Best Treble Unit?

[kensetsu]

Hi,



I beg your pardon?

First of all any ribbon tweeter has intrinsically much less moving mass than a dome tweeter.

Secondly, they usually are an order of magnitude more efficient too so that should help dynamic rendition as well.

Thirdly, since these tweeters are nothing more than a resistive element, their impedance curve is usually ruler flat making them extremely amplifier friendly.

IOW, they won't change much in frequency response due to changing impedance.

Since these are flat devices, dispersion can be a problem with some funny designs...
OTOH very wide dispersion creates room interaction problems so, take your pick...

Dome tweeters are more often than not heavier so will have higher moving mass. Hence slower transient response.

They also often contain ferrofluid material in the voice coil which makes the moving mass higher still and impedes high transient response.

Their dome shape can yield better dispersion provided it's not difracted by cabinet walls etc.

In short, anyone considering building a serious speaker should at the very least consider the virtues of a ribbon.
The difference between both technologies is NOT subtle as anyone having heard the difference will agree.

If you don't know how to correctly integrate all speaker factors into a good design then don't start with expensive dome tweeters, it's a waste of time and money IMHO.

Cheers,

Hi Frank, Did I remember to thank you... What would you think of the Esotars handing off the task to the Ravens taking at about 8000Hz? I heard thta that's where the Ravens really start to excel. They also like a real deep slope.

 

[kensetsu]

Well, I wouldn't want to say that poor measurements mean poor performance... this is simply not always the case with our current performance indicators like freq response, waterfall plots and distortion graphs. But in the case of ribbon's there eems to be a very lare gap between what the mic tells and what the ear and brain makes of it.

I beg to differ that most ribbons are better. I question if the kind of sound they produce does come closer to reality than similar priced domes. In a way I'd like to compare their virtues with electrostats, which also aren't inherently better than conventional speakers jus because they are so light. It just isn't that simple.

that's just mho.

Words of wisdom from the Doug-Man. I may not be a speaker expert, but I am a scientist. and just because some idiot chooses an arbitrary parameter to monitor, does not mean that it's the right one to measure. LOL... After Pharmacy school, I went to Purdue University for Nuclear Physics.. Yeah..and for all of you Purdue Alumni.. Dr. Stanley Shaw was still there, and just as mellow as ever. Anyway, you wouldn't believe some of the nonsense they used to believe about radiation safety.Not far off from falling off the edge of the world if you sail too far. Well.. for those of you who can't put your egos aside... Man is still not that smart. Trust me.. I see it all the time. With the results being that I have to let go of something that was thought to be fact when I was in College. Hey... Do you think that Dr. Bohrs should give back his Nobel Prize. After all, he was wrong about the structure of the atom. You know I never did like that Universal decay constant. You know. A = AsubO times E to the neg. Lambda T. Anyway.. Everything is theory. There are no real facts. Well Ok... in the USA if you don't pay your taxes the IRS will get you and lock you up. That is a fact.

 

[fdegrove]

Hi,

One of them is that it is a hundred times more difficult to find a good midrange unit wich integrates well with a ribbon tweeter. Ribbons are in basic better then domes in my opinion (and experience, I've had a lot of them) but using a ribbon doesn't mean that you will get a better overall sound from your speaker. If you want to use a ribbon you will have to start your speaker design with it, then find a good midrange for them (pretty hard job I can tell) and at last a suitable woofer. Most people work the other way around and most of those systems I have heard where actually better of with a dome tweeter.

Very well put, Sjef, you nailed it down.

It's precisely that problem most ESL designers face when they want to extend the bottom end of the speaker with a dynamic unit as well.

Or the difficulty of mating a solid state bass amp with a tube amp for the mids and highs for that matter.

Cheers,

 

[kensetsu]

Then what kind of distortion? I've never seen a ribbon tweeter perform "exceptionally" well in either linear or non-linear distortion tests. Acceptable, yes. World class, no. There is more to what we hear than simple frequency response graphs, but you must remember that if we can hear it we can measure it (though we may not be able to adequately correlate what we measure with how we perceive it). If you can hear a distortion, we should be able to measure it, and frankly frank ribbon tweeters just don't have the numbers to back them up.

