Thanks, Jim.
I had it from 120 to 2000 Hz in a test platform I've used with similar drivers that haven't produced the characteristic I've been picking up. Harshness might not be the right term--more a shrillness, especially attacks in female vocals. It could very well be that the problem is in the source material or signal path and that the Textreme is revealing it while other drivers cover it up. I do tend to favour smoother, less detailed drivers for mid bass-midrange.
Try lowering output between 600Hz to 1kHz (especially between 600-800Hz region) by 0.5 to 1 dB, play around that region. Personally I find the female vocal can be a bit "shouty" if those region is high. I used the MW16TX in a 2 way setup and crossover to the TW29TXN-B at 2.3kHz.
One thing that I noticed - with any driver - is that you really have to be careful with how you filter it and how you EQ it. As JamesTan point out, you have to make sure that there ain't' something "sticking" out somewhere.
Especially a larger bump or dip, of only 1 dB over a broad frequency range - can really be audible.
How do you measure - equipment, technique and setup?
Another aspect to consider: If the response is rising from 1k to 3k, many people will perceive a harshness or shrillness. It only has to be 1 to 2 dB rise, and not necessarily on-axis although that would be worst case. If the ER curve, the PIR curve, or the power response curve have a rising response from 1k to 3k, this can create the sensation of harshness that has nothing to do with the driver, but is entirely related to the radiation pattern.
This is the kind of response I am talking about. Note that the on-axis and listening window curves are flat +/- 1 dB. However, the Power response, ER, and PIR curves all have a rise from 1.5k to about 3k.
j.
This is the kind of response I am talking about. Note that the on-axis and listening window curves are flat +/- 1 dB. However, the Power response, ER, and PIR curves all have a rise from 1.5k to about 3k.
j.
You really think that is enough to cause a noticeable shrillness in music? How about the rise from 5k to 20k? That seems like a much sharper rise but you are saying the rise between 1k and 3k where it's at? I've have built speakers in the past and due to the cross over points and Butterworth filters there where bumps in the frequency response there and never had anything I would consider shrill before. But then again I've never really plotted the power response before.
When I had mine in the test box (I haven't finished my floor standers yet.
) they didn't have any shrillness to them but one thing I did notice was they where very revealing. If there where any flaws in the recording you where going to hear them. To some people this might come across as shrillness absent a better description, I don't know. I did have a good amount of roll off for the higher registers though but nothing that the midrange would of heard.
Yes, thats why people use a waveguide, to even out the off axis to make it sound more relaxed.
Some people think so, including people whose opinion matters. I don't seem particularly sensitive to this kind of radiation signature, and it may be to the type of music I prefer when I do critical listening. But many people are very sensitive to this phenomenon.
As @Jonasz mentions, a waveguide can control the directivity of a tweeter so that it blends well with the mid driver over the range from 1k to 4k, which is where a crossover is typically located. This allows the designer to avoid the 1k-3k rise in ER, PIR, and power. However, it can be done without a waveguide.
j.
Okay the wave guide on the tweeter I can understand. I have used them in the past to direct the high frequency energy in a built in cabinet/ speaker box that was above my head and near field, I know not ideal by any means, but it had nothing to do with harshness. I guess I'm having a hard time understanding the graph you showed. There is the rise in the power response and a corresponding dip in the off access frequency response so to me the rise in the power response is not coming from the off access frequency response of the tweeter but more likely coming from the room? So a wave guide will coral the off access energy and make for a more controlled off access response but how does that cause or remedy this perceived "shrillness"?
To me it would make more sense to increase the slopes on the cross overs. I don't know where they are set at now but if this shrillness is only at high volume I would think it is punching through and higher order x-overs are probably needed. That would be my first guess. If that doesn't work I would start with very narrow bands on the EQ and start pulling them down one by one to see if there is a corresponding frequency that is causing the shrillness and remove it with the EQ. That is how I would go about diagnosing the problem anyways. After that I would start looking at the drivers themselves. Are they both shrill or is it just one of them? It could very well be that he got a bad driver or maybe he is just driving them to hard. Another thing I would try is to take the speakers outside and see if the shrillness is still there. These are the first things I would look at. But that's just me.
@iamjackalope – our discussion might be going off the rails here just a bit. I am not saying that the radiation pattern I noted will always lead to harshness. I am merely pointing out that if harshness is perceived, and the speaker has the response trend I noted, this could be the explanation for the perception of harshness.
@svans asked if anyone had noticed a harshness in the MW16TX driver. I responded that I had not noticed any harshness. I then said that harshness at high SPL can be caused by many things, and it is not necessarily a driver issue. I did not offer any additional explanation of what those things might be. Later, I decided I should elaborate more on potential sources of harshness.
I became aware of this phenomenon during the most recent DSP filter development of my project
https://www.diyaudio.com/community/threads/new-active-satori-textreme.366347/page-10#post-7167183
post #181
Apparently, some people are more sensitive than I am to a rise in the 1k-3k region. I was able to mitigate the response with some careful DSP filter adjustments. For me, the difference in sound quality is small, but others might perceive a larger difference, and I think it might be program dependent. It is probably room dependent as well.
