Geddes on Waveguides

soongsc said:
Recently a few friends listened to and measured a Thiel CS5. It sounded a bit screechy, and when we did the measurements, sure enough there was a tweeter breakup mode a bit higher than 20KHz. So it seems that probably the Everest is using some tweeter of rigid material.

The implicit assumption is that the high frequency breakup is causing the "screechy" sound. That possibility can't be ruled out, but from the data given, it's not established as causal either.

Sheldon
 
soongsc said:
So it seems that probably the Everest is using some tweeter of rigid material.
Beryllium.
CLS said:
If a waveguide was done right in an elliptical shape, it should have a curvy front because of the constant countour path lengths. So, I imagine a pair of lips slightly protruding forward on the narrower side.

Or, I think it'd be very close in shape to the new AH 500. (He calls it " Conical Radial Horn " though.... And the sharp edge around mouth is somewhat questionable considering the diffractions... )
$2274/pr. Wow.
 
Sheldon said:


The implicit assumption is that the high frequency breakup is causing the "screechy" sound. That possibility can't be ruled out, but from the data given, it's not established as causal either.

Sheldon
Well, when you reduce it and it sounds much better, then you know that's it. This is one reason why you see many tweeters that use a mask in front of metal domes. This is no news in the industry, just whether you ignor it or not.
 
soongsc said:

Well, when you reduce it and it sounds much better, then you know that's it. This is one reason why you see many tweeters that use a mask in front of metal domes. This is no news in the industry, just whether you ignor it or not.

I believe that most metal dome tweeters that use a "phase shield" as some call it is not used to reduce the ultra-sonic resonance peak, it's to help reduce the droop seen in all hard domed drivers at higher frequencies below the peak well within the audible range caused by phase differences due to dome profile at these frequencies. Unless you have a way to alter that peak alone in isolation of any other changes (DSP comes to mind as the only possible method), then any changes made are affecting more than this resonance, so Sheldon's point is relevant.

Dave
 
gedlee said:


But isn't the point that you don't need to!? The drivers Don't make a difference, its the system design that matters.


This was my point for certain. I was trying to say that JBL went a bit nuts with the technology, which ramped up the price, and that in the end, my thinking is that similar results can be achieved for less and through a simpler approach. To further that point, I mentioned that I had toyed with the idea of over-building the Geddes based system by adding complexity through more drivers, more crossover points, more complexity, etc. In the end, I decided it would likely not buy me anything, not for all that work, and that I should give the Geddes system a shot as is.
 
Almost without exception people tell me that I sell my designs too cheap. That people want to pay big bucks for systems like this so that they think that they are getting something valuable. I think that this is BS. Its possible to make money and make quality, and that people are not so stupid that they need to be ripped off before they feel like they got something valuable.

Marketing people are a bunch of crooks IMO. Their goal is too cheat you to within an inch of the law, and then, sometimes, not even within the law. The whole audio business is suffering because of these kinds of beliefs. How many people have been convinced that you need ellaborate electronics at high prices to get a quality sound? How did this happen? Because, it simply isn't true. (Flame war starts now!)
 
gedlee said:
Almost without exception people tell me that I sell my designs too cheap. That people want to pay big bucks for systems like this so that they think that they are getting something valuable. I think that this is BS. Its possible to make money and make quality, and that people are not so stupid that they need to be ripped off before they feel like they got something valuable.

Marketing people are a bunch of crooks IMO. Their goal is too cheat you to within an inch of the law, and then, sometimes, not even within the law. The whole audio business is suffering because of these kinds of beliefs. How many people have been convinced that you need elaborate electronics at high prices to get a quality sound? How did this happen? Because, it simply isn't true. (Flame war starts now!)

Well if you do raise your prices, just wait until I'm done purchasing things. You could make really high retail prices. Then people not only would feel like they are getting a really good deal, but they could tell people: "Oh yeah I have these 50,000 dollar speakers, you have to hear them!"

I agree that Marketing companies are crooks, and I think that in many cases they have broken the law. They tell us what to like, how much we want to spend, how much to consume, and they are very effective at it. The laws of supply and demand died out a long time ago, demand does not drive the market at all. All you have to do is look at something like the Bose wave radio to see an example of something for which their was really no major market for, and yet through amazing advertising campaigns, Bose managed to create a huge market for what is really nothing more than an over priced alarm clock/radio. If you add inflation into the cost of the old KLH and Advent table radios, they came to roughly 150 dollars today. If you add inflation into the Bose Wave Radio (from when it was first introduced), it comes to over 500 dollars.

Another market that has been completely ruined by marketing is the music industry. They can claim its from the "punk kids" stealing their music, but I don't buy cd's much anymore because their aren't new musicians worth buying. As much as I look at the 60's as a heyday for music, I also think it was the beginning of the end. While I know that musicians were recorded based on marketability as well as talent even before the 60's, I think it became the corporate credo of the music industry starting when groups like the Beatles took hold. The logical end extension of that model, 40 years down the road, has been that the musicians now have no talent, or virtually no talent, don't write their own music, and are famous for no other reason than they are easily marketed. To make matters worse here, I think there are some really talented musicians out there right now, who are making albums, but you would never know how talented they are from the music they are trying to sell us.

