Well, yes, but there's a bit more to it than that.
In the vertical plane it's narrowing smoothly and gradually with frequency yes. In this plane the dispersion at the top end is set by the length of the ribbon, but at the low frequency end (down to about 2Khz) the waveguide takes control and restricts the excessive increase in dispersion that would otherwise occur.
In other words its more directional at the bottom end of it's range with the waveguide than if the ribbon was just suspended without any form of waveguide, thus reducing the total shift in directivity index with frequency. (If you by necessity have narrow dispersion at high frequencies, controlling the dispersion at low frequencies to reduce DI shift is a good thing)
In the horizontal plane the directivity is relatively constant. Yes it starts to narrow above 10Khz, but less than a typical dome, with directivity maintained down to about 2Khz.
It's probably helpful if I post a couple of other tweeters so you have something to directly compare against with the same scales and measurement setup.
The first is a Peerless 811815 fabric dome. A fairly typical dome tweeter, nothing fancy. Compare that to the horizontal response of the ribbon and tell me which one is narrowing more with frequency... in fact it's all over the place. Below 3Khz it's almost +/- 90 degrees (eg the angle of the baffle) while at the treble end above 10Khz it's narrower than the horizontal ribbon response. Total DI shift is far greater than the horizontal response of the ribbon.
As well as that there's a surprising amount of on axis cancellation, especially below 3Khz, as well as around 8Khz. On these graphs anywhere where you see orange along the middle and dark red above and below is an indication that the off axis response is greater than the on-axis response, which is never a good thing in a driver and shows poor directivity control. (Another way of looking at it is the response is not falling off monotonically as you go off axis) The off axis response is up to 3dB greater than on axis in many frequency regions. (Very obvious in polar plots too, although I wont bother posting those)
There is a small amount of this in the horizontal plane at 5.5Khz on the ribbon but it's quite minor in the order of 1dB and that's the only place it occurs, certainly not nothing below 3Khz, and nothing in the vertical plane, where it falls off monotonically.
The second driver I measured is a Peerless 811647 - an interesting second comparison because it's basically the same dome driver as the first one, but with a horn on the front instead of a nearly flat face plate. It's not a big horn, only about 40mm deep, but it is a horn profile rather than a waveguide profile.
If you compare the second graph you'll see that below about 4Khz the horn does absolutely nothing for directivity control - all the off-axis response increases at the bottom end are the same. This is an example of a horn doing nothing for directivity control when it's too short for the frequencies...All it's done to the directivity is to make the top end narrower, a lot narrower than the ribbon horizontally, and almost as narrow as the vertical response of the ribbon.
The polar responses of both those drivers are a mess IMO, (but quite typical) they're just not good drivers, and both have sibilance problems which is one reason I never used them in a project. I'll post the frequency response and CSD of them in another post because interestingly enough, both of them have a poor decay response right smack in the sibilance region, and it might help to get me back on topic
I think if you study the graphs I attached, you'll see that a horn that's too short does nothing for low frequency directivity control, but a waveguide can.
Baffled by this line of discussion - where did I say speakers outside sound the same as inside ? Why would I ever say something that ridiculous
Of course they sound different outside. Better - like most speakers do when room influences are removed...Attachments
Last edited:
Ok back on topic
Here are the CSD plots of the three tweeters up against each other, and a bit of analysis.
I've remeasured the ribbon since posting the previous CSD so all 3 were measured in the exact same test environment. I also lowered the cut-off level to -30dB to better show low level resonances. (I can't go much lower than this without room noise from the measurement computer showing up)
The first is the ribbon - a tweeter I consider to have no problems with sibilance, and plenty of "sparkle". Not a perfect CSD by any means, but 5-10Khz is pretty clean. (Maybe a slight cavity resonance at 15Khz - hard to tell if this is really there or not as I know my mic has a peak in that region that will interact with the speakers response)
Next up is the soft fabric dome (Peerless 811815) - It's by no means a high end tweeter and I'm sure there are plenty of much more expensive domes that will measure a lot better, but it's probably representative of an average mid price dome.
Subjectively it sounds like a "typical" soft dome tweeter - it's fairly crisp (I wouldn't say it had sparkle or air though) but sibilance is quite noticeable on many recordings, nothing terrible that non audiophiles would notice or complain about, but I certainly notice it and find it annoying.
