This information can be problematic.
Any given waveguide can be “good” but it largely depends on the overall design with particular attention paid to the high and low-pass results.
For instance you can quite easily get a bad result if the resulting dispersion at cross-point is higher in Spl off-axis than it is above and below the cross-over freq.. (..and it happens frequently with <=6” waveguides, and those looking for a crossover near 2kHz with a typical 6.5” driver.) That results in a narrow-band broader dispersion that screws with the power response (as a uniform result) and almost certainly is more reactive to baffle edge diffraction.
Yes you can lower non-linear distortion regardless, but unless it’s higher order distortion (5th) that’s “peaking” very narrow-band, then odds are you won’t really notice it. (..though you do need to do some Spl modeling to see if issues like this do become a problem within the intended play-back range.)
Again, getting a “good” waveguide result is often more complex that requires greater knowledge than many have. Of course for those that understand the various issues it can be fairly easy to get a very good result, and quite possibly easier than without a “good” waveguide.
Is this the time to suggest 'build a quality kit' to the OP (which has already probably been done, but I'm boss-eyed this morning)? It will save an awful lot of issues & provide a good basis to move forward from in design & understanding, if the urge even strikes. Troels, Zaph, PBN, Heissmann, even Pi if you're wanting a more pioneer-style design -plenty to choose from.
As rule of thumb - waveguide should be in the dimension of the mid driver. This Seas frontplate thing is not a proper waveguide.
@augerpro has great design for printing in different sizes: https://www.somasonus.net/waveguides
@augerpro has great design for printing in different sizes: https://www.somasonus.net/waveguides
Treble often sounds edgy when it's not right, but when it is right it sounds like the source. This means it can sound dull, sharp or just right depending on the program material.. however it will still be different to a bad speaker.
I disagree with the inference that narrow directivity is like having a focussed torch/flashlight shined in your eyes. It's not the same thing at all.
I did a bit of reading on this about six months ago, and it seems for a roundover to be effective, it should be much larger than one would expect; like say 50mm radius. But I'm damned if I can find the reference now - all I remember was that it was from a respected designer who posts on this forum and who'd done tests on various sized roundovers.
Lot of misinformation as usual. And some common sense posts too! "Silky" would indicate you want a warmth or colouration, but I get what you mean.
So, dome tweeters, .75" may have the extension but not the efficiency. 1", more the norm, but laws of physics dictate they beam around 12Khz upwards simply due to diameter. They also break up around the same frequency. The move towards beryllium, apart from cost, is to push the break up to a higher frequency. Silk dome, loved by companies such as ATC, smooth over the break up but still beam of course..less accurate than the more pistonic hard domes..but may have the "sweetness" you seek.
You can avoid beaming with wave guides/cd horns but you may not want to go that route.
Which leaves us with AMT's and true ribbons. If you are looking to crossover at higher frequencies these may offer the solution as you can choose a smaller "ribbon"..obviously the tall dimensions of the lower crossover variants will have reduced vertical dispersion which may be a good thing, but makes integrating their dispersion pattern with a cone driver more problematical.
But they are genuinely "fast" due to very light diaphragms and although not necessarily very extended in the hf display negligible break up at audio frequencies as can be seen on their burst decay and cumulative spectrum displays. And could be what you seek.
I have specific experience with the Visaton magnastat which has been around for many years and can confirm it does indeed sound very sweet with the "air and transparency" many seek.
Look at me getting all subjectivist there, better go and lie down..😀
So, dome tweeters, .75" may have the extension but not the efficiency. 1", more the norm, but laws of physics dictate they beam around 12Khz upwards simply due to diameter. They also break up around the same frequency. The move towards beryllium, apart from cost, is to push the break up to a higher frequency. Silk dome, loved by companies such as ATC, smooth over the break up but still beam of course..less accurate than the more pistonic hard domes..but may have the "sweetness" you seek.
You can avoid beaming with wave guides/cd horns but you may not want to go that route.
Which leaves us with AMT's and true ribbons. If you are looking to crossover at higher frequencies these may offer the solution as you can choose a smaller "ribbon"..obviously the tall dimensions of the lower crossover variants will have reduced vertical dispersion which may be a good thing, but makes integrating their dispersion pattern with a cone driver more problematical.
But they are genuinely "fast" due to very light diaphragms and although not necessarily very extended in the hf display negligible break up at audio frequencies as can be seen on their burst decay and cumulative spectrum displays. And could be what you seek.
I have specific experience with the Visaton magnastat which has been around for many years and can confirm it does indeed sound very sweet with the "air and transparency" many seek.
Look at me getting all subjectivist there, better go and lie down..😀
Effectiveness depends on wavelength, and wavelengths that diffract at the edge depends on the source and size of the baffle. The bigger the radius the longer wavelength it helps.
To make most effective roundover it must start immediately beside the transducer so that it is effective to all wavelengths that are supported by the baffle. For longer wavelengths the baffle/box is basically invisible. Basically sphere is the most effective roundover, it has biggest roundover that fits on the "construct".
So all you can do is have as big of a roundover as you can fit for the construct, the bigger the structure the bigger the roundover must be to be effective for the whole bandwidth that diffracts at the edge.
In practice, start your roundover immediately beside the transducer. If you make very small of a box/construct, no bigger than the transducer so that no roundover fits, its as big of a roundover as possible and best performance as can be for the construct 😉 To simplify even further, just minimize flat baffle area.