However, I also like the way they sound... the maggie ribbon is one of my all-time favorites. I also like the way domes sound, and in the right design they are superior to ribbons IMO. Remember, that's just IMO though, and is exactly what my previous post was above... we all have our preferences. A dozen people in this thread proclaiming that ribbons are hands down the best tweeters on the planet just doesn't make it so. Sorry.


Of course not. But that goes both ways... all the "low mass high transient response" and "flat impedance curve" paper specs don't necessarily translate into perceived superiority.


So?

And what do you do with they are also an order of magnitude more efficient than the midrange drivers?

A definite plus, but not something that can't be adequately addressed with dome designs (e.g., active crossover designs).

Sorry, but higher moving mass does not so directly translate into slower transient response. That's a myth of audio that I'm surprised you continue to perpetuate. Of course, you do tend to perpetuate a whole suite of myths, so perhaps I should not be so surprised after all.

Transient response is governed by f/m and inductance. Ribbons do have low m, but they also tend to have low f as well. If anything, it is the low inductance of ribbons that tends to give them extended frequency response (and remember, transient response can be clearly represented by a simple frequency response measurement... it isn't something mysterious and spooky that can't be measured).


Agreed. Anyone considering building a serious speaker should consider all the possible options available. But coming to the conclusion that ribbons are superior isn't a given. There are plenty of capable and knowledgable people and companies that have decided to use both ribbon and dome designs. Like most other things in audio, there are compromises and in general you can't optimize every performance metric simultaneously.

As always... choose your poison.

Whew!!! Nice response. Spoken like a true scientists . After all. Who chooses which parameters we measure. UUhh? The person selling the product that can brag about those parameters?? LOL.. Just kidding. Always remember. Many of todays facts are tomorrow jokes... and, NO!! I didn't go blind or grow hair on my palms.

Later, Nick

 

[kensetsu]

Re: Re: Esotar Tweeters vs Ravens

LOL, I don't think you really wanted to post that did you? Its pretty obvious what the answer is, unless you have yogurt for brains.

In nearly 24 years of building speakers I have never found an acceptable substitute for a great ribbon. Domes come and go in my designs but I noticed that I never keep them in my system for very long whereas I always sit down for serious listening using a ribbon.

A telling fact to where my preferences lie.

I've used quite a few domes over time and the only one that I really have any affinity for is the vifa XT25 because it offers such a great sound for next to no money. I actually prefer this to the average SS 9700.

Well spoken.. Obviously a man with class. you have my respect.

And i know quite a few people who would say those footsteps lead to a dead end (i can only speak wrt direct experience of the B&Ws)



You beat me to it...

The best tweeters i have yet heard are ribbons, i'm no big fan of domes in general, but it must be considered that the tweeter is just part of a much larger system -- other speakers, XO, amp(s), room, etc -- there is no best. But as a class ribbons have a lot of promise. I suspect that they are not widely used as much due to their fragility as anything else.

dave

 

[kensetsu]

It seems your the one who recently had unsuccesful brain surgery.

Read this, I DID NOT POST:

"Dome tweeters are more often than not heavier so will have higher moving mass. Hence slower transient response."

I think if you read back you will find fdegrove posted this. I have made no assumptions as to why IMO ribbons are better than domes, so I'm perpetuating no myths here.

You obviously have never riden the Shinkansen in Japan. a moving mass of about 1/2 mile long traveling at 180 miles per hour. Hmmn? I don't think that you want to get into a debate on Physics with me. Hmmn? F = MA? what the heck does that mean anyway. It means that a big mass is harder to stop. Well, I'm not going to try that transient response experiment out with my car on the tree around the corner. I suspect that my car will be real responsive. Yeah.. What I just said sounds stupid, but if you give it some thought.......

 

[kensetsu]

Yes, I meant to post precisely that. I was so flippant with the "so" because Frank either intentionally or unintentionally left out so much of the equation with his explanation. My "so" was intended to cause him and others to think a bit more about what role moving mass has. By itself it tells you practically nothing. I know car analogies are run into the dirt, and are generally not very useful, but in this case the point is well enough made: it's a bit akin to telling someone that your car only weighs 400lbs and is therefore "obviously" fast (oh, but you neglected to tell them that it is powered by a 12bhp 2-cycle engine...).