When I said "Some people think so, including people whose opinion matters" ... I was specifically thinking of @fluid and Kimmo. Both have discussed this subject, and that is how I became aware of it.
So I am not promoting a particular radiation pattern as superior. I am just offering some thoughts to svans on what he might be hearing in terms of "harshness"...
j.
@svans asked if anyone had noticed a harshness in the MW16TX driver. I responded that I had not noticed any harshness. I then said that harshness at high SPL can be caused by many things, and it is not necessarily a driver issue. I did not offer any additional explanation of what those things might be. Later, I decided I should elaborate more on potential sources of harshness.
I became aware of this phenomenon during the most recent DSP filter development of my project
https://www.diyaudio.com/community/threads/new-active-satori-textreme.366347/page-10#post-7167183
post #181
Apparently, some people are more sensitive than I am to a rise in the 1k-3k region. I was able to mitigate the response with some careful DSP filter adjustments. For me, the difference in sound quality is small, but others might perceive a larger difference, and I think it might be program dependent. It is probably room dependent as well.
When I said "Some people think so, including people whose opinion matters" ... I was specifically thinking of @fluid and Kimmo. Both have discussed this subject, and that is how I became aware of it.
So I am not promoting a particular radiation pattern as superior. I am just offering some thoughts to svans on what he might be hearing in terms of "harshness"...
j.
Yeah that's the problem with talking about sound. Like talking about wine or food. What is "shrillness" exactly. We need to define it better. Or do we even know?
according to wikipedia:
"Shrillness is a word used to describe the quality of sounds that have a high-pitched, strident, raucous, screeching or harsh character, such as those produced by a trumpet or piccolo, but it can also be used to describe a widely recognised and puzzling phenomenon whereby certain sounds are perceived as psychologically painful or aversive to a degree that cannot be accounted for simply in terms of frequency content or loudness. Such sounds include the sound of fingernails scraping a chalkboard, the sound of chalk on a blackboard, the sound of glass being scratched, and possibly the sound of a baby crying. There have been attempts to explain the phenomenon, often in terms of frequency content, or evolutionary advantage, but so far no complete explanation or mechanism has been found.
A 2011 study[1] by musicologists Michael Oehler and Christoph Reuter has led its authors to hypothesize that the unpleasantness of the sound is caused by acoustic resonance, as the shape of the human ear canal amplifies certain frequencies, especially those in the range of 2000 to 4000 Hz (the median pitches), at such a level that the sound triggers pain in our ears.[
Reference:
https://en.wikipedia.org/wiki/Shrillness
My interpretation of what Jim is saying is that we don't really know yet what's causing the shrillness. Before we blame the driver, perhaps more needs to be done to further define the problem. Take further measurements- not just on axis. On needs the listening window, all the horizontal off axis, all the vertical off axis, the averaged power response, and PIR, as well as the actual in-room response at a single mic point or using the moving mic method. Look at the amplitude response, look at the harmonics, the IMD. Or is there a non-static component to it eg. vibrato element.
And adjust by EQ or crossover to see when the "shrillness" goes away.
And then we can have something to talk about.
Have a look at this frequency chart:
https://alexiy.nl/eq/
I once complained about my 5 yo daughter's screaming temper tantrums bothered me so much. I whipped out an RTA and measured it- it's that piercing 4KHZ that is like an ice-pick to my ears.
It just doesn't seem to bother my spouse as much. She can ignore it. But she can't seem to ignore my deep voice yelling at my boy...
according to wikipedia:
"Shrillness is a word used to describe the quality of sounds that have a high-pitched, strident, raucous, screeching or harsh character, such as those produced by a trumpet or piccolo, but it can also be used to describe a widely recognised and puzzling phenomenon whereby certain sounds are perceived as psychologically painful or aversive to a degree that cannot be accounted for simply in terms of frequency content or loudness. Such sounds include the sound of fingernails scraping a chalkboard, the sound of chalk on a blackboard, the sound of glass being scratched, and possibly the sound of a baby crying. There have been attempts to explain the phenomenon, often in terms of frequency content, or evolutionary advantage, but so far no complete explanation or mechanism has been found.
A 2011 study[1] by musicologists Michael Oehler and Christoph Reuter has led its authors to hypothesize that the unpleasantness of the sound is caused by acoustic resonance, as the shape of the human ear canal amplifies certain frequencies, especially those in the range of 2000 to 4000 Hz (the median pitches), at such a level that the sound triggers pain in our ears.[
Reference:
https://en.wikipedia.org/wiki/Shrillness
My interpretation of what Jim is saying is that we don't really know yet what's causing the shrillness. Before we blame the driver, perhaps more needs to be done to further define the problem. Take further measurements- not just on axis. On needs the listening window, all the horizontal off axis, all the vertical off axis, the averaged power response, and PIR, as well as the actual in-room response at a single mic point or using the moving mic method. Look at the amplitude response, look at the harmonics, the IMD. Or is there a non-static component to it eg. vibrato element.
