Orions sound great because dipole?

[Patrick Bateman]

Have you discussed this with the guys at Harman? ;-) Maybe you should inform them about all the errors they have been making working on the Revel line of speakers.

It's not about on axis which can be made perfectly flat with even one driver. And as I mentioned earlier four way (four driver) may very well decrease thermal handling since with small drivers comes small voice coils. For many users it may be fine though. Horses for courses.


/Peter

Floyd Toole at Harman has written about this very subject. It's not like there's a divergence of opinion here. This is common knowledge. Hell, it's even in Dickason's Loudspeaker Cookbook, which is oriented towards hobbyists.

 

[Patrick Bateman]

How can you hear 2nd order harmonic distortion? Sounds very strange to me.



There are dome tweeters that are extremly clean and totally without sibiliance. But sure, I've heard a few that was not very fun to listen at. I have also heard horns/compression drivers and big paper drivers sounding unnatural and/or harsh. It's better IMO to speak of specific designs than painting with a broad brush and make general and sweeping statements that has no scientificall basis. There's to much dogma and ignorance as it is in the audio industry.


/Peter

Try listening to a proper wave guide for a few months. ALL tweeters will sound sibilant once you do. It doesn't have anything to do with the tweeters themselves; it has to do with the polar response. Which is what this thread is about.

Even the best dome tweeters suffer from beaming.

 

[gedlee]

This may be true, but I'm still interested in some nonlinear distortion data for his speakers, which I haven't yet found. Or how about some Gm graphs for the Summa?

What would be the point of posting a meaningless THD number? I don't take that data so there isn't any to post.

People really should read all the evidence about distortion.

Gm for a loudspeaker would be interesting although let me tell you it would be extremely low for all the reasons I state in all my publications on this topic. Loudspeakers cannot generate the kinds of nonlinearities that are audible. A post of Gm would also be meaningless out of context of what other loudspeakers do.

I have stated this many times and I'll say it again - current methods of evaluating nonlinear distortion are meaningless and I don't make meaningless measurements.

I find it laughable that everyone here seems to agree that "everyone knows THD is not meaningful" and yet they want to see the numbers. As the old joke goes "But the light is better here".

 

[gedlee]

John

If I'm not mistaken its 5:30 AM where you are? Whats up with that?

Also, you might find some of the discussion about thermal distortion on another thread interesting. There is a huge difference in how a dome tweeter and a compression driver behave thermally. I am becomeing convinced that this is a major factor in what I agree are serious sound quality differences between a waveguide and a classic piston tweeter.

 

[pelanj]

THD and box colorations do not matter. So then Patrick is a liar, when he states he can hear 2nd harmonic in normal program material?

And at low frequencies - hearing harmonics above test tone is nothing unusual. That does not matter? Then, does distortion matter in tube amps? Is speaker distortion "nice" for the ear? It is certainly for some guitarists...

Anyway, this debate is going in circles. Dr. Geddes does not provide measurements of THD and box colorations because he does not believe in that they matter - he believes, that they do not matter - therefore he does not measure. A perfect circle to me.

 

[peterbrorsson]

I'm maybe out of depth here, but... Speaking of standing waves in a box, I suppose the driver/s in Gedde's system have both stiff, strong cone and suspension. They are made for PA use with substantially higher pressures than compared with domestic levels.
Wouldn't these features make them less prone to be affected?

For sure, the wide rangers I have tried suffer from reflections in a box.

Peter

 

[auplater]

2nd harmonic.. weak argument

Anyone unfamiliar with 2nd harmonic audibility in reproduced or live music especially at lower frequencies, and their ubiquitous sound in all but the best systems, either hasn't listened to much of anything critically, is woefully untrained in music, or doesn't care about what they're listening to.

Hard to believe such a "technical" discussion is revolving around this issue.. as usual, disent about the "prevailing local opinion" is dismissed as irrelevant or uninformed.

 

[gedlee]

THD and box colorations do not matter. So then Patrick is a liar, when he states he can hear 2nd harmonic in normal program material?

