• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Does anyone else prefer a dipped midrange?

Mart68

Major Contributor
Forum Donor
Joined
Mar 22, 2021
Messages
2,609
Likes
4,860
Location
England
This is all very complex, sometimes I wish my hobbies were simpler. Do you have any simple hobbies, Jim?
The desire for things to be simple is what fuels the snake-oil industry - just but this magic cable/fuse/crystal/box of tricks and deploy it = better sound!

On the other hand the mantra 'If it sounds good to you then it is good' is inescapably true. Doesn't really matter whether we know why it sounds good to us or not.

Some of us still like to know the 'why' though. Then it gets complex.
 

Chromatischism

Major Contributor
Forum Donor
Joined
Jun 5, 2020
Messages
4,765
Likes
3,703
Toole has several times said that stereo is flawed. In the stereo triangle you will experience a dip around 2 kHz followed by a peak 3-4 kHz when using a speaker with flat frequency response. This effect is exaggerated in nearfield (and anechoic) conditions. The dip at 2 kHz can somewhat be filled with reflections from the room and the peak at 3-4 kHz can be damped with 1-2 dB dip.
Interesting. Many, many speakers have a rise at 3-4 kHz, especially so after a crossover dip. Another cause is cabinet edge diffraction, or so I've read. Seems we need speakers that overcome those issues so they don't add on to the effect you're describing.
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,422
Likes
2,407
Location
Sweden
Interesting. Many, many speakers have a rise at 3-4 kHz, especially so after a crossover dip. Another cause is cabinet edge diffraction, or so I've read. Seems we need speakers that overcome those issues so they don't add on to the effect you're describing.
Yes, many speakers have this "problem", i.e. a lower energy at 2 kHz vs 3-4 kHz, on- and/or off-axis. Dips around 2 kHz on or off-axis is not desired IMO, especially related to the 3-4 kHz region.

Some say this is adjusted in the mixing, but the 2 kHz hole can only be filled in using reflections. So near-field mixing and/or mixing in studios with "low-reflection" rooms, will not fix it. Also, if it would be in the mix, all listening tests using a mono speaker would have resulted in a frequency response which is not linear in this region.

So I would say the question is still open. I would however avoid speakers that exaggerate these stereo flaws.
 

sigbergaudio

Major Contributor
Audio Company
Forum Donor
Joined
Aug 21, 2020
Messages
2,639
Likes
5,397
Location
Norway
The overall energy in this frequency band can't be understood by looking at on-axis response alone. There's often interesting things going on off-axis here due to crossovers and the dispersion characteristics of the different drivers involved, so it's typically an area of compromise one way or another.

So you'd need to measure at least 0-60 degrees to get a useful understanding of the best target response.
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,422
Likes
2,407
Location
Sweden
The overall energy in this frequency band can't be understood by looking at on-axis response alone. There's often interesting things going on off-axis here due to crossovers and the dispersion characteristics of the different drivers involved, so it's typically an area of compromise one way or another.

So you'd need to measure at least 0-60 degrees to get a useful understanding of the best target response.
Totally agree.
 

goat76

Major Contributor
Joined
Jul 21, 2021
Messages
1,269
Likes
1,385
Yes, many speakers have this "problem", i.e. a lower energy at 2 kHz vs 3-4 kHz, on- and/or off-axis. Dips around 2 kHz on or off-axis is not desired IMO, especially related to the 3-4 kHz region.

Some say this is adjusted in the mixing, but the 2 kHz hole can only be filled in using reflections. So near-field mixing and/or mixing in studios with "low-reflection" rooms, will not fix it. Also, if it would be in the mix, all listening tests using a mono speaker would have resulted in a frequency response which is not linear in this region.