Between 5-10Khz there are two significant resonances, one at 6Khz, one just over 8Khz - both right in the sibilance region. Of particular interest is that the 8Khz resonance is an "anti-resonance" which actually occurs at a dip in the frequency response. (You can see the dip at the top of the waterfall, and as the response decays it then comes back up again - eg a decay that is not falling monotonically)
In a speaker an anti-resonance is often a sign of cone breakup. In the case of a dome tweeter it's almost certain that at this 8Khz resonance the middle of the dome is moving in the opposite direction to the outer part of the dome - thus causing partial cancellation. When the signal is removed and the outer part of the dome stops moving the middle continues to oscillate and flap around at this frequency and thus the ridge appears on the graph.
These resonances from 6-8Khz in my opinion are the cause of sibilance in this driver. If you look at the frequency response (at the top of the waterfall) it's not "hot" in this region at all, in fact there is a broad depression through the sibilance region which isn't there on the ribbon.
Ok onto tweeter number three - the Peerless 811647 soft dome horn - as mentioned in a previous post it's basically the same tweeter as the dome, but with a 40mm deep horn front plate instead of a flat face plate.
Subjectively I would say this is easily the worst of the three as far as sibilance goes. It's a very sibilant driver to the point where "essess" on just about everything sound bad. It's also got no sparkle or air at all, due to the rapid fall off in the treble above about 14Khz.
If we look at the 5-10Khz region again we can see that the resonances are even worse than the 2nd driver, and occurring at much the same frequencies - not surprising when it uses the same soft dome diaphragm. One critical difference though is that there are no "anti-resonances" visible, and much less of a dip in the response in that region.
Why ? Because as well as a horn it has a large phase plug in front of the middle of the dome which is blocking the out-of-phase radiation from the centre of the dome. You still have a resonance occurring because the rest of the cone is in cone-breakup as well, but you get a monotonically decaying resonance on the CSD because there is no out of phase cancellation from the middle of the dome.
In a way I think this actually makes it even more audible, because the overall amplitude in the frequency range where the resonances are occurring is increased.
Anyway that's my take on sibilance, I really do think it comes down to the decay response of the driver between 6-8Khz, (maybe even 5-10Khz) with the distortion characteristics possibly being a mitigating but secondary factor by increasing the excitation of the resonances depending on the spectra of the music. Frequency response balance only affects how noticeable or annoying it is.
Here are the CSD plots of the three tweeters up against each other, and a bit of analysis.
I've remeasured the ribbon since posting the previous CSD so all 3 were measured in the exact same test environment. I also lowered the cut-off level to -30dB to better show low level resonances. (I can't go much lower than this without room noise from the measurement computer showing up)
The first is the ribbon - a tweeter I consider to have no problems with sibilance, and plenty of "sparkle". Not a perfect CSD by any means, but 5-10Khz is pretty clean. (Maybe a slight cavity resonance at 15Khz - hard to tell if this is really there or not as I know my mic has a peak in that region that will interact with the speakers response)
Next up is the soft fabric dome (Peerless 811815) - It's by no means a high end tweeter and I'm sure there are plenty of much more expensive domes that will measure a lot better, but it's probably representative of an average mid price dome.
Subjectively it sounds like a "typical" soft dome tweeter - it's fairly crisp (I wouldn't say it had sparkle or air though) but sibilance is quite noticeable on many recordings, nothing terrible that non audiophiles would notice or complain about, but I certainly notice it and find it annoying.
Between 5-10Khz there are two significant resonances, one at 6Khz, one just over 8Khz - both right in the sibilance region. Of particular interest is that the 8Khz resonance is an "anti-resonance" which actually occurs at a dip in the frequency response. (You can see the dip at the top of the waterfall, and as the response decays it then comes back up again - eg a decay that is not falling monotonically)
In a speaker an anti-resonance is often a sign of cone breakup. In the case of a dome tweeter it's almost certain that at this 8Khz resonance the middle of the dome is moving in the opposite direction to the outer part of the dome - thus causing partial cancellation. When the signal is removed and the outer part of the dome stops moving the middle continues to oscillate and flap around at this frequency and thus the ridge appears on the graph.
These resonances from 6-8Khz in my opinion are the cause of sibilance in this driver. If you look at the frequency response (at the top of the waterfall) it's not "hot" in this region at all, in fact there is a broad depression through the sibilance region which isn't there on the ribbon.
Ok onto tweeter number three - the Peerless 811647 soft dome horn - as mentioned in a previous post it's basically the same tweeter as the dome, but with a 40mm deep horn front plate instead of a flat face plate.
