Member
Joined 2003
Bjørn
Labeling the horn and test conditions of your polar plots would be helpful, but from previous posts we can gather the horns below are the AH425 on the left, and an OSWG (with or without the foam or crossover filter normally used is not stated).
The AH425 goes from an omnidirectional pattern at 500 Hz, (wider dispersion than a 15" speaker at 15") progressively narrowing to 40 degrees at 8000 Hz.
The on axis response at 8000 is around +11dB, around twice as loud (to our ears) as at 500 Hz. Over a fairly narrow beamwidth, (less than the width of a couch at a 2-3 meter distance) the sound character will change from rather dull to overly bright.
The beamwidth to stay within +/- 3 dB is less than 40 degrees.
The OSWG has a peak of 7.5 dB around 1600 Hz which will require equalization or crossover attenuation to eliminate. The response stays within a +/- 3dB beamwidth to just under 80 degrees, double that of the AH425.
If relegated to a single listening position, the smoothness of the AH425 response is appealing, and at one specific angle off axis provides flat response.
Given varied seating positions, the overall uniformity of the OSWG pattern would be my preference.
Different strokes for different folks.
Art
Attachments
All else being equal it appears that AH425 can be crossed comfortably at 800hz and possibly lower while OSWG 1.6-2k . I don't know any 12-15" sounding good in that range but what do I know
I don't choose speakers by their design philosophy.
Different design philosophies can produce good-sounding speakers.
As much as I like my current speakers, I do know that speakers can, at least potentially, sound much better. On top of that, speakers of higher efficiency will enable me to use much better-sounding power amplifier.
Since I learned what Lynn's sound preferences are, I have good reason to assume that I'd like the speakers he will come up with. Yes, in a way it's a bet of mine, yet I'm perfectly willing to make that bet.
+2.
Marco covered my issue well, that one CAN infer some things about the sound of a loudspeaker from understanding the design. A 1" dome crossed at 2kHz with just a cap just won't do realistic levels.
Back to the OSWG data, the updated charts are more in-line. The comparison of a simulation (the LeCleach are simulated I believe) and measured data is still an advantage for the LeCleach, but at least it's closer. I'm betting the WG was smaller than the 15" model, and not necessarily mounted on a baffle. Gedlee's repeatedly made the point that a larger WG is smoother, and that the baffle is key, it's like the OSWG version of the huge lecleach roundover. My own OSWG-ish W/Gs are very smooth indeed, but are as big as the ones in the Summa, and have care at the mouth termination that a free-standing oswg wouldn't.
The unidentified OSWG in post 10123 is about -6 dB at 1000 Hz, the acoustic crossover used could be lower than that and still sum flat in the crossover region.All else being equal it appears that AH425 can be crossed comfortably at 800hz and possibly lower while OSWG 1.6-2k . I don't know any 12-15" sounding good in that range but what do I know
I have heard some 15" that sound quite good in the 1000-1600 Hz range.
Earl presently offers three OSWG using 8", 10" and 12" woofers.
Lynn,
Thanks for your answer to my comment about the fire bottle amplifier preference you have. I must say that I grew up with many tube amplifiers and it wasn't very hard to get away from some of those after many years. I have had more than one model of Mac amplifier and have also used some of the Cary amplifiers and I must say that those particular models all have harmonic signatures that had much to do with their design goals. So I have not tried to pursue any tube equipment after that. I have seem to many high dollar tube amplifiers at CES shows that just don't cut it as far as I am concerned, but you are in a different class of designer. I imagine that most of what I have heard are in fact Williamson type of tube amps. Now don't even get me going on some of the single ended low power amplifiers, those are to me nothing but tone controls.
I am working on a commercial design for a small format speaker and tube amplification is just not in the cards, but I have learned enough to not want to go down the monolithic amplifier trail at this point. I will add that I have an old Mac tube tuner that has to be one of the best tuners that I have ever heard, and the channel strength can be downright amazing.
Thanks for your answer to my comment about the fire bottle amplifier preference you have. I must say that I grew up with many tube amplifiers and it wasn't very hard to get away from some of those after many years. I have had more than one model of Mac amplifier and have also used some of the Cary amplifiers and I must say that those particular models all have harmonic signatures that had much to do with their design goals. So I have not tried to pursue any tube equipment after that. I have seem to many high dollar tube amplifiers at CES shows that just don't cut it as far as I am concerned, but you are in a different class of designer. I imagine that most of what I have heard are in fact Williamson type of tube amps. Now don't even get me going on some of the single ended low power amplifiers, those are to me nothing but tone controls.
I am working on a commercial design for a small format speaker and tube amplification is just not in the cards, but I have learned enough to not want to go down the monolithic amplifier trail at this point. I will add that I have an old Mac tube tuner that has to be one of the best tuners that I have ever heard, and the channel strength can be downright amazing.
What would you call "small"? Is there a threshold, do you think, or does the line blur between small and large? There was some discussion about this a few years back but I don't remember the outcome - or if there was an outcome.
Hello Art
That is a pressure response curve. There is no way that the actual frequency response looks like that. That's why I asked Earl if the pressure response was related to the DI curve. Looks to me like the DI curve.