So I would say the question is still open. I would however avoid speakers that exaggerate these stereo flaws.
I don't think we should exaggerate the problem with that dip at 2 kHz. Will anyone even notice that small dip when listening to the overall tone of a speaker, it's such a narrow band that is affected and will therefore hardly be noticed while listening to music that mostly contains instruments with much broader frequency response? It's not like listening to a frequency sweep where every single minor dip is easily heard. I don't think such small deviations are something you would be listening to while concentrating on the overall balance of a speaker.


1. Toole and his team came to the conclusion that most listeners prefer a speaker with a linear response.

2. If number 1 is true, we can assume that most mixing engineers also prefer that linear response, and have probably set up their sound systems with that in mind (except for the bass response which will probably deviate the most from the norm).

3. When the mixing engineer is making adjustments to the music mix, he will most likely, at least for modern productions, use EQ for every single instrument until they sound exactly like he wants them to sound in the mix. Most stereo flaws will therefore most likely and indirectly be addressed even if the engineer doesn't have any deeper knowledge of those particular stereo flaws, he just mixes the music until it sounds like he wants it to sound, and calls it a day.

4. That leads us to the full circle. If we end consumers have sound systems with a linear response, most recordings should sound fairly well balanced to most people. But there's nothing wrong if some of us have other preferences that don't go hand in hand in with "the middle of the road" as it seems to be with the OP of this thread. I prefer more bass than Harman's "trained Listeners" target curve, and opposite to the OP, I prefer a slightly forward-sounding midrange instead of a laid-back sound. To me, most music sounds more engaging and enveloping that way. :)
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,422
Likes
2,407
Location
Sweden
I don't think we should exaggerate the problem with that dip at 2 kHz. Will anyone even notice that small dip when listening to the overall tone of a speaker, it's such a narrow band that is affected and will therefore hardly be noticed while listening to music that mostly contains instruments with much broader frequency response? It's not like listening to a frequency sweep where every single minor dip is easily heard. I don't think such small deviations are something you would be listening to while concentrating on the overall balance of a speaker.


1. Toole and his team came to the conclusion that most listeners prefer a speaker with a linear response.

2. If number 1 is true, we can assume that most mixing engineers also prefer that linear response, and have probably set up their sound systems with that in mind (except for the bass response which will probably deviate the most from the norm).

3. When the mixing engineer is making adjustments to the music mix, he will most likely, at least for modern productions, use EQ for every single instrument until they sound exactly like he wants them to sound in the mix. Most stereo flaws will therefore most likely and indirectly be addressed even if the engineer doesn't have any deeper knowledge of those particular stereo flaws, he just mixes the music until it sounds like he wants it to sound, and calls it a day.

4. That leads us to the full circle. If we end consumers have sound systems with a linear response, most recordings should sound fairly well balanced to most people. But there's nothing wrong if some of us have other preferences that don't go hand in hand in with "the middle of the road" as it seems to be with the OP of this thread. I prefer more bass than Harman's "trained Listeners" target curve, and opposite to the OP, I prefer a slightly forward-sounding midrange instead of a laid-back sound. To me, most music sounds more engaging and enveloping that way. :)
Well even if my example is a bit exaggerated to show the characters of dips and peaks in the region the question is still open. There are, to my knowlegde, no published studies that investigate speaker preference in stereo specifically adressing these questions, and I doubt there ever will be.
 
Last edited:

Tangband

Major Contributor
Joined
Sep 3, 2019
Messages
2,994
Likes
2,789
Location
Sweden
You might be interested in this then:

This monitor seems to have 4-5 dB to high tweeter response at 4 kHz and upwards. Fortunately you can fix this with the internal switches inside the speaker .

No wonder Digby wanted to decrease the upper mid area , when it was already 4 dB to loud.
 