Subjectively I would say this is easily the worst of the three as far as sibilance goes. It's a very sibilant driver to the point where "essess" on just about everything sound bad. It's also got no sparkle or air at all, due to the rapid fall off in the treble above about 14Khz.
If we look at the 5-10Khz region again we can see that the resonances are even worse than the 2nd driver, and occurring at much the same frequencies - not surprising when it uses the same soft dome diaphragm. One critical difference though is that there are no "anti-resonances" visible, and much less of a dip in the response in that region.
Why ? Because as well as a horn it has a large phase plug in front of the middle of the dome which is blocking the out-of-phase radiation from the centre of the dome. You still have a resonance occurring because the rest of the cone is in cone-breakup as well, but you get a monotonically decaying resonance on the CSD because there is no out of phase cancellation from the middle of the dome.
In a way I think this actually makes it even more audible, because the overall amplitude in the frequency range where the resonances are occurring is increased.
Anyway that's my take on sibilance, I really do think it comes down to the decay response of the driver between 6-8Khz, (maybe even 5-10Khz) with the distortion characteristics possibly being a mitigating but secondary factor by increasing the excitation of the resonances depending on the spectra of the music. Frequency response balance only affects how noticeable or annoying it is.
Attachments
Last edited:
Unless they're designed well, with the frequency response optimized with the waveguide an integral part of the design, instead of being optional...
Here's the FR of the three tweeters from the previous test.
Dark blue is the horn tweeter, claimed sensitivity of 99dB/W/M - it reaches about 100dB, but only 3-8Khz, below 3Khz the horn is too short, above 8Khz, I'm not sure what the problem is, maybe related to the phase plug or the horn taper they chose...flatish across that narrow range but not exactly flat overall.
Purple is the dome tweeter, claimed sensitivity of 92dB/W/M. I did a double take when I measured this thinking there must be something wrong with the driver or my measurement set-up. Tried the other unit of the pair, similar result. Rechecked measurement set-up, everything seems ok, other models of driver measure as expected.
Doesn't measure nearly as flat as the obviously smoothed (or retouched?) manufacturers graphs, with a +/- 3.5dB variation between 2Khz and 15Khz. Actually measures considerably less flat than my 8" full range drivers! Bad design I think...
Cyan is the ribbon tweeter, claimed sensitivity of 96dB/W/M. Flat +/- 1.5dB from 1.4Khz to 15Khz. (Probably well beyond that too, I have doubts about my microphones response above 15Khz...)
I know which one looks flattest to me
No smoothing applied on any of them by the way. (Well, 1/24th octave, but that's essentially no smoothing)Attachments
Last edited:
I will check this again tonight... I definitely tried flipping the tweeter polarity and was rather bemused that it seemed to make NO difference whatsoever, what I don't remember was whether this was the 3rd order butterworth, or 4th order bessel... I'd have thought at least the 4th order should show a null when 180 deg out of phase... if it was the 3rd order, then it sounds like it is on the right track
Tony.
Well, if phase-flipping makes no difference, it must an odd-order crossover, never mind what is happening electrically. What I look for are 1/2 octave wide regions are elevated in the "test" connection - those are the areas where the drivers are starting to go out of phase with each other. If your 3rd-order crossover is free of that, good for you! That's not true of many commercial high-end speakers, which have obvious phasiness. I don't like going into the swamp of capacitor coloration, but if there is any, the tweeter caps are the parts where you'll hear it. I avoid Solens in my speakers - I heard colorations that were all too similar to low-quality dome tweeters.
Speaking of which, thanks for the CSD's of the Peerless 811815 tweeter, DBMandrake! Those remind me of the wretched domes from the 1970's, when there were not many choices. The Peerless fabric domes were badly damped and prone to pole-piece cavity resonances, and the hard-plastic ones from Philips were even worse. The best tweeters of that era were the Peerless HFC225 paper-cone tweeters, along with the early Audax 1" and 1.5" domes. Fairly decent, but nowhere close to modern tweeters.
The first dome I even used that measured decently - and sounded good - was the D9000 Scan-Speak, probably designed by Per Skanning, the one with the notorious stacky dome which was a magnet for the flaky packing material that Scan-Speak used. Rather than change the crumbly foam plastic, they changed the sticky stuff to materials that weren't quite as good! Argh!
