This is precisely why I do not do shows anymore. A speaker like the Summa is never going to grab someone attention right off, so showing them only tended to degrade their perception. To sell a speaker at shows they have to be tailored to that type of venue - make the sale, then who cares if they are fatiguing once you get them home. All the better, the customer grows to dislike the speakers so he goes back to a show to find new ones and the process starts all over again a new. This is why few people who buy my speakers sell them and buy something else, although several have upgraded models. But no one ever bought them after hearing them at a show.
I have had my Summas for about ten years now and I would not even think of getting anything else. Fatiguing - not in the least and that is just about everything to me.
Is it all diffraction? Maybe yes, maybe no, but I doubt it. Is diffraction major aspect of the problem, I would say certainly. But lack of audible compression (dynamics) and smoothness of response are also factors. What are the weightings of these attributes? That would be a great project for someone to do. Almost impossible, but very very interesting.
The only controlled subjective study on diffraction that I know of is the one that Lidia and I did about ten years ago. In that study I simulated a "diffraction like" signal and the listeners (dozens, blind) compared the diffraction contaminated signal to the uncontaminated one. By far the most interesting result was that the diffraction was more audible at higher SPL levels than lower ones. This means that the effect should be more audible at higher SPLs than lower SPLs. This was completely consistent with Brian Moore's result on the audibility of non-minimum phase where it becomes more audible at higher SPLs. This is in stark contrast to frequency masking which increases with level - non-minimum phase errors become unmasked at higher SPLs. How odd!
So if your speakers seem to "loose it" at high SPLs it could be diffraction and it could be nonlinearity, although the later tends to be masked for lower orders so the higher orders would have to be increasing faster than the masking - which can happen, clipping for example. But a well designed driver should not do this. A compression driver almost certainly never does this as our study of the perception of nonlinearity in compression drivers proved. A horn, on the other hand can have substantial amounts of HOMs which are non-minimum phase and hence become more audible at higher SPLs. They will sound like what people expect THD to sound like, but it's not THD.
It's all very complex, but it is also not completely unknown.
When it comes to audio, there aren't many 'free lunches', but it seems like this is one of them. On the Diyma thread there isn't a single post where someone tried a sphere and found that it sounded worse.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Here's some food for thought:
Bose sells a lot of speakers. I know that no one on this forum would defend them, but there's a lot of regular folks that thinks that Bose is HiFi.
So here's a possibility:
Do the small cubes that Bose use combat diffraction?
For instance, the frequency response measurement above demonstrates that the on-axis response extends to 20khz. But the driver will start to beam at 6khz, due it's diameter. Could the beaming, combined with the small baffle, prevent the radiation to the sides? And would that reduce diffraction, which in turn could give the lowly Bose speakers the 'listenability' that is a hallmark of low diffraction designs?
Of course the obvious way to reduce diffraction is to use a roundover, but you can also do it by a careful combination of driver beaming and baffle width.
For instance, a 2" driver will beam at 6750hz. So put it in a 2" baffle, like Bose does, and everything above 6,750hz will 'avoid' the baffle edges due to beaming. As we go lower in frequency, the wavelengths basically won't "see" the enclosure, because they dwarf it. By the time we get down to 1700hz, the wavelengths are completely swamping the enclosure size; it's as if the enclosure simply isn't there.
I would need to do some sims to confirm, but it appears that diffraction would only be an issue in the two octaves from 1700hz to 6,750hz, and the effects would get less and less as you get closer to 1700hz, as the wavelengths are beginning to swamp the enclosure's size.
Last edited:
A compression driver on damped plane wave tube does not perform the same as compression driver on horn/waveguide.
Nobody listens to music with plane wave tube.
Unloaded driver in horn at low frequency does just what direct radiator does: moves more, in non-linear fashion, generating low frequency IM that dirties up response.
Not convinced
I'm just not buying all this diffraction business.
Have you read Chapter 9 of Toole's book? He focuses on diffraction and comb filtering from reflections, like those you might see from a baffle or from the room surfaces. He examines what happens when a reflection is delayed 1 ms (1.135 ft) like you might see from a baffle edge. It is one of those cases that because we think we can measure it, do something about it, and then see an improvement, we think we are improving it.