Last edited:

Tangband

Major Contributor
Joined
Sep 3, 2019
Messages
2,994
Likes
2,789
Location
Sweden
Well even if my example is a bit exaggerated to show the characters of dips and peaks in the region the question is still open. There are, to my knowlegde, no published studies that investigate speaker preference in stereo specifically adressing these questions, and I doubt there ever will be.
My own conclusion after investigating this with my HYBRID dsp DIY loudspeaker and Genelec 8340 is that if the loudspeaker has good directivity, a flat frequency response are gonna be prefered , at least by me and some friends . Even in a stereo setup.
This is also what Dr Toole teaches that most people prefer .

This also means that a loudspeaker with less good directivity at the crossover region , often at 3 KHz , gonna need some compensation with less energy in the direct sound because at 30 degree angle there might be a peak in the crossover region if no waveguide is used.

A very good loudspeaker such as pi60 with a big 8 inch woofer and a tweeter with a very small waveguide gonna have less good directivity , thus need those small stereo system corrections in the frequency response so the sound will , overall , be balanced.

This is not needed in a Kef Ls50 Meta or Genelec 8351 , because the directivity is optimal .

John Atkinsson at Stereophile has shown us many examples of bad directivity loudspeakers that measures less than flat on axis, because of a compensation of the sound off axis .

Further , as goat76 has written earlier: - 9 out of 10 recordings are made and shaped in the studio , listening in stereo , where every instrument is shaped to sound good with eq in this stereo setup.

The recordings are already in this way stereo compensated .
 
Last edited:

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,424
Likes
4,030
Location
Pacific Northwest
After testing a few different speakers, I have come to the conclusion that I don't like a completely flat midrange, I find it to sound unnatural and somewhat boosted to my ears. I am trying to compile a list of reasons why this might be, that aren't related to personal preference, here are some I have come up with:

1. A lot of, particularly older, classical music is recorded with a midrange boost, because of microphone placement (according to Linkwitz), so a flat speaker sounds boosted
2. Listening in an untreated room. Perhaps recordings mastered in a heavily treated studio will sound boosted in the mids when transferred to a typical, untreated room?
3. Listening in the nearfield. It just seems much more relaxing/natural to have a dip when the speakers are within 1m of you. The same dip was appreciated at distances of 3m or more, but a somewhat shallower dip may be preferred at these distances.
I believe there is some truth to each of these points. And some other good points raised during this discussion. A lot of classical is close miced which emphasizes mids & treble. That is what it sounds like on stage with the musicians but not what it sounds like to the audience. Yet many people prefer this hyper-detailed sound having more "zing".

In my treated listening room, the tube traps have reflective material around half their circumferance. I always had this facing out into the room for dispersion. A few months ago I rotated them 180* so the reflective sides are against the walls. This made an audible difference in the sound that I find to be an improvement. It took a slight edge or emphasis off the mids & treble for a smoother more natural sound.
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,422
Likes
2,407
Location
Sweden
My own conclusion after investigating this with my HYBRID dsp DIY loudspeaker
is that if the loudspeaker has good directivity, a flat frequency response are gonna be prefered , at least by me and some friends . Even in a stereo setup.
This is also what Dr Toole teaches that most people prefer .

This also means that a loudspeaker with less good directivity at the crossover region , often at 3 KHz , gonna need some compensation with less energy in the direct sound because at 30 degree angle there might be a peak in the crossover region if no waveguide is used.

A very good loudspeaker such as pi60 with a big 8 inch woofer and a tweeter with a very small waveguide gonna have less good directivity , thus need those small stereo system corrections in the frequency response so the sound will , overall , be balanced.

This is not needed in a Kef Ls50 Meta or Genelec 8351 , because the directivity is optimal .

Further , as goat76 has written earlier , 9 out of 10 recordings are made and shaped in the studio , listening in stereo , where every instrument is shaped to sound good with eq in this stereo setup.

