Member
Joined 2003
Voicing Constant Directivity
Earl,
The constant directivity waveguides I've been experimenting with over the past few months sometimes sound overly bright when voiced flat.
This is almost certainly a problem with the source material rather than the speaker. With good sources, for example AIX 24/96 DVD-A and many new SACD recordings, they sound so good they are simply stunningly. However, mediocre material sounds too "hot"...the speaker is clearly revealing flaws in the recording process. Unfortunately, mediocre sources are far too common. Do you strike a compromise when voicing CD (and if so, what is it) or do you let the sources stand on their own merit?
Paul
Count me in
Earl,
The constant directivity waveguides I've been experimenting with over the past few months sometimes sound overly bright when voiced flat.
This is almost certainly a problem with the source material rather than the speaker. With good sources, for example AIX 24/96 DVD-A and many new SACD recordings, they sound so good they are simply stunningly. However, mediocre material sounds too "hot"...the speaker is clearly revealing flaws in the recording process. Unfortunately, mediocre sources are far too common. Do you strike a compromise when voicing CD (and if so, what is it) or do you let the sources stand on their own merit?
Paul
I modified my simple oblate spheroid spreadsheed so you can set the initial throat angle using the Goal Seek function in Excel. No effort was made to make it "pretty" but it works.
http://www.crestviewcable.com/~catapult/Geddes.xls
http://www.crestviewcable.com/~catapult/Geddes.xls
gedlee said:
I am the guilty for the math part. I didn't take into account the slop of the exit of the CD because..well I didn't know I should have. Makes sense, but I didn't think of it.
I changed your formula in that spreadsheet a bit so that the function gave a roundover at the mouth (going asymptotic) while not affecting (too much) the profile at the throat and conical section.
Re: Voicing Constant Directivity
This is a big issue - one that we wrestle with all the time. We are so used to directional highs with very low power responses that the speakers can certainly sound too bright when voiced to flat. But exactly as you say, a good recording does not suffer ths problem. SO what is one to do? It is a delemma. We at Ai do voice down the highs but just a bit. We look to getting a +-2 dB sound field but we always shoot for the low end of the 2 dB. I have found that if the HF polar response field does not rise in power response - and believe me it can - then it is probably OK. But we almost always find that this requires a subtle rolloff of the response right on axis.
Remember I said that I do the EQ and crossover for the best polar and power response NOT the best axial response. A flat axial response will almost certainly sound too bright. We are just not used to hearing a flat power response to HF.
As to your comments about the waveguides sounding "stunning" - this has been my experience too. Dynamics that no small tweeter can even come close too and yet as clean and uncolored a response as I have ever heard from even the best tweeters.
Oh, by the way. This impression holds up very well even when the speakers are blaring sound at ear splitting levels in a disco. We have some disco installations and it is simply amazing how well these speakers hold up under extreme sound output levels. The music sounds like crap, but thats the way its supposed to sound!!
Paul W said:
Count me in
Earl,
The constant directivity waveguides I've been experimenting with over the past few months sometimes sound overly bright when voiced flat.
This is almost certainly a problem with the source material rather than the speaker. With good sources, for example AIX 24/96 DVD-A and many new SACD recordings, they sound so good they are simply stunningly. However, mediocre material sounds too "hot"...the speaker is clearly revealing flaws in the recording process. Unfortunately, mediocre sources are far too common. Do you strike a compromise when voicing CD (and if so, what is it) or do you let the sources stand on their own merit?
Paul
This is a big issue - one that we wrestle with all the time. We are so used to directional highs with very low power responses that the speakers can certainly sound too bright when voiced to flat. But exactly as you say, a good recording does not suffer ths problem. SO what is one to do? It is a delemma. We at Ai do voice down the highs but just a bit. We look to getting a +-2 dB sound field but we always shoot for the low end of the 2 dB. I have found that if the HF polar response field does not rise in power response - and believe me it can - then it is probably OK. But we almost always find that this requires a subtle rolloff of the response right on axis.
Remember I said that I do the EQ and crossover for the best polar and power response NOT the best axial response. A flat axial response will almost certainly sound too bright. We are just not used to hearing a flat power response to HF.
As to your comments about the waveguides sounding "stunning" - this has been my experience too. Dynamics that no small tweeter can even come close too and yet as clean and uncolored a response as I have ever heard from even the best tweeters.
Oh, by the way. This impression holds up very well even when the speakers are blaring sound at ear splitting levels in a disco. We have some disco installations and it is simply amazing how well these speakers hold up under extreme sound output levels. The music sounds like crap, but thats the way its supposed to sound!!
Josh
As a representative of the village idiots, Mongo says this is just way too much fun!
I'll bet a function can be added to the worksheet. One that will set the origin of the profile equal to a point on the curve that is tangent to a selected angle.
I haven't forgotten my commitment to you....just looking for the right profile.
As a representative of the village idiots, Mongo says this is just way too much fun!