Ribbons are wonderful above 4~5 kHz, but I have reservations about the current fad of using them in audiophile minimonitors (or whatever you call a compact stand-mounted speaker that's 85~87 dB/meter efficient). The combination of a low-efficiency and breakup-prone midbass driver combined with a fashionable low-slope crossover results in a speaker that is pretty clean at low levels and carefully chosen audiophile discs, but goes to pieces at higher levels. What's missing is a ribbon-quality driver that covers the 700 Hz to 5 kHz range - those are not easy to find.
Last edited:
Yes. The driver centres were about about 13cm below the the middle of the top edge of a 39cm wide by 63cm high baffle. Some of the off axis peaking of the two Peerless drivers I complained about below 3Khz may be baffle edge diffraction effects, because neither driver has any real directivity control below about 4Khz. (I didn't try them free field as I don't have a good way to mount them mid air at the moment)
Even if so, it's representative of what you'd get with the drivers actually mounted in a real cabinet.
The ribbon on the other hand measures nearly the same on the baffle or just sitting largely un-baffled on the top edge of the cabinet as it's directivity makes it largely immune to baffle edge diffraction. (A big advantage of having a bit of directivity IMO)
Last edited:
Simon,
You obviously know quite a bit about speakers.
However when I make a point you seem to miss it rather entirely.
That's ok. I don't mind. I'm frequently incorrect anyhow. No harm no foul.
Let me try again. I suggested that you listen to your tweeters outside so that you could notice what happens to the highs once you move off axis. It's generally not so apparent inside. That's what I said. Is that difficult??
Perhaps you are aware of this, but most people, even experienced people are not and have not made the direct comparison.
The comparison of your ribbon - proving that it is better than the other drivers you show is entirely unnecessary as far as I am concerned. I have no doubt that it is a very good tweeter and outperforms most if not all domes so long as they are run in the same frequency range and within the linear power limits.
I think where we truly differ is that you appear to claim that "sibilance" can be eliminated by the use of a tweeter, like yours, alone. I think that is what ur saying? Maybe not.
Also I think you said something about running your tweeter from 4kHz. up.
That requires a very high quality driver to meet it. I don't think you offered any suggestions on that. Although that may work, it seems to me that you are into a 4 way or better system with xover points only two octaves or so apart. My experience says that I'd prefer to listen to a system without xover points in the middle of the critical "voice" spectrum, but that's only my preference and my experience. For example a high quality ESL meets that requirement.
_-_-bear
You obviously know quite a bit about speakers.
However when I make a point you seem to miss it rather entirely.
That's ok. I don't mind. I'm frequently incorrect anyhow. No harm no foul.
Let me try again. I suggested that you listen to your tweeters outside so that you could notice what happens to the highs once you move off axis. It's generally not so apparent inside. That's what I said. Is that difficult??
Perhaps you are aware of this, but most people, even experienced people are not and have not made the direct comparison.
The comparison of your ribbon - proving that it is better than the other drivers you show is entirely unnecessary as far as I am concerned. I have no doubt that it is a very good tweeter and outperforms most if not all domes so long as they are run in the same frequency range and within the linear power limits.
I think where we truly differ is that you appear to claim that "sibilance" can be eliminated by the use of a tweeter, like yours, alone. I think that is what ur saying? Maybe not.
Also I think you said something about running your tweeter from 4kHz. up.
That requires a very high quality driver to meet it. I don't think you offered any suggestions on that. Although that may work, it seems to me that you are into a 4 way or better system with xover points only two octaves or so apart. My experience says that I'd prefer to listen to a system without xover points in the middle of the critical "voice" spectrum, but that's only my preference and my experience. For example a high quality ESL meets that requirement.
_-_-bear
I'm surprised nobody mentioned what I call "The Abby Road Disease." The sound engineers at EMI's Abby Road studios seem to have invented one of the most irritating techniques in all recording, adding reverb to only sibilance of pop misic singers. Horrible and ruined a lot of otherwise fine recordings. There's no way to get rid of it, it can't be filtered out. It's like the Stuxnet computer virus, it's there forever.
I probably do miss the point of what you're saying sometimes - some of the semi-rhetorical questions you've posed through the thread like "but how do they sound outside" have left me scratching my head thinking "what is he trying to get at here", so I probably have gone off in long tangents due to misunderstanding the reason for your comment. I'm someone who likes to be precise so if a question or situation is a bit vague I like to account for all possible interpretations. I probably should just ask for clarification.