And as I pointed out, motor strength isn't typically high on the list of ribbon qualities. The two do balance out (motor strength and moving mass) to net a good result as is seen by the extended frequency response and typically high sensitivity, but there is always more than one way to skin a cat. Dome tweeters usually have a much higher BL, so the additional moving mass need not dictate low sensitivity or "transient response." Again, if anything it is the low inductance of ribbons that allows them to more easily play into the higher kHz range than most domes. The trade-off appears to be distortion products... the modal behavior of a very thin and very compliant ribbon of any significant length is much more complex than a dome, and due to the low stiffness the amplitudes will be much higher as well. The added mass of a dome may in fact be its greatest benefit if it prevents high-order modes from being excited to any significant amplitude.

As for the "yogurt for brains" I thank you very much for the mudslinging. I now know how to treat any future posts you may make.

Very respectful and mature response. My hat's off to you . I'm certainly glad that you actually take the time to swat flies. I'm learning alot from the way you swing the swatter.

 

[RHosch]

lol, that's about the most entertaining string of posts I've read in a while. It's also nice to see a bit more laid back attitude... helps remind is all that this stuff is usually just for fun and shouldn't be taken quite so seriously.

And man... you managed to agree with everyone on both sides of the discussion! Are you running for any political office? You should...

 

[CeramicMan]

After much reading, I'd like make a few points.
1) I've looked long and hard and have not been able to find any waterfall plots for ribbon tweeters that have more than about 15dB of resolution. Such a low resolution is useless because all of the useful data about resonances is below the threshold of the graph - are they trying to hide something or what? By contrast, wf. plots of dome tweeters often have 20 or 25 dB resolution, which makes their measurements appear worse than they really are. Same thing with impulse response graphs: the time-scale is often downright crap in order to show that ribbon tweeters are extremely quick and clean sounding without much ringing, whereas it may just be that the statistics are worse than lies.

2)A gentle reminder that humans can only hear up to about 20kHz (on a good day), and anything above that can only cause problems. If ordinary dome tweeters can have a flat frequency response up to 20kHz, then a faster transient response is not necessary. A response up to 100kHz (as is too often the case) could be responsible for perceived problems and differences between amps, cd players and cables. The inevitable ultrasonic noises, tones and harmonics may cause audible changes due to intermodulation distortion, and trick people into thinking that they're hearing extra detail or that there are huge differences between cables.

3) There's no such thing as a linear ferrite or iron transformer. They all have non-linearities across their range of signal levels, and may therefore contribute to a lot of IM distortion, even if audible harmonic distortion is very low.

4) When making comparisons with dome tweeters, can people please note whether it's a hard or soft dome tweeter? A Vifa XT25 was mentioned earlier as being quite good, but the measurements show that the break-up frequency is below 1.5kHz (that's right: 1k5 Hz), which in my books doesn't even make the grade, so the comparison with ribbons is pointless.

Don't get me wrong, I have ******-all (damn that censoring! ) practical experience with ribbons and I want to try out a pair, but at $100s cost, round-the-world shipping to NZ plus mafia duties, I also don't want to make a mistake. A local hifi shop might have some on a loudspeaker, but those sales-people can smell a DIYer a mile away, and I'd never get to hear it. So convince me with measurements someone!

CM

 

[phase_accurate]

I have read a review on a Stage-Accompany Studio monitor (using the "Compact Driver"). They deliberatly used a higher resolution than usual in order to be able to find anomalies within the tweeter's response (there weren't many)! So much to cheating in favour of ribbons !

About high cutoff-frequencies causing problems: Do you prefer to decrease transparency in order to intentionally hide problems caused BY cd players, amps etc

Regards

Charles

 

[Hans L]

It seems your the one who recently had unsuccesful brain surgery.

Read this, I DID NOT POST:

"Dome tweeters are more often than not heavier so will have higher moving mass. Hence slower transient response."

I think if you read back you will find fdegrove posted this. I have made no assumptions as to why IMO ribbons are better than domes, so I'm perpetuating no myths here.

WOOPSadaisy... I should have said 'ageed with' instead of 'stated' when I quoted fdegrove.
And even that assumption seems to have been premature, because you say you make 'no assumptions as to why' a ribbon sounds better.

Soooo...... excuse me pretty please .... the stitches come out next wednesday...

 

[Vortex]

Don't stand for ribbon or dome ones...

I think they sound much different at all, everybody should choose his favorite one.

The more interesting thing would be "ribbons vs. ribbons" and "domes vs. domes" comparison, I think.