Not only that but those are SIM's where are the real measurements and polar measurements that are not normalized.
Rob
Last edited:
If it's in your home then its small (unless you have some kind of mansion. But I have been in some pretty big homes and none of them used the "big" rooms for audio. So I'd stick to my claim that if we are talking about home audio then its all small.)
Yes, exactly. I have updated my web page with the figure as well. I have tried to do the comparisons at my site as fair as possible, but not having the exact dimensions of your wave guides (since you have not wanted to give them to me), the OS simulations cannot possibly be exact replicas of any of your speaker configurations. And, the axisymmetric BEM can of course not simulate your rectangular boxes.
I do not consider myself responsible for how others use the data I have made publicly available. And... it may come as a surprise, but if a simulation of an OSWG does not behave exactly like one of your models, it is not necessarily an attempt to make the OSWG look bad. They may simply lack your experience and knowledge on how to make the best out of the OSWG.
Art, you can find captioned figures, details and comments on my web site: Bem Simulations. There are also simulations of several other LeCléac'h horns and OSWGs, in addition to other horn types, all of them (IIRC) with full polar maps. Note that these are all simulations, without driver or crossover, intended to show what the horn does. The frequency response will of course depend on the driver, but the directivity will not change much.
Regards,
Bjørn
Thanks, Lynn. It's OK, I know this OS vs LeCléac'h discussion can get heated sometimes. So perhaps it is preferable to link to my BEM page, where all the data is. All the charts are numbered and labeled, so they should be easy to refer to.
Regards,
Bjørn
Interesting viewpoint. What I'm always working towards is that the venue acoustic completely takes over the listening acoustic, and quite often this works especially well for live rock performances - the venue become enormous because there was enough pickup of crowd noises and other cues. At the very best performance levels the house or room you're in completely ceases to exist acoustically, the soundscape you're listening to is fully dominant.
So a large orchestra is no special problem, the walls in front of you cease to exist, the space expands to whatever size is encoded in the recording ...
I use a BD designs BD15 up to near 1200 and it's great with a big OSWG above.
JBL's 2206h is a 12" that actually sounds pretty good unfiltered- I recently sent a buddy a pair like that, with a CD horn loaded soft dome "super" tweeter (is it a supertweeter if XO is 2500?) and he's super-happy.
He's a lowther fan- I'm certainly not advocating that filterless is always the right way to go or even often the right way to go.
This is going back to cheaply knocked up loudspeaker enclosures like the Lockwood monitors the BBC dropped back in the seventies. Adding resonances is not hifi in any vernacular. aaaagh
Bjorn
I quite understand your position and I do not blame you for how other use your data.
I would say the simulations look reasonable given the pure sine nature. Most polars are shown 1/3 octave which washes away a lot of problems (most of them actually). I use critical band smoothing which is about 1/20th octave above 1 kHz. This is still going to smooth out things that you will see with a simulation at a fixed frequency. All in all I would say that your sims - when properly considered - show precisely what the issues are. When taken out of context they can show whatever the poster wants them to show. Its important that people see both sides.
By the way I tried your website but I could not load it. Do you show polar maps for the OSWG. Those are what I know best and could tell if the sims reflect reality or not.
I would have no issue with giving you the dimension of one of my waveguides. I don't remember you asking. I have a ton of very accurate data on them that could be used to validate a model.
I plan to take my measurements up one more notch to better quantify the performance. My hope is that with the next installment I can begin to sort out the HOMs. It's clear that something like that is occurring in your sims as the angular features are not at all constant - a key feature of HOMs. The question that is very hard to sort out is what are caused by the mouth and what are from down into the device. No far field measurement could ever sort that out.
Are you aware that I can reconstruct the velocity at the mouth by a reverse (ala holographic) technique from measured data. It works down to about a 1/2 wavelength now because the boundary conditions are not precise. My plan is to use a large flat baffle which should allow for about a 1/4 wavelength resolution and then possible a reconstruction of the waves inside the device from that. A lot of work but the only way to get to the nitty-gritty of these things.
The only way that this is possible is for the room to be completely dead - an anechoic chamber. Otherwise the small room queues will always dominate because they are shorter (longer is never a problem.) But then this leaves a completely lifeless local acoustic and only music with a lot of venue acoustic will sound even remotely correct - but then the acoustic is still coming from the wrong place. So this approach simply kills the kind of music that I prefer and that is simply 2 channel studio work. So a dead local acoustic is simply not an option for me. I make the local acoustic as live as possible.
Rock concerts don't have any venue acoustic since the feed is generally taken directly from the performers and not the room. Played loud enough, the local acoustic kind of goes away and the playback realism becomes quite convincing. Its gone at lower levels, probably because the expectation is for loud.
I would guess that a significant transition would occur once you could push the earliest reflections beyond 10 ms. Things would start to change significantly once that happened. But with reflections < 10 ms., even suppressed and the queues are all wrong. And no recording can remove those early reflections no matter what the engineers do - the brain knows. (Actually it knows too much! That's always a problem!)