A 'nude' B&G Neo3 is what most of the dipole people have been trying. Here are my measurements, a few iterations back on the crossover work. That's a Neo3-PDR with no baffle, 1800 LR4 XO to a 6.5" midrange on a 7.5" wide baffle.
The PDR version of this tweeter has felt strips that reduce the radiating surface width at higher frequencies, which I think accentuates the 7-8kHz problems on axis - the 'driver' becomes narrow compared to the 'baffle'. I've since added a notch filter to reduce the off-axis peaks, which of course made the on-axis curve worse. I've been planning to switch to the non-PDR Neo3 for some time now, that should help the on-axis measurements around that region.
Some of the non-CD'ness around the crossover region probably also comes from the electronic delay I have on the tweeter to align its impulse with the midrange. Better physical driver alignment should help that. Hopefully on my next build.
The PDR version of this tweeter has felt strips that reduce the radiating surface width at higher frequencies, which I think accentuates the 7-8kHz problems on axis - the 'driver' becomes narrow compared to the 'baffle'. I've since added a notch filter to reduce the off-axis peaks, which of course made the on-axis curve worse. I've been planning to switch to the non-PDR Neo3 for some time now, that should help the on-axis measurements around that region.
Some of the non-CD'ness around the crossover region probably also comes from the electronic delay I have on the tweeter to align its impulse with the midrange. Better physical driver alignment should help that. Hopefully on my next build.
Attachments
Last edited:
I've never seen the Emerald Physics speakers so I don't know. The measurements are mine done in my living room.
It would be nice if Brandon would post the measurements on the HPR-122i and PE WG w/ the Seleniums. Last time I was at the drivervault, he hadn't and his measurements of the Seleniums look different from mine, but that would be expected d/t the different implementation.
No offense, but I don't really put a lot of weight in subjective opinions w/o evidence and I'm sure you wouldn't either. I wish you had some for your implementation, and wouldn't be surprised if the Selenium was your worst. It should be for the cost. I also wouldn't be surprised if your impression of its sound wasn't influenced by its cost and your implementation.
How far does the HPR-122i go down to? My guess would be 1500Hz range based on previous experience. What I'd be more concerned with for an improvement is the throat and how it mates w/ the CD.
I'd certainly be interested in a test drive.
Thanks,
Dan
Putting an orchestra in a big box shaped room is artificial. I've never heard that Linkwitz is interested in any other acoustic than that of classic concert halls. That's a pretty narrow view on sound reproduction.
I was always wondering why he's not promoting binaural recordings and headphone playback.
Best, Markus
Nonsense. He simply wants the speakers to accurately reproduce the spatial cues that are on the recording. If you have to close your eyes to suspend disbelief and imagine you are sitting in a big hall, so be it. He's not trying to create a concert hall where none exists on the recording. On his demo/test CD he has a number of tracks that have nothing to do with concert halls or music. The goal is for Your speakers in Your room to make the tracks sound like the real thing (with your eyes closed if necessary).
7 - Leaving Concert Hall, 4:12
Walking from the hall through a narrow, low ceiling door passage into the large foyer and out to the street. All in a large group of people.
SFO '95. -6 dB
8 - Walking to car in garage, 4:57
Continued walk to parking garage, up the stairs to the parking level and to my car. Inside the car and driving off. Door slams with large very low frequency content.
SFO '95. - 6 dB and low cut filter.
9 - Parking garage stairwell, 0:44
Doors closing.
SFO '96. -6 dB and low cut filter.
12 - Ripping paper, 1:43
Outdoors in backyard about 60 cm from sphere microphone.
Corte Madera '04. -9 dB
13 - Ocean surf on open beach, 2:36
On a calm early morning to avoid wind noise.
Sea Ranch '04. -17 dB
14 - Ocean surf near cliff, 3:45
Other end of same beach, same time.
Sea Ranch '04. -17 dB
15 - Passenger train, 3:06
Passing diesel engine train from same location as Track 15.
Lincoln '93. -2 dB
16 - Trucking and Welding shop, 5:22
Door slam with large very low frequency content.
Lincoln '93. -3 dB and low-cut filter.
17 - Men's room, 2:46
Large room with open hallway.
SFO '96. -9 dB
Thanks for some "real" data. I rest my case. What are the angular spacings?
I'd love to plot some data like that in my new package. Any chance of that?
Why would the very same single strong lateral dipole reflection make recordings of all those very different aural spaces sound more realistic?
Does it work for you?