 

[Hans L]

Aurum Cantus G2 published distortion specs.

1. Does this bear any similarity with reality?

2. Is this good?

Unless there are some nasty high order distortions that aren't shown, if the real figures are close to what is below that looks pretty darn good to me.

That's true, although it would likely be ultrasonic. I wonder how many people have measurement equipment that goes up past 60kHz to measure K3 at 20kHz.
I am interested to know if anyone has data to back up the "high distortion of ribbon tweeters".

The scale of this graph is linear, which helps making it look good. Most graphs I'v seen show a log scale which shows what happen under 1% much better. At which spl is this graph taken? It doesn't mean a think without that piece of info.

I have a graphs of a similar model, the Aurum Cantus Model G3Si in front of me from the magazine HobbyHifi (one of two great speaker diy magazines in Germany). This model is priced slightly below the model 2, priced at 200 EUR, where the model 2 costs 250 and the 3 costs 375 eur. The graphs show K2, K3 and up to 20KHz and K5 up to about 8,5KHz, at 90db. I can't post their graphs so you have to do with my description.

Below 2KHz, distortion skyrockets like in your graph, but including K2. K3 is (substantially) higher than K2 below 4KHz, which I find strange. Between 3KHz and 7KHz, K2 and K3 linger between 0.03% and 0,3%. Now... about &KHz all distortion products skyrockets again. There goes the idea that ribbons are at their best high up in the tweeter range (neo 3, same story above 10KHz for K2).
If I compare this to the Seas Excel (the cheaper T25CF001 version, sells for 120 eur), the dome performs better on every test at virtually every freq range: freq response, waterfall, impuls, distortion, low and higher order.

A few pages further on in the test we see a Expolinear RT-5 Pro (360eur) and the hiquphon OW1 (100eur). This ribbon does perform better at very high freq. Above 5KHz it has better dist figures than the OW1. Below that point they perform virtually the same. Freq response for the OW1 is slightly flatter and extends to about 30KHz instead of the 40KHz of the Expolinear. Waterfall and impuls measure equally well.

That all for now. Maybe I can get the magazines writers to agree to post a few graphs in case the above isn't enough to support: Ribbons don't measure well.

 

[CeramicMan]

I have read a review on a Stage-Accompany Studio monitor (using the "Compact Driver"). They deliberatly used a higher resolution than usual in order to be able to find anomalies within the tweeter's response (there weren't many)! So much to cheating in favour of ribbons !...

Ok, fair enough.

...About high cutoff-frequencies causing problems: Do you prefer to decrease transparency in order to intentionally hide problems caused BY cd players, amps etc

Regards

Charles

One can't miss the sound of something they can't hear in the first place, and a pretty standardized upper threshold for hearing is 20kHz. However, I was recently experimenting with a phono de-emphasis filter that had multiple faults, and at one point (while in the audio chain) it produced a funny grinding noise that was made extremely horrible by a dominant 22kHz tone, so by chance my best estimate of the range of my hearing went up by about 4kHz. Even then, I only heard it because the signal was 10V peak at the amplifier, and I was 0.5m away. Obviously, the 20kHz limit could be extended, but then where do we stop? At a -80dB cutoff point?

CD players have a theoretical bandwidth of <22.05kHz, but in reality it's anywhere between around 16kHz and 20kHz, so that the A to D and D to A processes are simpler. With vinyl the filtering is mechanical as well as electrical, so that also has a limited bandwidth (believe it or not ).

In normal situations, a 20kHz limit is already pretty conservative considering that most people can't hear much beyond 15kHz, and that such frequencies are easily masked by lower ones anyway. IMO most of the so-called "detail" in recorded music is between around 5kHz to 12kHz, and the remainder above that is devoted to triangles and recordings of squeaky car disc brakes.

As I said before, having bandwidth far beyond 20kHz will just cause problems. It's not so much the actual ribbon's bandwidth, but because of the transformer and whatever else is non-linear in the system. The thing with transformers is that they have hysteresis and that causes distortion even at low power.

CM

 

[ShinOBIWAN]

Its interesting because people are placing so much on the measurements. It almost sounds like those are the people that prefer to read the back of the book rather than get stuck into the story.

Why are there so many lovers of ribbons if by way of measurement, they are deemed to be insuperior to domes?

Listening to good ribbons is a joy to me, never heard any distortion or compression, even at high volumes and so transparent, refined and subtle. I cannot find a suitable dome that would offer me all this in a similar presentation.