First, the ear's equivalent rectangular bandwidth will prevent you from hearing most comb filter effects. Second, that puny radius on the baffle edge sure looks good, but it is too small for any real effect. Third, by your admission you say that the Vandersteen's have good imaging. And they certainly don't appear to have any rounded edges. There are countless examples of mini-monitors and other speakers with sharp edges that have stellar imaging.
At the other end of the scale, I have heard big front horns with sharp edges, altec horns with sharp edges that produce extraordinary imaging.
These are just anecdotal claims. I'd hazard that the difference lies in the crossover design, frequency response, and smoothness of the off-axis response in the Summa. Check that first before chalking it up to diffraction.
Another thing is that diffraction from a rectangular baffle changes at different points. You will not see the effects of diffraction in a power response, i.e., it cannot be categorized as a resonance. This is completely different on a round horn, where the source is equidistant from the baffle edge.
I'm just not buying all this diffraction business.
Have you read Chapter 9 of Toole's book? He focuses on diffraction and comb filtering from reflections, like those you might see from a baffle or from the room surfaces. He examines what happens when a reflection is delayed 1 ms (1.135 ft) like you might see from a baffle edge. It is one of those cases that because we think we can measure it, do something about it, and then see an improvement, we think we are improving it.
First, the ear's equivalent rectangular bandwidth will prevent you from hearing most comb filter effects. Second, that puny radius on the baffle edge sure looks good, but it is too small for any real effect. Third, by your admission you say that the Vandersteen's have good imaging. And they certainly don't appear to have any rounded edges. There are countless examples of mini-monitors and other speakers with sharp edges that have stellar imaging.
At the other end of the scale, I have heard big front horns with sharp edges, altec horns with sharp edges that produce extraordinary imaging.
These are just anecdotal claims. I'd hazard that the difference lies in the crossover design, frequency response, and smoothness of the off-axis response in the Summa. Check that first before chalking it up to diffraction.
Another thing is that diffraction from a rectangular baffle changes at different points. You will not see the effects of diffraction in a power response, i.e., it cannot be categorized as a resonance. This is completely different on a round horn, where the source is equidistant from the baffle edge.
Last edited:
Another 'data point' might be studio monitors.
It seems logical that people in studios spend a lot of time listening to loudspeakers.
An externally hosted image should be here but it was not working when we last tested it.
And I see a lot of B&Ws and Dunlavys in studios. With diffraction reducing enclosures. This pic is from Abbey Road.
An externally hosted image should be here but it was not working when we last tested it.
Here's a set of Dunlavys for sale on CL in LA right now
If the 'audiophile' speakers were ideal, I'd think you'd see more Wilson and JMLab in studios, but they don't seem to be as prevalent as B&W.
Could it just as well be that B&W does a much better job of marketing? Could it also be that you don't have all the data and that the market of studio monitors is mainly occupied by companies like Genelec and even Behringer?
It's very easy to run away with conclusions... perhaps you need to read Kahneman too.
It's very easy to run away with conclusions... perhaps you need to read Kahneman too.
When I met Geddes at the RMAF I noticed that we both drove the same car.
And he mentioned how this particular car brand was very popular with engineers in the automobile industry.
That's probably a fairly valid way to choose a car. Don't ask your friends, or your partner, or the internet. Ask a mechanic. Or someone that works in the automotive industry. They'll probably have some of the best data on what cars are a wise investment.
By the same token, I'm more likely to trust the engineering decisions of a recording studio.
I worked in a half dozen recording studios in Western Canada for 10 years as a mix engineer and every control room had these bridge monitors: The Yamaha NS10 Story as did virtually everyone else in the biz at the time.
For large format these were very popular: 1977 UREI 813 Studio Monitors
My poor ears
For large format these were very popular: 1977 UREI 813 Studio Monitors
My poor ears
If you assume brick wall cutoff then quater wave length is fine. In reality, you still have cut off slop concerns.