And at low frequencies - hearing harmonics above test tone is nothing unusual. That does not matter? Then, does distortion matter in tube amps? Is speaker distortion "nice" for the ear? It is certainly for some guitarists...

I have only said that box coloration is not a factor in MY designs. Of course it can be a factor in poor designs. let's not mistate what I am saying here.

Patrick is not lying, but on the other hand I'm not so sure that what he hears is 2nd order harmonics either. What he is certain of is how clean the Summas are and how everything else sounds "cloudy" after listening to them. This is very very true! I can confirm that. But I would not jump to a conclusion about what this difference is due to. Since our work in nonlinearities I have been looking for the reasons that the Summas sound as good as they do. This lead me to HOMs, which is now becoming clear is a major factor in sound quality. I am also convinced that how the driver handles thermal loads is another major factor. Cabinet diffraction is a factor and in my cabinets, box coloration is not.

That musical instrument loudspeakers are highly nonlinear is a well known fact, but those speakers are part of a musical instrument NOT a musical reproducer - the two things are not the same, so don't get them mixed up. Nonlinearity in musical instruments is virtually always a good thing. Its what makes the tone.

The second order distortion in a tube amp or any part of the system is not a major factor. Tube amps are usually class A and as such they don;t suffer from the single most audible form of nonlinearity, crossover distortion.

 

[pelanj]

Well, I must say, that as a guitarist I can blindly recognize ear pleasant distortion (preamp and power tubes) to unpleasant transistor (or DSP-based) distortion. And I can hear distortion of guitar speakers - but they are actually designed to do that.

Argument with stiff PA cones seems valid enough to me. But it needs a proof by measurement. And still, the box adds some stored energy, does it not?

 

[pelanj]

Dr. Geddes, you write faster than I do Thanks for clarification - it makes your position more understandable (to a non-native speaker). I can now follow your arguments and together it makes sense.

Could you please post a link where HOM is explained in summary? There was a lot of discussion about that and this summary would be difficult to find.

 

[gedlee]

And still, the box adds some stored energy, does it not?

As a blanket statement of course all structures have stored energy, including OB!! The point is what is the level of that energy. One needs to consider this in any design - the enclosure should not be made of rubber or soft plastic like some of the pro stuff - but thats very different than saying that "all boxes are bad because of stored energy". I use Constrained Layer Damping in every box that I make in strategic parts where I have found it to be effective.

And as far as OB NOT having stored energy, what about that major source of energy, the mechanical vibrations of the driver? This gets into an OB box just as readily as it does a closed box. This idea that "all *closed* boxes are bad" is the ridiculous part of the discussion. That there are bad boxs is invariably true, just as true as the fact that good boxs can be design and made.

 

[Pan]

Floyd Toole at Harman has written about this very subject. It's not like there's a divergence of opinion here. This is common knowledge. Hell, it's even in Dickason's Loudspeaker Cookbook, which is oriented towards hobbyists.

Well that's my point.. I think you did not get me right there.

Try listening to a proper wave guide for a few months. ALL tweeters will sound sibilant once you do.

Not at all. I have listened many hours on dome tweeters in waveguides and there was no hint of sibiliance. I also have listned many many hours on non-waveguide dome tweeters which has no sibiliance at all either.

It doesn't have anything to do with the tweeters themselves;

Yes it has, distortion and stored energy can casue sibiliance.

it has to do with the polar response. Which is what this thread is about.

Not at all, a slight beaming towards high frequencies does not casue sibiliance.




/Peter

 

[Pan]

What would be the point of posting a meaningless THD number? I don't take that data so there isn't any to post.

Why don't post graphs instead? You know.. a picture says more than a thousand words.

People really should read all the evidence about distortion.

People do read and people do measure and study and the conclusion is that distortion is audible down to very low levels.

Gm for a loudspeaker would be interesting although let me tell you it would be extremely low for all the reasons I state in all my publications on this topic.

What does Gm stand for?

Loudspeakers cannot generate the kinds of nonlinearities that are audible.

You wanna make a bet?

