as well (cutting 80-150 Hz). Often fixing one thing leads to adjustments in other areas.

Sometimes listening to pink noise can help highlight problems like this, especially if you have a known nicely balanced speaker or set of headphones to instantly switch back and forth to.
 
  • Like
Reactions: 2 users
I tried moving the speakers a little (not much room to do that), I also tried EQ:ing them more flat from the listening position, but that also sounded wrong.
They measure nice nearfield, but I can still hear the same thing when listening closer to the speakers.. Since I had similar issues with the DXT tweeter, I started thinking I should try without a waveguide.
 
Room response "hotness" around 3kHz is typical consequenze of tweeter's wider radiation pattern. 2-3dB notch in on-axis at 1-2m should tame it remarkably. Ears are very sensitive around there, even 1 dB can be heard. 2-4 or 3-6 kHz means one octave! Many instruments have lots of harmonic distortion there (their sound signature)
 
Last edited:
  • Like
Reactions: 1 user
I would also continue with some applied EQ. Try to get the voicing right - it always needs some listening to the finished speaker and often some fine adjustments.
The DXT has a way to small wavegudie and pretty sure also pronaunces this area in the room, esp when used with a bigger midrange.

T34A would be a tweeter which can cope with an 8" midrange and has very wide radiation.
 
Hi,

I have had similar kind of problem at about 2.5-3 kHz which was very close to cross over frequency. Closer examination revealed that while on-axis response was more or less flat, at some angles off-axis(top/side) there were peak about 2-3 dB in magnitude. Modeling software did not show this problem because it did not take into account combined directions, like both at top & left/right. So cross over was changed a little bit and now there is a sligt dip on axis @ 2.7 kHz, no peak off axis and the sound is pleasant to listen.

Regards
 
I would also continue with some applied EQ. Try to get the voicing right - it always needs some listening to the finished speaker and often some fine adjustments.
The DXT has a way to small wavegudie and pretty sure also pronaunces this area in the room, esp when used with a bigger midrange.

T34A would be a tweeter which can cope with an 8" midrange and has very wide radiation.

I will try again.. :)

The DXT was another project with a smaller mid, 5,25" if I remember correctly, on a large curved baffle, maybe it was in the background on some pics here before, not sure. FR looked nice off axis too, but still did not like it.
 
FR looked nice off axis too, but still did not like it.
I don't like the sound of DXT driver as it is. It's a pretty old Seas tweeter, I always found them intrusive. Listened to a few of them and owned one speaker with these tweeters.
But I really like the sound of ring radiators, just don't like the beaming >10kHz. A waveguide can balance this a little, so that should not be the problem here.