Super Tweeters Suck... Here's Why

Comb filtering happens when you have two drivers (speakers) playing the same signal and the signal from one of those drivers reaches your ear before the other one. Depending on the delay time, some frequencies are reinforced while others will cancel out, making the frequency response look like a comb with teeth (peaks and dips). This is a very audible (and proven to be audible) effect.

So while the super tweeter can give a substantial boost to the high end response, it can also degrade the midrange response. And our hearing is most sensitive to these midrange frequencies.

Not shown in this video is how much worse the response can become when listening off axis, especially vertically. As you move above or below the speaker's tweeter, the delay increases, causing more of that comb filtering effect.

Getting the super tweeter as physically close to the speaker's tweeter as possible and using a higher order filter (crossover) will help to mitigate these unwanted effects. But a better "fix" for a perceived lack of high frequency detail is to either crank up the treble knob (if your setup has one) or get yourself fitted with hearing aids.
Believe it or not, hearing aids will make a massive difference if your hearing has degraded due to age or damage. And it's a difference that you can't match by boosting the high frequency response of your speakers or by adding a super tweeter.

You can help support the work I do in making these videos:
Support this channel on Patreon:

#diyspeakers
#johnheisz
#audio

My "Scrap bin" channel:

My main channel:

[IBuildIt]

Comb filtering happens when you have two drivers (speakers) playing the same signal and the signal from one of those drivers reaches your ear before the other one. Depending on the delay time, some frequencies are reinforced while others will cancel out, making the frequency response look like a comb with teeth (peaks and dips). This is a very audible (and proven to be audible) effect.

So while the super tweeter can give a substantial boost to the high end response, it can also degrade the midrange response. And our hearing is most sensitive to these midrange frequencies.

Not shown in this video is how much worse the response can become when listening off axis, especially vertically. As you move above or below the speaker's tweeter, the delay increases, causing more of that comb filtering effect.

Getting the super tweeter as physically close to the speaker's tweeter as possible and using a higher order filter (crossover) will help to mitigate these unwanted effects. But a better "fix" for a perceived lack of high frequency detail is to either crank up the treble knob (if your setup has one) or get yourself fitted with hearing aids.
Believe it or not, hearing aids will make a massive difference if your hearing has degraded due to age or damage. And it's a difference that you can't match by boosting the high frequency response of your speakers or by adding a super tweeter.

[CDN_Torsten]

The problem with having 2 tweeters/speaker is that they need to be placed within one wavelength of each other to be additive. This becomes problematic at higher frequencies as the wavelength may become quite short (less than 2cm at 20kHz for example). The second potential problem will be that the tweeters will become much more directional (which may or may not be a desired feature). For example stacking them vertically will narrow the sound image in the vertical direction while keeping the horizontal image a wide as it was with only a single driver.

[amdenis]

Well explained, and fun to watch and follow process. Two little FYI things. First, that speaker is a standard dome, not a horn speaker. If that was really a horn, virtually all speakers would be a horn. Second, super tweeters are generally 20kHz and above, and serve a different, but related purpose.

[yurodivy]

A more accurate title would be "Here's Why My Crossovers For Super Tweeters Suck". Are people really trying to have this much overlap with their stock tweeters? I am in the process of adding dipole super tweeters to my Magnepan LRSs and they will have a third order crossover at 10k, maybe a little higher so the overlap with the Magnepans will be minimal.

[robertgates]

In my observations the comb filtering characteristics change just as much by moving the microphone a few inches as it does the supertweeter. Moving back from 1 meter to 3 meters there are there are a dozen or so node changes, each at slightly different but distinct frequencies. But in the end, the negative effects can go away when listening to complex waveforms (music). With the right crossover and Lpads on both drivers, you can really add a nice new dimension to your speakers. I use a second order xover at 12KHz with a JBL ST450 supertweeter.

[vicf]

I really enjoy the material and information that you have on this channel.

[mkshffr]

On my next build I am considering a single tweeter firing up into a bi-directional reflector.

[Poppinwheeeeellllllieeeeez]

Turn the tweeter backwards or straight up or angled 45 degrees towards the rear. Or position the speaker back from the baffle in 1/2 " increments until the correct phase is achieved. That is the fun of the quest of the best sound. Baffle-less speakers are the next step. Use a bread board to build an external crossover jjg and test caps inductors and crossover variation. Lots of cool tests to run.

