Horn Design Experiment - Does Shape Really Matter?

[@IBuildIt]

How much does the shape affect the performance of a horn or waveguide? This video is made to answer that. Each of the four are exactly the same size and thickness, making all the horns the same depth. Changes in shape are to the flare from completely flat to fully rounded. As noted in the video, the flat one has a significantly larger hole at the throat, and that's the biggest difference between the samples. Best overall is the fully rounded, with the smoothest response and the most horn loading gain centered at around 2800Hz. BUT, and this is important, "best" in this instance probably won't translate to audibly better for most serious listeners. Younger people who still have the capability to hear those high frequencies may hear a difference, but preference for sound quality is subjective, even though one measures objectively better. The response is virtually identical below 8KHz and there's not a lot of musical content above that in most songs. As the horn gets deeper the shape may have a more pronounced effect on the response, but the changes would have to be large to result in an easy to hear difference. One finding is that it seems a more important factor to avoid cancellations in the higher frequency (like that 13KHz null in the flat horn) is to make the throat as close to the tweeter dome as possible. This requires more tests, and I'll do that the next time I design a horn. Watch the video showing the horn I made for my speakers here: kzbin.info/www/bejne/oGfPeJJ7hM6pbrs

[@woodyTM]

As far as i'm concerned, I'm not a member of the chiroptera, and most engineers begin rolling off overtones anywhere between 12-16khz anyways and since human beings are only REALLY GOOD at telling the difference between stimuli rather than specifically identifying which is actually objectively better, I figure with most of these sorts of experiments: the die hards will continue to make fools of themselves, while everyone else, based in reality, (and those of us specifically in untreated listening rooms) will continue to be happy as "none-the-wiser". I so ever much love the content you produce and hope you do more just like this one!

[@larry999y]

Thank you. you are doing very important and useful work

[@matthewchen1392]

Thoughts on the effect of throat depth in horns? How do things change for high frequency vs low frequency drivers?

[@ppdan]

Horn design is really critical for compression drivers, the slighest error in the horn can have catastrophic results. As an example : I had a client/installer that had a frequency missing (+20dB dip at one specific frequency) on a cinema setup (big horn loaded cinema set). The HF horn and driver come in seperate boxes and when mounting the driver on the horn he used the 1-2mm rubber seal from the package (those drivers come mounted on a wooden board in a box). After removing that rubber (wasn't supposed to be used when mounting the driver on the horn) everyhting was normal. Funny how a 1-2mm rubber seal can mess up the frequency response on a +50cm horn.

[@bogu2359]

Scary. There's no fun in audio without a decent measuring set!

[@scottwolf8633]

Very impressive unbiased empiricism sir. Did you take a response measurement of the tweeter sans horn loading as a control, to compare with?

[@IBuildIt]

I showed that in the original video: kzbin.info/www/bejne/oGfPeJJ7hM6pbrs

[@hubbsllc]

Appreciate the experiment, thank you. I’m about to try one of my own to see if placing baffles to either side of vintage Acoustic 807 cabinets will extend the bass response.

[@picassoimpaler3243]

Very interesting video. But when you put up the overlaid graphs at the end, you really should have either put up a legend as to which graph was what waveguide. It was interesting to see the differences, but i really have no idea which is which. Other the one with the enormous diffraction dip. Anyway, thanks for the vid!

[@Frisenette]

A longer flat horn perhaps with a non abrupt edge (spiky, felt, slanted etc.) would perform better. That is, the same performance as the rounded. So you can make the horn out of thick paper if need be.

[@kenanderson2216]

I'd like to see how Augspurger and JBL design their mid and/or high horns(on the high end stuff). They use a wide flat horn. Augspurger says even the design of the little dividers in the horn make a difference.

[@lesliesmart2879]

Perhaps consider brass band instruments, trumpet, baritone, tuba, etc - all using similar (or switch around) mouth pieces, playing variety of notes, to then compare measured results.

[@rcpmac]

Would like to see the curve for the tweeter without horns. Perhaps that was in the original video.

[@IBuildIt]

It was, I showed that in the original video: kzbin.info/www/bejne/oGfPeJJ7hM6pbrs

[@dhariniparker]

Can you please design a shelf in wooden chest which mounts to the lid .

[@scharbon4089]

Would you try concave curves or even straight angles? 😮 Why only convex curves?