I have stated this many times and I'll say it again - current methods of evaluating nonlinear distortion are meaningless and I don't make meaningless measurements.

Not at all. It's well investigated and documented what kind of levels of distortion that is audible at different SPL's. Loudspeakers typically produce levels way over the threeshold.

I find it laughable that everyone here seems to agree that "everyone knows THD is not meaningful" and yet they want to see the numbers. As the old joke goes "But the light is better here".

Is it really THD numbers that people want to see? I want to see graphs that show how severe the nonlinear distortion products are. I can judge that for myself. I want distortion sweeps and a couple of singel/dual/multitone FFT views.


/Peter

 

[cuibono]

We aren't talking about THD.

What would be the point of posting a meaningless THD number?

Okay, I'll say it again - I'm not interested in THD. How about some other representations of nonlinear distortion, like some spectra, or Gm data, which shouldn't be beneath you. Call it whatever you want, but if you're going to claim that your speaker is "clean", or that someone guessing Summas have high distortion "guessed wrong", the onus is on you to support your claim.



Really though, I thought the discussion on binaural recording was more interesting.

 

[Pan]

The second order distortion in a tube amp or any part of the system is not a major factor. Tube amps are usually class A and as such they don;t suffer from the single most audible form of nonlinearity, crossover distortion.

Saying that crossover distortion is the single most audible for of nonlinearity makes no sense at all. There are optimum biased A/B amps that switches over the crossover region constantly that is close to audibly transparent.. which can't be said about many of the class A tube amps that exists.

A tube amp (or any amp, class A or not) does not have "2nd order distortion", some amps have more or less only a 2nd order product on a single sine wave but music tend to be a bit more complex than that. All nonlinear systems (read all audio gear) produce sum and difference tones, IMD, which end up all over the place.

The same goes for speakers which is the most nonlinear devices in a well designed audio chain.


/Peter

 

[gedlee]

Okay, I'll say it again - I'm not interested in THD. How about some other representations of nonlinear distortion, like some spectra, or Gm data, which shouldn't be beneath you.

Really though, I thought the discussion on binaural recording was more interesting.

I thought that discussion interesting also. I really think that this sort of thing could be very useful. I'd much rather spend the time to do binaural recordings and post those than do nonlinear measurements.

As I have said before - I don't have any nonlinearity data - I can't post what I don't have!! And setting up to take this kind of data is difficult to impossible for me. I don't have an anechoic chamber and nonlinearity measurements are steady state so I can't gate them. That makes nearfield the only option. With nearfield I can really only do the woofer and waveguide seperately.

If you told me what it is that you would wanted to see and its something that I could do then I might make those measurements the next time I am setup to run data. I also don't have a permanent measurement room and with the holdays getting it (the living room) will be difficult. I lack a high powered amp to really push the drivers as I only test at listening levels -much lower than the speakers are capable of. Hence, you see, its kind of difficult for me to do something that I see as pointless - even if you see otherwise.

I COULD do a multitone near field at a typical SPL level. But what basis of comparison is there here? Where is the scale of audibility? Or are these just pictures of something which has no context to anything? I really see this as a meaningless exercize.

 

[john k...]

Earl,

You can interpret my 1/3 octave data as you like. The points I would make are that 1) for a given speaker position the result is independent of mic/speaker orientation 2) independent of mic position in the room, and 3) is flat through out the midrange. In fact, the data is also fundamentally unchanged if the speaker position is changed as well.

If I make that measurement for a small box 2-way with similarly flat on axis response (see figure below, blue = box, red = dipole) I see a consistent downward slope from 200 Hz through the midrange. Seems to me this has to be a result of DI approaching 1 as the frequency drops (i.e. baffle step effects). Also apparent is the dip and peak for the box in the crossover region due to a poor DI match between woofer and tweeter (centered around 2k Hz). A wave guide could improve the mid/tweeter transition.

An externally hosted image should be here but it was not working when we last tested it.