If you have a horn/waveguide, it would diffract as well with the bandwidth that is not beamed and reaches the edge. Because the source is small compared to construct the edge must be rounded over for the waveguide as well, a rollback. This is not trivial, but with BEM simulation it is possible to tailor the edge diffraction to such extent that it can be utilized to help narrow response without having too much effect on listening window, or almost completely prevented but even then some sound goes all the way around the construct and some ripple remains, close to listening window.
For direct radiating tweeters with the standard 4" flange there would be some because the flange prevent you starting the roundover where it should. In this case you could extend the "source" with a waveguide past the flange, but one must roundover the waveguide as well.
50mm is reasonable because it can have an effect an octave below the most sensitive region (2kHz from 4kHz), and some effect as low as 1kHz where we are becoming less sensitive to diffraction. (Although we're still sensitive to the response issues it creates.)
Yeah, I don't know what bandwidth diffraction would be audible, and where less so. If it's less audible below ~1kHz then that would make another waypoint to navigate.
i just feel like i wouldnt learn much from a kit, so i might as well start constructing a "easy" speaker first
im currently planning to build a WAW, hence the question about different tweeter types, specially compared to (small) fullrange drivers and what gives the "smoothest" (but still uncolored and great clarity) treble
Also some think the trebles should be corralated to bass extension too for a good sounding equilibrium . For illustation a 2 ways with short bass and too much extended trebles may sound too treble-ish (fatiguing, "piercing "?).
toeing the loudspeaker to fine the best equilibrium at listening position can solve some problems sometimes : pad down some areas like the 5k to 8k if bumpy. Also some others have noticed the magnitude at listening position (power response of the speaker) should slowly decrease from the bass : 5 db to 10 db difference between the low and the highs is not a ridiculous targett for a good sounding loudspeakers. Not too high odd harmonics numbers for the treble unit as well or some "harsh" poly caps can ruin things also !
Many things to consider ! Personal tastes and quality of sources and material not the lasts !
My rule of thumb : One should have a good loudspeaker first at home for reference and listening to music to face to the long journey of diy route 🙂 !
Seas 22 TAF/G : the Hiquphon of the poors (with beyrillium like high end break ups)
BC25TG : a treble that is not "there" but can be "spatious"
ring domes
maybe waveguide with very narrow baffle around with appropriate roundover or smooth recess (around 30 degrees?)
... may please the op if I try to understand the "silk" word as explained. With the appropriate filter and front baffle and.... and .... !
toeing the loudspeaker to fine the best equilibrium at listening position can solve some problems sometimes : pad down some areas like the 5k to 8k if bumpy. Also some others have noticed the magnitude at listening position (power response of the speaker) should slowly decrease from the bass : 5 db to 10 db difference between the low and the highs is not a ridiculous targett for a good sounding loudspeakers. Not too high odd harmonics numbers for the treble unit as well or some "harsh" poly caps can ruin things also !
Many things to consider ! Personal tastes and quality of sources and material not the lasts !
My rule of thumb : One should have a good loudspeaker first at home for reference and listening to music to face to the long journey of diy route 🙂 !
Seas 22 TAF/G : the Hiquphon of the poors (with beyrillium like high end break ups)
BC25TG : a treble that is not "there" but can be "spatious"
ring domes
maybe waveguide with very narrow baffle around with appropriate roundover or smooth recess (around 30 degrees?)
... may please the op if I try to understand the "silk" word as explained. With the appropriate filter and front baffle and.... and .... !
Last edited:
You'd learn a lot about driver selection, how a good designer integrates them, what tradeoffs they make (since every speaker is about selecting compromises), how much of your beliefs, if any, they show to be accurate, and have a quality performance baseline from which you can assess future projects. If you don't think that's learning much -well, your choice, but you're badly mistaken. Because starting by 'constructing an "easy" first speaker' when you don't really have a good handle on what you're doing isn't likely to get you very far, since it's not in fact all that easy. Especially if you're new to speaker design, remember the olde adage: 'some say he only knows two facts about ducks, and both of them are wrong'. Nothing wrong with that -we're all learning all the time, but it can get very frustrating, and very expensive, very quickly, so a few shortcuts are usually a sound investment. You pays your money...
Last edited:
Well I never made your comparison, But
You would be wrong. 😉
Here is another thread that I mentioned the same subjective result:
https://www.diyaudio.com/community/...xial-driver-thread.276803/page-7#post-4412174
-if you continue to read through that thread you'll note the same conclusion for the finished design.
I would note that it is dependent on just how different the dispersion patterns are: if we are only talking about a very shallow waveguide with something approaching 140 degrees then the effect isn't really noticeable.
Also notably: when you get into extreme directivity with an array that's only about 15 degrees wide in directivity that it doesn't just sound closer to you (and as a result louder), but that it actually sound's as if the source is at whatever ear it's aimed at - it's creepy (particularly if its a "whisper" sound).
Last edited:
true but imo i can have most by just looking at the designs, building it will not tell me much more, of course listening expierences are not a bad thing, but i feel like i should really build my own to learn something (im also more of a practicle guy...)
what i find currently most difficult is finding drivers without being able to look at waterfall and distortion measurements (since there are in most cases non)
Perhaps those things are not as necessary as they seem. Drivers are more or less minimum phase in their passband, making a waterfall plot redundant there. Flat/smooth response takes care of that.
im just surprised that there are not many talking about different sound of different type on tweeters, as some sound its just frequency response and dispersion/distortion
dont have AMT`s like more attack or something?
dont have AMT`s like more attack or something?
danny of gr research makes it look like a waterfallplot can reveal resonances sometimes better than the on axis frequency response measurement, same goes for distortion if i understood things right
A good tweeter will add all the attack of the source material, but you'll hear no attack unless it's there. If you want more attack, you are asking for more distortion (not necessarily HD).