[@Aswaguespack]

I can see a new publication in the making “Speakers Design for Senior Citizens “ 😂

[@AdrianHiggins83]

Wonder how different material would affect the results. Wood plastic aluminum

[@daveg8htfadlibaudio250]

Hi John, How much gain did you actually get from the best shape waveguide compared to the tweeter on it's own, just curious as you then could use a bit more resistive attenuation in a passive system. Nice work Regards Dave. 😀

[@IBuildIt]

Around 6-8db with a horn this shallow.

[@davidkclayton]

A plot would have been nice on that to compare

[@IBuildIt]

I showed that in the original video: kzbin.info/www/bejne/oGfPeJJ7hM6pbrs

[@daveg8htfadlibaudio250]

Cheers john, I did it with 3 layers of 15mm plywood using a B&C DED10 and got about that sort of gain.@@IBuildIt

[@daboyakasha101]

nice video to get brains churning! might be out of the scope of your intended video, it would be interesting if you took more off axis measurements and presented normalized polar maps. Sometimes one waveguide might have poorer on axis response, but a more well behaved off axis radiation pattern. Especially if you will equalize the tweeter anyway. Of course there's so much more to this. I've basically just been doing this lately (except measuring up to 90degrees in 10deg increments) because I have no idea how to theoretically design a horn/waveguide so I've just been brute forcing it with mdf/3dprinted prototypes and its helped me learn a tremendous amount more intuitively

[@craigenputtock]

Wow youre an amazing speaker builder. Thanks for this.

[@chrisferris3303]

This is great John. Love your videos.

[@JukeboxAlley]

Heres maybe a good question, but does the felt and foam edge trim around some speakers like the ls3/5a, proacs, dunlavy, jbl and many others actually make some sort of difference, or would they sound just as good without the felt/foam edges?

[@IBuildIt]

Felt or foam around the tweeter is there to help stop diffraction. But while diffraction is easy to measure, there haven't been any definitive tests that prove it's audible while playing music. In other words, it probably sounds exactly the same with or without it.

[@JukeboxAlley]

@@IBuildIt appreciate your take on it, that's pretty well how I looked at it as well, I'd say if you did a blind test of say an ls3/5a with and without the felt, I'd almost guarantee you couldn't tell any differences.

[@IBuildIt]

As a technology matures it becomes heavy with marginal improvements. I'll use racing as an example and how in the beginning it was just to make and race a fast car. Now it's computers and hollowing out bolts to save weight and every other tiny tweak that can be done to maximize what you can get. But that's competitive racing, while audio reproduction for listening enjoyment is something completely different. Tweaks in audio that make a measurable difference - an objective improvement - may not translate to better sound quality. That's what a lot of guys aren't getting these days.

[@dy1an]

I notice on your open baffle speakers the woofers are also set back from the face with a bevelled surround, and I wondered what (if any) difference that makes compared to mounting them flush with the face? I don't know enough about speaker design, but figured you'd have more insight on that. I'm planning on building my own at some point, and my current design has the woofers set back like yours purely because I preferred the way they looked like that 😁

[@IBuildIt]

Has no effect on lower frequency drivers.

[@dy1an]

@@IBuildIt that's great to know. Thanks 👍

[@KravchenkoAudioPerth]

Agree with your findings John. And also can point you to a few other things. I have designed quite a few of these. And to get right is not exactly easy. Especially when you are trying to keep your top end fairly smooth and create a boost around 800 to 1000 hertz where I would argue is most important for a loading of this nature. I have spent month son different varieties of horns and most aggravating the coupling at the throat. That is where you fight the most to get past 15 khertz. And I agree, we can't really hear up that high most of the time. Happily by the time the music gets to the mics there is not all that much content up that high in the first place. Mark

[@IBuildIt]

The top end roll-off is an inherent problem with horns and the only effective way to reduce it is to make the horn less deep or configure a phase plug. The deeper the horn, the more gain it will provide, so it's all a trade-off. Lower range gain would come with making the mouth larger in diameter, but then you have to make the horn deeper to do that and in turn make the throat cancellation more extreme. A horn like this gives the boost centered at around 2.5KHz, and you can then pad that down to extend the coverage for a lower xover point and reduce distortion. The cost is that roll-off in the upper treble that probably won't be an audible problem. In the original video ( kzbin.info/www/bejne/oGfPeJJ7hM6pbrs ) I show the response without the horn and also play recorded samples to hear the differences.