It seems to me that this is the biggest difference between a box speaker and a good dipole. A box speaker midrange will tend to have a higher level of energy in the reverberant field around the Schroeder frequency, rolling off there above, while the dipole has a more uniform midrange reverberant field. This seems consistent with the general, subjective observation that by comparison dipoles sound more open while box speakers sound warmer or heavier even though the direct sound is nominally flat for both. I know you have said repeatedly that the Summa maintains CD form about 700 Hz upward to some cut off. And you have stated that what happens below 700 Hz isn't of concern, or doesn't matter, or something of that nature. However, with all due respect, if still above the Schroeder frequency I would say it does matter, and saying it doesn't seems to me to be sweeping the issue under the rug because nothing can really be done about it with a direct radiator, box speaker. The baffle step is going to occur above the Schroeder frequency and the resulting change in DI for flat on axis response can only result in a reverberant field with excess low frequency energy. Some may prefer this sound, other not.

 

[john k...]

John

If I'm not mistaken its 5:30 AM where you are? Whats up with that?

Also, you might find some of the discussion about thermal distortion on another thread interesting. There is a huge difference in how a dome tweeter and a compression driver behave thermally. I am becomeing convinced that this is a major factor in what I agree are serious sound quality differences between a waveguide and a classic piston tweeter.

I assume this was addressed to me. I think 6:30 was probably the time.

You know, I have never really found a dome tweeter I truely liked. I use them, but they all suck. The best I have heard is the SS Ari Circ. I only wish it was 8 ohms and a lot less expensive. Still, you may see a compression tweeter in the ICTA. I've been leaning that way for some time for a number of reasons.

 

[cuibono]

As I have said before - I don't have any nonlinearity data - I can't post what I don't have!!

If you told me what it is that you would wanted to see and its something that I could do then I might make those measurements the next time I am setup to run data.

Very understandable - I don't actually put much weight into any sort of distortion data, unless there is really a large body of data, particularly between drivers. The number of variables in measurement of nonlinear distortion are so many that I prefer to watch trend that can be seen in groups of data.

Really, my contention was with claims of the "cleanness" of the Summa's and/or absence of distortion, followed by a lack of data. That's fine if you don't have any measurements of nonlinear distortion, but you do have your own metric, and you agreed the Summa's are "clean" and that you "can confirm that". I think we both value the scientific method, and so I would assume some evidence is in order.

Interestingly, near the end of your 2007 ALMA US presentation, you state that "In a blind test of some 15 subjects no one detected the difference between the Summa waveguide and a “Hi-End” direct
radiating soft dome device." That goes against what you've said above, and I would enjoy knowing more about this.

 

[Pan]

I don't have an anechoic chamber and nonlinearity measurements are steady state so I can't gate them. That makes nearfield the only option. With nearfield I can really only do the woofer and waveguide seperately.

Which is perfectly fine! Actually in room measurements (say at one meter and speakers + mic out from walls) will tell a lot even though there will be ups and downs in the curves. You can still see the general trend by smoothing with your own eyes. Not perfect but better than you'd expect. Also moving the mic closer to 25-50cm would mostly eliminate the room effect.

I lack a high powered amp to really push the drivers as I only test at listening levels -much lower than the speakers are capable of.

No need for high power amp. 1W is enough and since you have an amp that you use with them for listening, why would you need to measure at higher levels than you ever listen at?

I COULD do a multitone near field at a typical SPL level. But what basis of comparison is there here? Where is the scale of audibility? Or are these just pictures of something which has no context to anything? I really see this as a meaningless exercize.

Or just use sines at some intervals. The basis of comparison.. do you mean how other speaker typically measures?

Scale of audibility, If the harmonic distortion is at or below -80dB we can start to consider the nonlinearity being inaudible. At levels hihger than -60dB i think it's safe to consider the effects of the nonlinearity being audible.

The human hearing is such that we have the greatest sensitivity to distortion/spectral components at around 80dB SPL. Above or below that we can tolerate more distortion relatively. At 1W I want low distortion since most listening is at 70-90dB SPL and at higher levels it's important that the distortion does not increase all of a sudden.


/Peter