The angles were eyeballed, I'd say about 15 degrees each, and I wasn't being careful about rotating around a fixed point. Sorry, I didn't save the measurements. I'm using the demo version of ARTA so I just saved a screenshot.
I'll be changing the tweeter and maybe the midrange soon. I can try and save the individual measurements once I've worked up an initial crossover.
And I'm a complete novice at this, I'm sure better results can be achieved by someone with a better idea of what they're doing.
I'll be changing the tweeter and maybe the midrange soon. I can try and save the individual measurements once I've worked up an initial crossover.
And I'm a complete novice at this, I'm sure better results can be achieved by someone with a better idea of what they're doing.
Now Earl, be nice. 🙂
He said he was using the PDR version where the membrane is too narrow compared to the baffle width. Someone else posted curves for the non-PDR where the wider membrane decreases the problems at 8K and it looks pretty good until you start to get beaming above 10K due to membrane width. (I'm too lazy to search for the curves.)
The rule of thumb that's being worked out in these experiments (with theory to back it up) is the membrane/piston needs to be at least 1/2 the effective baffle width (including depth effects) if you want it to play wavelengths shorter than the effective baffle width. The idea is that driver directivity starts taking over as it loses directivity due to dipole cancellation. Of course, in a perfect world, you'd never let the driver play that high, but there aren't any commercially available tweeters small enough to get all the way up to 20K with perfect dipole behavior.
What 'strong lateral dipole reflection' is that? Dipoles have a null to the side.
Thanks, but I think that I understand the theory.
Why is it that whenever we see "real" data its always couched in "well this isn't right", or "I've already changed that". I can only judge by what I see and I have not seen anything that would make me change my mind about dipoles. Its more like the problems from 5kHz - 10 kHz isn't it?
As I said, I would love to get some "real" raw data and plot it in the same way that I plot all of my data. That makes it fair to everyone.
http://www.diyaudio.com/forums/multi-way/166411-measurements-when-what-how-why-33.html#post2192065
Also of interest:
An externally hosted image should be here but it was not working when we last tested it.
Fair enough. I don't have data that's current or ready enough to compare against your measurements. If/when I do, I can get back in this discussion.
I've done recording using modified earphones. It was the most immersive playback I have ever experienced when played back throught the recording MD with earphones. I did the listening right after the recording in the same hall, and was continously looking behind my back to find the person that was talking.
Okay, fine, I'll do the search for you. 🙂 Neo3 + a 6" mid 'naked' with the crossovers. 0-90 with 15 spacing. Green is a power average of all the curves. Yes its polar response is kinda ragged above 6K and really ragged above 10K but it's much better than the typical waveguide speaker below 1K. Add in the dipole woofers and it's going to have similar directivity as low as you want to go. Personally, I'm more concerned with the directivity of the stuff below 6K than the stuff above.
Edit: compare the 15 degree design axis to the green power response. Pretty dang similar.
http://www.diyaudio.com/forums/multi-way/160938-unbaffled-dipole.html
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Ipsilateral
Ipsilateral (Latin ipse; self/same): on the same side as another structure. Thus, the left arm is ipsilateral to the left leg.
I had to look it up.
Ipsilateral (Latin ipse; self/same): on the same side as another structure. Thus, the left arm is ipsilateral to the left leg.
I had to look it up.
Cool Soongsc! I've been doing it for many years and have the same experience. It's amazing stuff. I've done it just walking around town, doing the morning routine, concerts, etc... SL may be a bit crazy, but in a very cool way. My wife gave me the idea. She recorded her lectures that way. I thought she just had bad taste in glasses.😎
Dan
I'm trying to create a suite of polar maps to show people what they can expect from certain design approaches. I'm not looking to "compare" them to mine, but of course that will be a natural occurance. I have shown two or three very typical two-ways, a coax and a typical two way woofer and horn system. All are professional made and commercially available systems (or were) - some highly reputed. (I think it best to just refer to the design approach and not the manufacturer). What I do not have is a dipole system and try as I may I cannot get anyone to part with any data - which in and of itself is a curious event.
I guess that we see what we want to see.
An average set of curve about the power response does not mean CD.
Any chance of that data in raw form? Or better yet, the whole system?
Yup, I think SL's notion is playing that back through phones can give you an imaging reference. Then you try to build speakers and a room that will give you the same experience without having to wear the phones. He likes his dipole and omni speakers spread pretty wide and far from the walls but it doesn't really matter how you get it done as long as it works. Toole prefers multichannel speakers and surround processing -- lots of ways to skin the proverbial cat.
- Status
- Not open for further replies.