Speakers are only a hobby for me by the way and only actively for about 10 years, so I still have a lot to learn. One of the main reasons I come to forums like this apart from enjoying the chance to discuss my favourite hobby with like minded people is to learn from others and from the discussion. There are many people lurking on this forum alone who have decades more experience than me, some of them in the commercial realm so being able to read what they say and interact with them is like gold for me.
In the physical world I know almost no-one who is even vaguely interested in speaker design let alone skilled in it, maybe 1 or 2 people, and they're now half a world away since I moved to the UK. Let's face it, as hobbies go speaker building is one of the most obscure you can find...
(Whenever I tell someone that's one of my hobbies a puzzled look crawls across their face followed by "why would you want to do that?", so I usually don't bother now)Yes it'll change off axis outdoors naturally. Practically speaking with that small size of ribbon anywhere between +/- 45 degrees horizontal and +/- 20 degrees vertical is the ideal listening window, which I'd say is fine. For a ribbon twice as long (such as the AC G1) the vertical window would narrow to +/- 10 degrees, which is a bit narrow for my liking, one reason I didn't buy them. (The other being the price!)
I wouldn't use a ribbon tweeter of this type for any outdoor speaker design though, it's one thing to take your home hi-fi speakers and drag them out onto the deck facing into the yard (which I've done many times, when I lived somewhere which actually had a yard...) but they're not a good choice of driver for outdoors in general for many reasons including - the ribbon element can be damaged by strong gusts of wind, the magnetic gap is open and exposed apart from the protective mesh, so any metallic filings picked up outside are going to find their way into the gap, and they don't have the necessary maximum SPL for playing really loud outdoors - they're not a PA driver. At that size they're strictly a home hi-fi driver, but it's these reasons not directivity which are the main reason they're not suited outdoors.
Now I think it's you that's getting the wrong impression. I'm not here to say "this tweeter is better than anything else", nowhere have I suggested that, and it's certainly not true. They have their limitations, and as foil ribbons go they're relatively small and inexpensive ones as well. (I bought mine in 2003 when they first came out and since then they've more than halved in price, but they're still a current production model)
The only real claim I've made relevant to the thread, is that in my opinion this is one example of a driver that doesn't have problems with sibilance, and also that any well designed aluminium foil ribbon is also very likely to be the same in this regard, as the freedom from cone breakup in the sibilance region is inherent in the mode of operation of a foil ribbon. You can screw up the foil and waveguide design of a ribbon and get a non flat response, high distortion, and cavity resonance problems but cone breakup is unlikely and maybe even impossible even in a poor design.
This is not the case with a planar driver with a hybrid diaphragm, which may or may not have dome like cone breakup depending on how the former (kapton etc) is bonded to the voice-coil and damped etc. I'm sure some of them are really good, and some quite bad.
(I actually had a number of planar drivers on my short list when looking for a ribbon, including your Stage Acompany drivers, possibly the exact same model, as it looks very familiar, but in the end wanted to go with an aluminium foil design, price might have been a factor for the SA drivers too, I can't remember how much they were)
I think it's also unlikely that any soft dome tweeter will ever be 100% free of cone breakup through the treble, damping will only reduce it to hopefully low levels but I can't see how it can be eliminated, so it seems like a design dead end to me in this regard.
Rigid domes like titanium and now diamond show promise - in theory they can be so stiff and light that the dome wont bend below 20Khz and thus won't have any cone breakup in the audible region let alone sibilance region.
I do worry about the strong ultrasonic breakup resonance though and what IM effects this might have in band. (Let alone whether a strong ultrasonic resonance can somehow be perceived, even if not heard) Until this problem is solved I'm still weary of this type of driver, and I imagine most of them are rather expensive as well, and out of my price range.
I remember listening to the B&W Nautilus 802's a few years ago (the original model with the titanium domes not the current diamond ones) and there was just something I didn't like about the treble that I couldn't put my finger on, a brittleness of some kind, as well as what seems like a rather hot upwards sloping tonal balance. Although they were good sounding speakers overall I did actually find them a bit fatiguing to listen to.
Years later I read that the midrange to tweeter crossover on the 802's was 1st order, something that really surprised me. Even though crossed over at 4Khz it would seem to me that especially in light of Lynn's comments on tweeter IM that crossing them over 1st order is a really bad idea. Perhaps I was noticing IM distortion at higher volume levels...who knows.