One of the greatest lessons I ever had in music was from one of my early music teachers. I asked him tot each me this Eric Clapton lead, I think it was the lead to Layla if I recall right. He told me, why would I want to teach you that, Eric Clapton never played the lead the same way twice, why should you. I don't think he meant that as a deep statement to live by. He was trying to teach me to improvise on my own, and not be afraid to explore the musical notes of a song the way the original artists had. However, I have taken it to heart, and in all my work and life, i try to be inspired by other people by doing things in the spirit of what inspired me, not by directly copying it.
 
soongsc said:
When you reduce the peak, normally the phase changes. So what's the difference?

The phase at issue is not the minimum-phase response, it's the phase delta between the signal at the tip vs. the signal at the rim, for the extreme portion. It's a result of the integrated response over the whole area of the dome. At high frequencies the extra distance due to the depth of the dome causes a droop in response.

This is OT in this thread. You may be interested in a page with tests at my site on the Dayton RS52 hard-dome midrange. The principle is the same. I'd suggest a new thread if you want to discuss this more.

Dave
 
gedlee said:
Almost without exception people tell me that I sell my designs too cheap. That people want to pay big bucks for systems like this so that they think that they are getting something valuable. I think that this is BS. Its possible to make money and make quality, and that people are not so stupid that they need to be ripped off before they feel like they got something valuable.

Marketing people are a bunch of crooks IMO. Their goal is too cheat you to within an inch of the law, and then, sometimes, not even within the law. The whole audio business is suffering because of these kinds of beliefs. How many people have been convinced that you need ellaborate electronics at high prices to get a quality sound? How did this happen? Because, it simply isn't true. (Flame war starts now!)


Earl, you are a good guy 🙂
 
dlr said:


The phase at issue is not the minimum-phase response, it's the phase delta between the signal at the tip vs. the signal at the rim, for the extreme portion. It's a result of the integrated response over the whole area of the dome. At high frequencies the extra distance due to the depth of the dome causes a droop in response.

This is OT in this thread. You may be interested in a page with tests at my site on the Dayton RS52 hard-dome midrange. The principle is the same. I'd suggest a new thread if you want to discuss this more.

Dave
If you are talking about the Dayton Reference Series Midrange RS52AN-8, I don't see significant reduction in the peak, whereas in the tweeters I measured there were significant reductions in the peaks. So unless you have used a driver with the design and then you changed it, we are talking about two different things.
 
soongsc said:

If you are talking about the Dayton Reference Series Midrange RS52AN-8, I don't see significant reduction in the peak, whereas in the tweeters I measured there were significant reductions in the peaks. So unless you have used a driver with the design and then you changed it, we are talking about two different things.

Of course, there is only one RS52 and I noted the number on the page, did you not read it? I'll say again, the shields are not used to reduce the peak, though they will have an effect on it. The purpose is to reduce the droop in response below the peak. Why do you ignore this point and insist on focusing only on the peak, the least important aspect? I understand Dr. Geddes focus on 10K and below, that's of prime importance.

Here are measurements for the RS28 hard-domed tweeter that shows the effect with/without a phase shield.

rs_t_noshield_shield_feltshield_spl.gif


In any case, this will be my last post on that topic in this thread. Start another one if you want to continue. It's way OT here.

Dave
 
I did not see this driver on your web site. In the graph you have posted, the yellow line looks quite well behaved to me. Basically when you start putting things in the wave path, it creates a more complicated wave pattern. This is normally what I like to avoid. Since I can't see the CSD, phase plot, and off-axis response, I really can't understand what you are trying to accomplish.
 
Dr. Geddes,

I have a general question about the CD behavior of the Oblate Spheroid. When running simulations in David McBean’s Hornresp program it appears that the CD coverage angle actually narrows to below the waveguide angle, and then widens back up to the waveguide angle at a higher frequency. Does this really happen with the Oblate Spheroid? See graphic below. Thanks.

l_44eec18069c1b718a4cef795271fea72.jpg
 
There is a narrowing of the response at the lower frequencies, but nothing as extreme as shown in your plot. The sharp changes in pattern with frequency shown in that curve are unrealistic. Things don't change that fast. And looking at just the -6 dB points is not very informative either. One needs to look at the total coverage pattern.
 
Waveguide crossover

How do you crossover a waveguide that is used close to its low end roll off?

For what I can understand, a compression driver + waveguide rolls off acoustically with 24 dB/oct at the low end. Adding even more roll off in the crossover would make it very steep, making it almost impossible to create equal roll off characteristics in the woofer network.

Do you use asymetrical crossover slopes or do you run the tweeter "fullrange", trusting that the resistive sensivity padding is enough protection?