The recordings are already in this way stereo compensated .
Could be. But again remember that the preferred loudspeaker has a linear response when evaluated in mono. Puttng the same speaker in stereo causes a timbral shift for the central phantom source and which can be fixed by adding a center speaker. In stereo you are in a dilemma and you can only partly do something. Also remember that there are quite a number of speakers that show dips on and/or off-axis around 2 kHz.
 

ryanosaur

Major Contributor
Joined
Mar 17, 2022
Messages
1,501
Likes
2,375
Location
Cali
Just weighing in on the original question, my preference is flat FR on the Speaker, and ideally flat in room.
I did experiment once by introducing a mild version of the BBC dip, shallow (only -1.5dB) and wider... It did take some "glare" out of some recordings, but also still made certain passages a bit lackluster.
I listened on and off with this setting for a month or so, ultimately ditching it for maximally flat.
 

Tangband

Major Contributor
Joined
Sep 3, 2019
Messages
2,994
Likes
2,789
Location
Sweden
Just weighing in on the original question, my preference is flat FR on the Speaker, and ideally flat in room.
I did experiment once by introducing a mild version of the BBC dip, shallow (only -1.5dB) and wider... It did take some "glare" out of some recordings, but also still made certain passages a bit lackluster.
I listened on and off with this setting for a month or so, ultimately ditching it for maximally flat.
I have done about the same , with small dsp corrections ( - 1,5 dB ) and after long time listening come to the same conclusion as you have .
 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,424
Likes
4,030
Location
Pacific Northwest
I did experiment once by introducing a mild version of the BBC dip, shallow (only -1.5dB) and wider... It did take some "glare" out of some recordings, but also still made certain passages a bit lackluster.
I listened on and off with this setting for a month or so, ultimately ditching it for maximally flat.
I have done about the same , with small dsp corrections ( - 1,5 dB ) and after long time listening come to the same conclusion as you have .
Ditto here. I had no idea this was such a common thing!
Rotating my tube traps to reduce reflection perceptually provided a similar smoothing as the EQ, without the loss of detail. More linear & natural.
 

goat76

Major Contributor
Joined
Jul 21, 2021
Messages
1,269
Likes
1,385
Ditto here. I had no idea this was such a common thing!
Rotating my tube traps to reduce reflection perceptually provided a similar smoothing as the EQ, without the loss of detail. More linear & natural.
You practically made your tube traps into normal absorbers, I think. Maybe you should think of getting normal flat panels instead, which will more easily cover a larger area of reflective surface?
 
Last edited:
OP
D

Digby

Major Contributor
Joined
Mar 12, 2021
Messages
1,632
Likes
1,555
My own conclusion after investigating this with my HYBRID dsp DIY loudspeaker and Genelec 8340 is that if the loudspeaker has good directivity, a flat frequency response are gonna be prefered , at least by me and some friends . Even in a stereo setup.
I think the problem is you (maybe ASR generally?....shots fired, BTW :p) are at risk of making a god out of the source material, when the source material is itself a gross distortion of live sound. If you look at points 1 & 2 from my OP, they are objective facts, not my preference. Microphones are unnaturally placed and receiving signals in a way different from our ears, and studios are 9 times out of 10, much more soundproofed than a home living room, so what sounds right in a studio may not in a typical home. These are just two distortions I can bring to mind, likely there are others.

TL;DR If a Stradivarius or Guaneri sounded the way it is reproduced by anechoically flat speakers, then you'd have people running from the concert hall. This may be no fault of the speakers themselves, but...(and also in response to this post by abdo123)
It's just that these issues should be sorted out in the production stage, you can't expect the end user to play DJ in their free time.
we have to deal with things as they are, not as they would be in a perfect world.
 
Last edited:

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,422
Likes
2,407
Location
Sweden
Experiment with a center and good downmix to three-channel audio.
 

goat76

Major Contributor
Joined
Jul 21, 2021
Messages
1,269
Likes
1,385
The tube traps geometry enable them to affect much lower frequencies than flat panels would.
That depends on the thickness of the panel and the distance between the panel and the wall. I think you can get an even better result with a panel.
 
Top Bottom