As for measurements comparing to two other drivers, the whole point of that was to take a driver I subjectively don't think has sibilance, compare it to two other drivers I have kicking about that I do think have sibilance problems, and measure them to see how it fits in with my theory, not to bash them or show how much better some other tweeter might be.
I bought those Peerless drivers around 2001 for a small project and never ended up using them because I wasn't happy with the sibilance. At the time I knew what the sibilance frequency region was, but I had no software that could measure CSD, nor did I have a microphone flat enough in the treble to determine whether the sibilance might be due to FR emphasis. All I knew is I didn't like their sound, and EQ didn't seem to help much.
I've only had software that can measure CSD since last year, so when I measured their CSD for this discussion thread it was the first time I've measured them in well over 6 years, the first time I've measured them with my current test set-up, and the first time I've ever measured their CSD, so I didn't really know what to expect.
I was quite prepared to be proven flat wrong by the measurements, but almost as if on cue they both have bad resonances in the sibilance region confirming what I suspected about them for years but had never got around to re-measuring. One of the two drivers even has a large depression in the frequency response in the sibilance region while still managing to sound sibilant, which I think does help back my theory that the root cause is not just an excess of output in that frequency range.
Of course you only have my word for it that those two drivers sound sibilant, but whether something sounds sibilant is a subjective judgement anyway, and two people might not agree on the same driver during the same test.
No, I wouldn't build a 4 way system as I don't believe in putting a crossover in the midrange. As I mentioned I'm using a full range driver (see my sig picture) on the basis that I'd like the driver covering midrange to cover the widest range possible with any crossover to a woofer or tweeter occurring fairly low or high respectively. For a long time I used them as a midrange driver in a modular 3 way from 250Hz-4Khz, but since I moved and have a lot less room now I've been using them 2 way 40Hz-4Khz then crossing over to the ribbon tweeter.
Yes it does take a driver with an unusually good high frequency extension to cross over really well to a ribbon tweeter but I think it's worth it, not just to be able to use a ribbon, but because putting the crossover above the presence region seems to give a lot better coherence more akin to a full range driver, as well as minimizing the midrange requirements of the tweeter drastically improving IM and dynamic performance.
If a designer is obsessed with polar response and dead flat frequency response they'll usually push the crossover frequency as low as possible, (seems to be a trend in a lot of modern speakers along with midrange drivers with huge out of band resonances that need notching out) but if they really care about dynamic performance, low distortion, coherence and transparency, the higher the crossover frequency the better to a point, as long as the midrange driver is well behaved to a high enough frequency...
Last edited:
Oh, so THAT's how they get that effect...... Yuck....
If you listen for it, I think it's easy to identify. Every time I hear it I cringe. It's unavoidable, like a bad taste in your mouth you can't forget.
I've got a number of EMI cds recorded in the 80s and 90s that suffer Abby Road Disease.
I think I found out today why it was I was getting confused with my odd and even crossovers (at least the electrical side anyway). The tweeter is 90 deg out of phase with the mid-bass' when both are measured raw!
That's why when I only put a cap on the tweeter and a notch filter on the midbass' I get a deep null one way and ok the other. ie to get even order behaviour I need odd order electrical on one speaker and even on the other (I think)...
the sim I'd done had 3rd order butterworth slopes (acoustically) for both drivers but flipping resulted in quite a deep null. The one I simmed that had 4th order bessel slopes (again acoustic) made no difference whether I flipped the tweeter...
So now I'm wondering, can the acoustic slope appear to be even order, but the phase is actually odd order?
Also I did some measurements today and changed over to swept sine (which worked significantly better than the previous mls measurements I'd done), and even with the cap on the tweeter I was seeing very high levels of distortion in the tweeter at around 900Hz, aprox 12%! I think I do need to up the slope on the tweeter crossover... I'd attenuated it to cut back on the sibilance, but I think that was just treating a symptom, rather than finding a cure. Probably a higher order crossover (as I can't go higher in freq) and a resonant peak filter may go a long way to fixing the sillibance, which I think will allow me to not attenuate the tweeter and end up with a better balanced sound
Tony.
Very true Bear! Mine is only supposed to be used down to 1800Hz resonant freq is supposed to be 700 but seems to be closer to 1000.
If the manufacturer specifies a range of 1800 - 22000 then how low can it theoretically be crossed? Distortion at 2K and higher is 0.5% or less. Should 1.8Khz be say 12 or even 24db down?
old rule of thumb is best if 2 octaves above resonance, unfortunately I can't cross over that high without some serious respsonse shaping of my mids (and resultant drop in sensitivity).