I quite agree that more drivers can make a unity horn better. I have actually done the opposite to create wider dispersion in a single driver by causing it to break into an array of sources, which also helps dissipate energy on the cone faster. But that happens only in a very limited high frequency range. The fact that you can listen to the Summas longer than the Synergy says something about the trade off in design.
I still think one can understand more thoroughly when you do a series of measurements inside and outside the horn. I can repost one series that I did if you can post what you did with the Synery. Then it would make the discussion more clear. Note that early in this thread we talked about how to spot HOMs, well, I now don't refer to the terminology HOM, but rather just plainly try to figure out what to fix. Just not into all the academic terminology.
Yes, I do disagree. Because your claim is based on purely anecdotal information. You have no data on whether your claim is universally true. We cannot draw any conclusion on diffraction, which is a small small feature of those speakers performance anyway, if we don't know for sure if B&Ws are the most popular studio speakers.
Earl promotes the fact that his speakers don't have any early reflections. Also, why would it be imperceptible if the second wave is less delayed compared to more delayed?
Again, suggest you read Chapter 9 of Toole's book. There are plenty of wide baffle speakers out there with good imaging. The baffles of your Summas are pretty wide too.
How is diffraction reduced by making a horn larger?
Let's do the thought experiment of taking the baffle (or other edge) between tiny and infinity. Tiny - no diffraction. Infinite - no diffraction.
In between, if diffraction is an issue, then it's likely to be more or less audible at some frequencies than at others, and would probably depend on frequency range of the source and the distance between the main wave and the diffracted signal.
Sheldon
Hi soongsc,
The ability to specify multiple drivers was added to the Wavefront Simulator in late July.
http://www.diyaudio.com/forums/subwoofers/119854-hornresp-365.html#post3579785
Other features were also added to the Wavefront Simulator around the same time.
http://www.diyaudio.com/forums/multi-way/217968-my-summer-ruined-2.html#post3569570
Kind regards,
David
Lots of good stuff here.
I read your posts with interest PB, many good ideas. But I also think the "B&W is popular in studios" is a marketing thing. Abbey Road has got them, so they must be good, right? Good marketing strategy there from B&W. Abbey Road has also got other speakers, but we don't see pictures of them posted as often on the internet.
I read your posts with interest PB, many good ideas. But I also think the "B&W is popular in studios" is a marketing thing. Abbey Road has got them, so they must be good, right? Good marketing strategy there from B&W. Abbey Road has also got other speakers, but we don't see pictures of them posted as often on the internet.
That diffraction is audible is pretty clear in my mind, but I am not sure it is responsible for everything. Here are the points where I think that you are wrong.
1) ERB does not mean that we can't here comb filters. Maybe we can't but you can't use that argument because it does not apply. Consider a swept tone for example. You will hear the peak and dip as the tone seeps through the comb.
2) Toole never really looks at diffraction, reflections yes, but not diffraction and they are different things. But it is precisely the audibility of reflections in Toole that I (and David Greisinger) object to. I don't think that his tests completely nail down the audibility of reflections or more importantly diffraction.
3) The radius on a Summa is 2 inches, which is not "puny". The radius at the waveguides mouth is 4 inches. These are large radi when compared to the wavelengths involved and will substantially reduce diffraction from these edges.
4) most diffraction from a loudspeaker occurs at less than 1 ms. where the ear is in the region known as "summing localization" and results above a few ms. do not apply to these very short time delays.
5) Diffraction effects will be more audible at higher SPLs. Toole never did tests as a function of SPL to determine if the audibility changes with level or not. It may well be that all his tests were at very low SPLs where the audibility is at its minimum.
6) Your claims are anecdotal as well. The only scientific study is the one that I quoted and it does not conclude that diffraction is not audible - quite the opposite. Either exclude anecdotal data or include it, but don't mix your rejection or acceptance to suite your argument. Just remember that when you accept anecdotal data that the discussion can never conclude anything because all anyone has to say is: "Well it sounds that way to me."
Last edited:
I have a new car now - a Honda Civic SI coupe - best low priced sports car on the market - no contest. Incredible performance, 30+ MPG. What's not to like!?