[jdlech]

what I most commonly see with supertweeters is that people plop it on top of a box with a simple 1st order filter and little to no effort to phase match the tweeter.
A super tweeter should not be used unless the tweeter rolls off at the high end. Second, a super tweeter needs a high order filter set to a point that takes over for a tweeter with a high freq. rolloff. They are not meant to compete with each other. Third, plopping on top of the box is just stupid. It needs to be as close to the tweeter as possible - literally cutting the baffle of one to accommodate the other. In the example of the Elac, the supertweeter baffle should be cut to butt directly up to the tweeter's horn. In other systems, the two baffles should be integrated to minimize the space between drivers.
As you know, electronically phase matching high order filters is notoriously difficult. But there's another way. Physical phase matching.
You have the opportunity to phase match the drivers by physically mounting the supertweeter forward or back from level with the tweeter (the Z axis). You could use 1/16" or 1/32" shims until you get good phase matching. This informs you of how well the two will work together, and how deep you want to cut the speaker baffle on the final build. Basically build a test baffle to get it right before cutting the real speaker box up. Even so, I understand a lot of people don't like the idea of cutting into their factory spec speakers and modding their drivers. Too much work and planning, and it takes some confidence in your skill level. But you don't have to worry too much about phase matching high order filters when you control the mounting depth of the supertweeter. And since we're talking about a crossover point that should be no lower than 12KHz, it doesn't take much along the Z axis to make a dramatic difference in phase matching.

All in all, you hit on some of the right things to do with a super tweeter. But you can take it one more step. Try a 3rd order filter on the ST, add a lowpass filter to the tweeter, and cross them over so they don't compete much. Then play with physical mounting on the Z axis to phase match them. I'm sure you'll love the result.

[vinylcabasse]

i'm 38 now but have been messing around with all things audio since i was a young kid. i remember playing with a signal generator program (can't remember what it was, perhaps the one built into winISD?) and being able to hear the super high pitched squeals at 19khz, approaching 20khz. now its my tinnitus that covers that range and basically everything past 16khz. i hope that one day modern medicine will come up with something to revive or build new hair cells in our ears... seems like it should be possible

[terryilett]

Although not DIY, you can purchase Aperion aluminium ribbon super tweeters with built in crossovers that sit on top of your existing speakers, very good products, I have mine sitting on top of vintage Altec Santiago's, i'm happy with the result, all the best.

[Mrstevebro]

Help me to understand the PureAudioProject Duet15 no box open feel?

[woodyTM]

I always appreciate the time and effort you put into these experiments; best part is that you'll always have people in the comments finding ways to dispute against everything your experiment shows, which contrary to popular belief, is a good thing because theres always room to expand our collective knowledge and refine the experiment process.

[PirxthePilot]

One thing that was not mentioned i think ...volume level...at the time of test. These days i listen to music at around 55-65 db. At that level my tower tweeters were definitelly sat backward and lacked detail. Application of diy tweeters (aluminium) vs (speaker tweeter soft dome) with L pad set at 6db and high pas filter at 3 khz appeared to remedy the issue. I can now fully hear strings, platters, bells, air in the room...so to each of their own. I did play with the spectrum analyzer and it does appear that I am experiencing some frequency rolloff pass 8khz - but honestly for what i gained which is almost linear response to that point - I am really happy. With that said - all was done on tower speakers with unknown level of degradation off crossover circuit...it just does sound great now...😂

[SuperAmazingAnt]

I have suggestions for videos and maybe they are fun also. 1. Testing/measuring how adding neodym magnets to speaker elements affect them. Elements with neodym are pretty expensive, magnets not so much.
Should make them, more sensitive faster and more power efficient(?) 2. Measuring / example how double of cabinet size affect the sound.
3. Full range speaker elements, does adding tweeter make them better.

[eleckson]

Kinda crazy to not mention the model of ELAC in this video at all, nor the description.

[larryy]

thanks, you helped me a lot. I had an idea to add a super tweeter to the acoustics, but I ran into the calculation of the band-pass filter and phase matching. Now I understand that the game was not worth the candle

[alaingingras]

May I ask what Measuring Program you are using please ?

[Audio_Simon]

Nice video! Interesting the comb filtering was more at low frequencies. I guess at the higher frequencies it is there too, but so narrow and evenly spread it is difficult to detect.

For what it's worth my supertweeters have a 2nd order acoustic high-pass at 16khz. Also recommend to keep some distance between them and the main tweeter or even point off-axis a bit.

[perlman]

Well done video with something to think about before leaping. From what I know of them, my EPI 100s don't go above 10Khz and so a super tweeter set above 9Khz might be a good thing in my situation. What model Elac speakers are those?