And adjust by EQ or crossover to see when the "shrillness" goes away.
And then we can have something to talk about.
Have a look at this frequency chart:
https://alexiy.nl/eq/
I once complained about my 5 yo daughter's screaming temper tantrums bothered me so much. I whipped out an RTA and measured it- it's that piercing 4KHZ that is like an ice-pick to my ears.
It just doesn't seem to bother my spouse as much. She can ignore it. But she can't seem to ignore my deep voice yelling at my boy...
Last edited:
I think we have a winner. This a sound that I think we can all agree on. That little girl screech. There is always one of these little ear demons😱 on every playground. They scream at everything. They see a lizard and it's time to pierce the ear drums. If this is the sound of the drivers at volume then there is definitely a serious problem.😮
Without knowing first hand what this shrillness sounds like it's kind of hard to diagnose and that's why I listed how I would go about dealing with it. Because it is at volume my first suspect would be the x-overs and I would think they need to be a higher order but it could be a rouge frequency that is causing it and in that case start pulling things down in the EQ and see if it goes away. You can also take the speakers outside to rule out the room. After all of that and you can't find the culprit you can always do the full gambit of measurements and everything that goes with it but sometimes you just want it to sound good and not have to go through all of that.
It's all good man. I wasn't suggesting anything I was just trying to understand what you where saying because it wasn't clicking in my mind and I was trying to explain why it wasn't clicking. No worries it's all academic. I actually respect what you have to say on these matters as you have been delving much deeper into theory then I have ever bothered to go. I think it's that engineering background of yours. I'm more the guy that knows how to make it sound right but can't always explain it. This has especially been the case with car audio where I have a lot more experience. As a matter of fact I'm in the process of putting a system together in my Tundra right now. It's shaping up to be a real nice install. I have a build thread going over in the car audio forum.
Speaking of waveguides I have used them in car audio installs to direct sound off access and up to the listeners ears. In this case you are trying to get more off access energy not less. When I first started using waveguides I was amazed at how much of a tweeters energy can be deflected with a simple wire the size of an 8 penny nail.
Do other speakers in that room also show "shrillness" ? Because we are very senstitive between 2-4 kHz and this is also range where modern-empty-refletive rooms have problem with flutter echo...Female voices, violine, trumpet and right buttons of piano can sound painfull at higher volumes....already a small absorption - diffusion treatment on right spots will reduce this a lot...
And yes, increased power response from typical two way in this area is just making it worse....
And yes, increased power response from typical two way in this area is just making it worse....
And that's the beauty of waveguides and 4-5" midranges. The WG helps the tweeter to reach lower while also getting a power response matching the midrange. And the smaller midrange can reach a little higher than the 6-7" versions, while still keeping a nice and even response at all angles, helping to have that cleaner transition in the cross-over region
There is definitely a tradeoff and a balancing of performance priorities. Certainly, any tweeter benefits from a higher filter crossover than a lower one. But the flip side of that coin is the performance of the midrange driver going ever higher.
I have found that a tweeter can be crossed based on the following criteria:
(1) Based on Fs and filter level: For 2nd order filters 4 x Fs, for 3rd order filters 3 x Fs, for 4th order filters 2 x Fs. The idea is that the electrical drive signal should be about -20 dB down at Fs.
(2) Based on the point where the 2nd Harmonic Distortion reaches -40 dB (1%)
(3) Based on the point where 3rd Harmonic Distortion level reaches -50 dB (0.3%)
The minimum crossover is the highest of any of the above criteria.
This criteria has worked for me for several projects. If the tweeter can produce 1.7k with sufficient SPL and low distortion, I have found that I almost always subjectively prefer the tweeter to handle it rather than a cone midrange driver.
I don't expect everyone to agree with this. We all differ in how we hear things, and what sonic tradeoffs we are willing to make.
I did not mean to imply that the MW16TX should not be used above 1.5k. In my application, when combined with the TW29TXN-B tweeter which can play down to 1.6k with low distortion, I believe that the best crossover is 1.6k 4th order. I have also used 1.8k 3rd order. But with a different tweeter, a 2k or 2.5k might be best.
Here is the measured polar response of the MW16TX-4 in my cabinet, with no filtering applied except a +6 dB baffle step compensation. We can see that there is a cone resonance at 5k. This is something that any crossover will have to deal with, whether active or passive. I found a notch filter works very well. The higher the crossover, the more difficult this peak becomes, but a 2.5k crossover is quite practical.
There is also a bit of misbehavior at 2.7k-3k, and after looking at the spectral decay plots, I don't see much stored energy at this frequency, so I think it is the rubber surround moving in antiphase. The off-axis cancellation null should also happen at about 3k, so this is also what we are seeing here. This small burble at 2.7k - 3k does not need a notch filter, in my opinion.