Also, I think Ceramicman posted about the frequency range above 20khz. Virtually all manufacturers are now using designs that reach far above this frequency, hell even SACD and DVD-A do this. If it was fundamentally flawed why would they do this?

UK mag What-hifi did a little test whereby they used a pair of KEF Reference 203 and disabled the super tweeter, in doing do they found the sound to have less air and spaciousness.

 

[Hans L]

Its interesting because people are placing so much on the measurements. It almost sounds like those are the people that prefer to read the back of the book rather than get stuck into the story.

That's a very poor comparison.

Why are there so many lovers of ribbons if by way of measurement, they are deemed to be insuperior to domes?

I don't know. Why do som ppl like high distortion, flea power valve amps? Why do ppl buy $10.000 interlinks?

Listening to good ribbons is a joy to me, never heard any distortion or compression, even at high volumes and so transparent, refined and subtle. I cannot find a suitable dome that would offer me all this in a similar presentation.

Maybe you haven't looked far enough? This is an honest question. Have you for instance ever heard the Scan-Speak ring radiator in a proper 3/4-way?

UK mag What-hifi did a little test whereby they used a pair of KEF Reference 203 and disabled the super tweeter, in doing do they found the sound to have less air and spaciousness.

What-hifi really has no place in a diy forum imho. But apart from that, the outcome of this 'experiment' is hardly surprising!! The supertweeter is crossed at 15KHz, with an unspecified xo slope. Of course you're going to hear that

 

[mark.carline]

JM labs / Focal

HEY!!

Has nobody here thought about the Focal / JM Labs option:

Audiom TD5
http://www.focal.tm.fr/gb/compo/tweeter/a_td5.htm

A snip at £448 each, see:

http://www.falcon-acoustics.co.uk/pl28p11.htm

 

[CeramicMan]

SACD and other such things

....Also, I think Ceramicman posted about the frequency range above 20khz. Virtually all manufacturers are now using designs that reach far above this frequency, hell even SACD and DVD-A do this. If it was fundamentally flawed why would they do this?....

People like bigger numbers. Who cares that in scientific measurements with a descending sine sweep, people only start hearing it below 20kHz? Unused performance like 30kHz, 40khz or 100kHz is way more impressive! Besides, as efforts continue towards improving the sound of tweeters in the range that people can actually hear, manufacturers noticed that sounds that we can't hear improved as well, so why not put that data onto the graphs just in case someone thinks they can hear ultrasound and are willing to pay more?

Now that you've mentioned SACD and DVD-A, "fundamentally flawed" is just the operative term. The SACD format is fundamentally flawed, and in more ways than one. Instead of using binary weighting like everyone else does, it uses an equal weighting system that is so extremely inefficient that the music is stored even less accurately than with the old CD format but takes up several times as much space, just so it can have an unnecessarily wide bandwidth of about 50kHz. It relies on high-frequency switching that assumes that cheap delta-sigma DSD-compatible DACs will be used. This makes digital manipulation of the signal a downright PITA, even for relatively simple things like digital volume control. Ordinary delta-sigma DACs can automatically convert a PCM signal to a similar high-frequency 1-bit signal too, but the only possible benefit I can see from pre-recording this onto disc is that a superior algorithm may be used to bring the DAC switching noise out of the audible frequency range more effectively. This could also be done with a bit of extra processing power in a cd player. Why they had to do it the hard way and invent a new format I don't know. DVD-A has the flaw that it unnecessarily stores ultrasonic data, when the only place that a sampling rate of 192k samples/s is acutally needed is for an oversampling ADC in the recording process. Why they decided not to down-convert the 192k to 48k is beyond me.

CM

 

[phase_accurate]

It is simple as that: Any filter outside the hearing range will definitely have an impact within the hearing range (Attention: a tweeter is also a filter !!!!).
An increased sampling rate, even when not used up to it's Nyqvist frequency will have useful impact. One can greatly relax the requirements for the necessary anti-aliasing filter.

Regards

Charles

 

[RHosch]

It is simple as that: Any filter outside the hearing range will definitely have an impact within the hearing range (Attention: a tweeter is also a filter !!!!).

Is that just an overgeneralization meant "within reason" or do you actually mean that a 6db/oct filter at 100kHz will be audible???