[@KravchenkoAudioPerth]

@@IBuildIt Without giving away to much of the engineering work we have pulled off I can tell you that we have a similar horn on a 28mmk dome that has reasonable response out to 17 kilohertz. It took quite a lot of work to get that rabbit out of the hat. So details stay in-house. We were after a rise in efficiency at 1kilohertz to facilitate a crossover where the 6.5 was still giving us a hemispherical dispersion. Getting some gain, and then a reasonably extended high frequency response was no small bit of work. If you want to simulate you can use Hornresp for this. Mark

[@IBuildIt]

As shown with measurements, I'm getting what I'd consider reasonable response out to 17K on the horn I made in my speaker. It starts to roll off around 12K and is only down by 8db at 18K. Given that there are some very highly regarded tube amps that have that kind of roll off in response, I'd say that's well within what's acceptable for audiophile listening. And, I'm padding it down equally - across the full response - when I could be padding down just the boosted range from 2K to 8K and make the response even flatter. That's the way most handle that natural roll off with a horn, by selectively padding the response. But like I also said, a lot of that is for bragging rights only. Most serious listeners are fooling themselves if they think they can hear the difference, because their ears have lost that ability to hear the high frequencies involved.

[@paulhirst3548]

Well done and I agree that old people would struggle to hear much of a difference. I do not think that it should stop people from trying to build the "best" speaker that they can ( according to industry standards ) as long as they realize that they may not be able to tell the difference in the end. If you have hearing aids that have settings for music, try switching between the music and automatic settings to see if you can tell the difference. For me, I turn mine on to music and I like to think that there is an improvement but in all honesty, I doubt that I can determine one way or the other.

[@IBuildIt]

All shallow horns like this will have the roll off in response in the upper treble, unless they use a phase plug to try to stop the throat cancellation that naturally happens. So any time you see a speaker that has this kind of waveguide, it will have that drop-off above 10KHz, depending on the depth of the horn. What can be done is pad down the response below the roll off to make the drop less of a factor. In other words, take the extra efficiency the horn provides and throw it away (while reducing distortion in the lower frequencies) to better match the woofer and make the response flatter. The measurements I did here are raw, before any crossover magic happens, so you need to keep that in mind when watching.

[@k-sell4065]

Good info thank you. New sub here.

[@st170ish]

Great video John, thanks mate

[@kennethnielsen3864]

Thanks for sharing.

[@dennismiller5725]

Great test!

[@cougar1861]

What's the difference between a horn and a waveguide, if any?

[@IBuildIt]

I made a video on that a few months ago: kzbin.info/www/bejne/Z6O9c3-ImcZ0pdE

[@cougar1861]

@@IBuildIt Thanks

[@misinformationwithrandy]

Being a noob, how does it impact the tweeter when the horn has that later widening that I've seen in some speakers? Is that predominantly about targeting sound?

[@IBuildIt]

That flattening is to make a less abrupt transition to the front baffle. The flare changes from high slope inside to low slope outside and that may be more important for a deeper horn.

[@BostonMike68]

I don't know if you watched Joseph Crowes channel but he's an expert on horns and wave guides . The horn is usually set back if designed correctly for time alignment because the waves of a tweeter are faster than a woofer and a big reason is it raises the SPL of the tweeter again if done correctly. Imagine cupping your hands and talking through them or someone yelling through cupped hands its louder and that's done so you can crossover the tweeter at a lower frequency because the larger the woofer it starts to beam the higher the frequency for instance you have a 10" woofer it will start beaming at around 1400 Hz . So you want to crossover low but a lot of tweeters can't be crossed low so that raises the resistance frequency which determines how low you can cross the tweeter. He's able to use a 1' order simple crossover also which the less parts in the path the better so they say. Idk the best I have I hope I make sense but check out Joseph Crowes channel if you want to see more about horns that's his thing he makes a living off of making horn designs

[@bogu2359]

What have you done?! If it is D2905/950000, it has almost flat characteristics (on-axis) between 2 and 20 kHz. With horns, you have 6 dB difference between 3 and 10 kHz. It is even bigger (~14 dB) between 3 and 18 kHz. It is tragedy ;)

[@jacquesbolduc1070]

Tanks for the nice job show to us 👍

[@nathandaniels4823]

I’d love to see if there’s a difference between guide and no guide. Perhaps you’d find a waveguide is not of much benefit whatsoever.

[@IBuildIt]

I showed that in the original video: kzbin.info/www/bejne/oGfPeJJ7hM6pbrs

[@kwilj]

Joseph Crowe collab pls? Or maybe you could test some of his designs (if Patrons bought the CAD files from his site)? I know you've already got what you want but I'm curious about the differences and whether it's worth the extra $

[@IBuildIt]

I have the parts to build his Sabourin design, and I may get that done over the winter.

[@jonatanxzavier1445]

*promo sm* 💃