Tony.
If the manufacturer specifies a range of 1800 - 22000 then how low can it theoretically be crossed? Distortion at 2K and higher is 0.5% or less. Should 1.8Khz be say 12 or even 24db down?
old rule of thumb is best if 2 octaves above resonance, unfortunately I can't cross over that high without some serious respsonse shaping of my mids (and resultant drop in sensitivity).
Tony.
>>> The tradeoff is made more difficult when a rigid-cone (Kevlar, metal, ceramic, etc.) midbass driver is chosen, because the onset of breakup commonly falls in the 3~5 kHz region, right where the ear is most sensitive to distortion.
>>> As you can see, the worst possible solution is a 1st-order crossover combined with a midbass driver that has a severe breakup region (Kevlar drivers, I'm looking at you). The 1st-order crossover fails to control out-of-band excursion, so program material in the 700 Hz-2.8 kHz region results in IM distortion in the tweeter's working range, while plenty of midbass breakup in the 3~5 kHz range gets through as well. And midbass breakup sounds the same as a bad tweeter, since the distortion and resonances fall in the same frequency range.
Yup, i never had good luck with kevlar midbass and domes. I do enjoy full range drivers tho and think off axis they can sound fantastic and fatigue free.
>>> As you can see, the worst possible solution is a 1st-order crossover combined with a midbass driver that has a severe breakup region (Kevlar drivers, I'm looking at you). The 1st-order crossover fails to control out-of-band excursion, so program material in the 700 Hz-2.8 kHz region results in IM distortion in the tweeter's working range, while plenty of midbass breakup in the 3~5 kHz range gets through as well. And midbass breakup sounds the same as a bad tweeter, since the distortion and resonances fall in the same frequency range.
Yup, i never had good luck with kevlar midbass and domes. I do enjoy full range drivers tho and think off axis they can sound fantastic and fatigue free.
Tony,
1800Hz. is your target minimum.
The electrical slope, imo, would do well to be 4th order, since you are already having IM and breakup issues.
To do the xover "right" the Q of the xover ought to taken into account, especially the composite response +/- the xover point, which is effected by the acoustic response in that region & so the Q of the xover. But having said that, any 4th order should probably reduce the sonic issues somewhat...
Oh - "reduce the sonic issue in the tweeter" - might make some issues in the midrange driver appear... no free lunch, just strike the best balance possible!
If you don't have too much invested in the speakers, this is maybe a time to think about some other ideas and designs, drivers. Also a good point to try different things with this set of drivers as a learning experiment...
Oh that old rule of thumb - imo - doesn't work because no drivers exist that really meet that criterion!
_-_-bear
1800Hz. is your target minimum.
The electrical slope, imo, would do well to be 4th order, since you are already having IM and breakup issues.
To do the xover "right" the Q of the xover ought to taken into account, especially the composite response +/- the xover point, which is effected by the acoustic response in that region & so the Q of the xover. But having said that, any 4th order should probably reduce the sonic issues somewhat...
Oh - "reduce the sonic issue in the tweeter" - might make some issues in the midrange driver appear... no free lunch, just strike the best balance possible!
If you don't have too much invested in the speakers, this is maybe a time to think about some other ideas and designs, drivers. Also a good point to try different things with this set of drivers as a learning experiment...
Oh that old rule of thumb - imo - doesn't work because no drivers exist that really meet that criterion!
_-_-bear
Last edited:
Crossing a tweeter at 2Khz with only a cap in series is asking for distortion imo.
A 1st order crossover looks good on paper, the tweeters i've used sofar all sounded strained when using a 1st-order (electrical) crossover at or below 3Khz.
That includes a Dynaudio d28, which is kinda old by now, but was being advertised as a tweeter that could handle 1st-order crossovers.
The worst about this tweeter-distortion to me is that it is dynamic in nature.
Gentle cone-breakup sounds less objectionable to my ears than a strained tweeter.
A 1st order crossover looks good on paper, the tweeters i've used sofar all sounded strained when using a 1st-order (electrical) crossover at or below 3Khz.
That includes a Dynaudio d28, which is kinda old by now, but was being advertised as a tweeter that could handle 1st-order crossovers.
The worst about this tweeter-distortion to me is that it is dynamic in nature.
Gentle cone-breakup sounds less objectionable to my ears than a strained tweeter.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.