I'll bet a function can be added to the worksheet. One that will set the origin of the profile equal to a point on the curve that is tangent to a selected angle.
I haven't forgotten my commitment to you....just looking for the right profile.
gedlee said:
Very smart to "aim" the lobe at the floor bounce. I like that idea. I do wonder about the horizontal response and axial response due to combing. You don't fine this a problem, or at least a high priority problem? A lot of fuss has been made over combing on and offaxis of MTM's and it's resulting coloration and I don't see how they would differ in principle with a large WG mated to large woofer at a conventional frequency.
BTW this is a great thread! I'm sure I'll it'll end up "favorited" by many by the time your done!
pooge said:Earl,
I think you would do well to rename the foam phase plug to something like "HOM absorber foam" or the like to distinguish from the phase plug in the compression driver. You are welcome to use my trademark "HOM absorber" if you like.
I've never called them a "phase plug". I called them a "refractive index absorber" in the patent, but thats a bit complex. Nothing really comes to mind.
augerpro said:
No we don't find this at all. When I get to Thailand I'll try to post the actual measurements as they clearly show that there are no issues like this. I don't have the measurements here with me now.
There is a single axial hole that is caused by the mouth diffraction, but nothing (horizontally) caused by the driver to waveguide spacing.
I am going to try the idea of foam on the mouth radius to cut down on the diffraction - see if it works.
catapult said:I modified my simple oblate spheroid spreadsheed so you can set the initial throat angle using the Goal Seek function in Excel. No effort was made to make it "pretty" but it works.
http://www.crestviewcable.com/~catapult/Geddes.xls
Unfortunately this doesn't quite do it. Thats because when you simply shave off the end, when the angle is correct the throat radius is no longer correct. You have to find a new throat radius and offset such that when the angle is correct the radius is also correct. I doubt that Excel has a simple function for this. It took me a couple of pages of math to get it right.
I'm curious whether it may be possible to satisfy your equation using graphical synthesis in a parametric CAD program. I took a graduate level Mech Eng class which used parametric CAD to generate advanced linkages that satisfied certain position, acceleration and velocity constraints. This may end up being an easier solution.
Lee
Lee
FWIW I recently had a 15" oblate spheroid WG CNC machined out of wood/plywood. Coverage angle is 120 degrees, throat is 1.4", and the WG profile worksheet is JoshV's found in
this post,
using for all I know Earl Geddes' formulas. I then added a LeCleac'h calculation for the flare and matched it by hand to the diameter necessary for the desired directivity lower cutoff of 800-1000 Hz. Drawings are in
this post.
Since I have no compression driver available I tested it with a dome tweeter. The exit of the nominal 1" tweeter happens to match the 1.4" throat (the surround takes up that much space). I made no attempt to correct for entry angle, it is at 0 degrees, but with the dome tweeter all bets are off anyway and the test is only very preliminary, to get a handle on lower directivity frequency limits. I did test it without the faceplate, so that the diaphragm is actually in the throat, but differences in response were small.
Measurements were done with WG mounted on a large baffle to prevent edge diffractions, and are shown
in this post
and the posts following it. I find polar response good, though with a dip on-axis that may or may no be significant. Directivity matches expectations. Comments welcome.
And one more question to Earl Geddes: most if not all 1" compression drivers I have seen were rated for a 1.6k crossover minimum. How did you manage to have the 1" driver go below 1000 Hz?
this post,
using for all I know Earl Geddes' formulas. I then added a LeCleac'h calculation for the flare and matched it by hand to the diameter necessary for the desired directivity lower cutoff of 800-1000 Hz. Drawings are in
this post.
Since I have no compression driver available I tested it with a dome tweeter. The exit of the nominal 1" tweeter happens to match the 1.4" throat (the surround takes up that much space). I made no attempt to correct for entry angle, it is at 0 degrees, but with the dome tweeter all bets are off anyway and the test is only very preliminary, to get a handle on lower directivity frequency limits. I did test it without the faceplate, so that the diaphragm is actually in the throat, but differences in response were small.
Measurements were done with WG mounted on a large baffle to prevent edge diffractions, and are shown
in this post
and the posts following it. I find polar response good, though with a dip on-axis that may or may no be significant. Directivity matches expectations. Comments welcome.
And one more question to Earl Geddes: most if not all 1" compression drivers I have seen were rated for a 1.6k crossover minimum. How did you manage to have the 1" driver go below 1000 Hz?
Member
Joined 2003
MBK said:
I'd love to see your polar measurements.
This rating for drivers is a safety factor put on by the manufacturer to keep "Dummy's" from overexcurting the diaphragm and hitting the phase plug - goodbye diaphragm! In a true CD waveguide you have low end to spare and it is no problem to take the driver down to 900 Hz or so. In my systems the waveguide is padded out at the lower end by at least 10 dB (more like 12 - 15 dB) That's already a lot of protection for the driver. We have not seen a problem yet. But there are a lot of "dummy's" out there who will drive these systems with 1000 watt amps - clipping! We will see.