Looks better than average, but suspicious. Way too smooth, but thats a 70 dB scale which makes everything look good, and then there is the very odd response arround 5 kHz which almost seems impossible. The response appears to drop 15 dB in 10 degrees, which is pretty much impossible.
Hi Earl
Are you sure you read the scale right?
Those plots are 30 degrees apart - from 0 to 180 degrees. The sudden drop in treble is going from 90 to 120 degrees (i.e. round the back of the box).
Doesn't look like a waveguide to me, given that the plots up to 90 degrees off-axis are within 8dB all the way to 10kHz. Maybe a 2 or 3-way with drivers mounted flush on the front of the box?
Regards - Godfrey
Are you sure you read the scale right?
Those plots are 30 degrees apart - from 0 to 180 degrees. The sudden drop in treble is going from 90 to 120 degrees (i.e. round the back of the box).
Doesn't look like a waveguide to me, given that the plots up to 90 degrees off-axis are within 8dB all the way to 10kHz. Maybe a 2 or 3-way with drivers mounted flush on the front of the box?
Regards - Godfrey
Last edited:
Those curves look like belonging to a well designed two-way box speaker, with probably some facets or round-overs at the edges - there's hardly any diffraction visible (not smoothed out I hope?). The decrease in directivity around 5 khz is normal for a dome tweeter in a baffle. You could reduce the on-axis response a bit in this range to compensate for the effect.
Last edited:
FYI Earl,
ARTA can produce an array of polar plots formats:
An externally hosted image should be here but it was not working when we last tested it.
The limits on angle are controlled by the data which the user loads and the response can be normalized as desired:
And for completeness, that is NOT one of my speakers. It is an example form the ARTA Help manual.
First off, thank you all for the comments.
a number of points:
1) this is a simmed response using visatons Boxsim program, not ACTUAL measurements. If the program is accurate, then i feel id be more than satisfied with the result.
2) on axis scale is 5dB per major division, 1 dB per minor division
3) Off-axis is 10dB per major, 2dB per minor...(This is what i dont like, as i agree, the coarser scale doesnt show the detail id like)
4) @ Keyser : Youre right, it is a 2way design, with 130mm bass driver and 20mm dome tweeter, baffle edges rounded to minimise diffraction, drivers flush mounted. no smoothing applied, unless it happens to be native to the simulation. very astute! I hoped it WAS well designed, we'll see when i build it!
5) the angular increments are 0-30-60-90-120-180 again here 15 degrees would be nice, but it is JUST a sim.
a number of points:
1) this is a simmed response using visatons Boxsim program, not ACTUAL measurements. If the program is accurate, then i feel id be more than satisfied with the result.
2) on axis scale is 5dB per major division, 1 dB per minor division
3) Off-axis is 10dB per major, 2dB per minor...(This is what i dont like, as i agree, the coarser scale doesnt show the detail id like)
4) @ Keyser : Youre right, it is a 2way design, with 130mm bass driver and 20mm dome tweeter, baffle edges rounded to minimise diffraction, drivers flush mounted. no smoothing applied, unless it happens to be native to the simulation. very astute! I hoped it WAS well designed, we'll see when i build it!
5) the angular increments are 0-30-60-90-120-180 again here 15 degrees would be nice, but it is JUST a sim.
Last edited:
And speaking of off-axis measurements - where is the pivot of that axis? Face for the speaker? Middle of the box?
The reason I ask I because I'd like to to some off-axis measurements of multi-cell horns. But the things are so large that I don't know where to put the pivot point of the 1M radius. Pegged to the front center of the horn it would seem too close to the horn if swung around the sides.
Gonna have to think about this one....
The reason I ask I because I'd like to to some off-axis measurements of multi-cell horns. But the things are so large that I don't know where to put the pivot point of the 1M radius. Pegged to the front center of the horn it would seem too close to the horn if swung around the sides.
Gonna have to think about this one....
An other question, how to use the polar response graph ?
If I can improve a loudspeaker from its spl/phase on axis, I don't have any idea how to interpret the polar response and how I can improve the overall response of a loudspeaker
Great thread!
The simple idea for me is to look for the nulls and make sure they are "tilted" out of the listening area. Im definitely no expert though so I stick to the simple things
Like even getting proper measurements. Going to try HOLMimpluse with my newly built QSC waveguide clones.
Isnt it going to be the same, if you set the speaker on the center of a "lazy suzy" the center and the middle are both on 0 and still would be in line at 15, 30, 45, etc.
You can buy turntables to support 1000 pounds for around $30. This is the one I bought a while ago:
McMaster-Carr
The easy answer is to measure farther away. Then small positioning errors become less important. Personally I would rotate multicell horns around a point back at the throat - roughly where the cells all converge. Then measure in the 3-5m range.
McMaster-Carr
The easy answer is to measure farther away. Then small positioning errors become less important. Personally I would rotate multicell horns around a point back at the throat - roughly where the cells all converge. Then measure in the 3-5m range.
Yes, rockler has them even cheaper
Low Profile Lazy Susans - Rockler Woodworking Tools
12" is $8.29 and supports 1000 pounds.
HD sells 12" round pieces of wood so I just use that and place markings on the wood.
Good ideas, guys - thanks!
Main problem for me is that I'd have to haul them to some big warehouse or a very quiet outdoor location.
John - I agree about using the cell join point as the pivot point. And about 3 meters away ought to do it.
Maybe I can do the horns only. That would be much easier. In HOLMImpuse it seems that gating needs to be under 6mS (1.7 meters) for good resolution down to about 200Hz. Is that right?
Main problem for me is that I'd have to haul them to some big warehouse or a very quiet outdoor location.
John - I agree about using the cell join point as the pivot point. And about 3 meters away ought to do it.
Maybe I can do the horns only. That would be much easier. In HOLMImpuse it seems that gating needs to be under 6mS (1.7 meters) for good resolution down to about 200Hz. Is that right?
Thank you rudolf, i was not aware of this as my german reading isnt very good
1/6th octave isnt too coarse, better than 1/3, but worse than none at all.
I recommend you read Dr. Floyd E. Toole's classic white papers on Loudspeakers and Rooms (esp part 2).
What we hear in a room is very very much dependent on the off axis response. The room affects this however and makes the importance different for different rooms. But a "good" room is not very dead and so the off-axis becomes very important.
My "turntable" is just two sheets of greased melamine coated wood, with a 1/2" pin as the axis. I can load 100 lb Summas on it and it truns just fine. I don't even like the bearing turntables because they aren't stable enough.
My "turntable" is just two sheets of greased melamine coated wood, with a 1/2" pin as the axis. I can load 100 lb Summas on it and it truns just fine. I don't even like the bearing turntables because they aren't stable enough.
Yes I understand but this appreciation is qualitative, not quantitative. If some people have some knowledge from their experiment, at my knowledge the interaction polar response - room is not a domain actually known ? and seems very difficult to put in equation.
If we have models to explain the behavior of a loudspeaker, these models are not enough precise to describe the actual response, in particular the polar response on the baffle, by extension in the room.
I cannot say if a polar response is good enough to have a good speaker. I can say one thing, the thinner is the baffle, better is the soundstage. Polar response seems very dependent on the geometry of the baffle.
Actually, it is quantitative. Toole & Olive have done the research. I'm sure Dr Geddes has to, but i haven't seen any papers.
dave
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.