If you need any color correction or some easy ideas done hit me up.
You told me this elementary stuff before. But it's more that the off axis should roll off at a natural 6dB per octave slope?
You told me this elementary stuff before. But it's more that the off axis should roll off at a natural 6dB per octave slope?
Yea, I suppose you're right. Still, you can't do much better than that--well as far as I can think. Should you then want the tightest pattern you can achieve to minimize the room?😕 Seems logical to me right now. Of course if the pattern is too tight the sound will be more like headphones. Too loose and all sorts of room issues have a chance to rear their ugly heads.
Dan
Last edited:
😕Hmm. This makes me think. There's no sense in having pattern control in the bass sense your ears can't tell where it's coming from, but if high frequencies hit a lot of VERs, you'll mess everything up--image, timbre. So perhaps you want a speaker that gets progressively narrower as frequency increases. You'll want high frequencies to have a clear path to your ear as that's where timbre and image come from.😕 My brain hurts. I was just trying to do the opposite in this last build. I was trying not to excite room modes as to not have exaggerated bass.
Ugh, time for bed.
Dan
Ugh, time for bed.
Dan
Pattern control in the bass or modal region effects which, and by how much, modes are excited, for better or worse.
Well the ones that people say are audible anyway. I think the truth is there are more modes the higher you go it's just we are very skilled at ignoring hf modes.
Correct. Furthermore there's a transition zone where room dominance dimishes and the loudspeaker begins to dominate. This depends on room size.
I wonder how a one-size-fits-all directivity approach will meet different room/ speaker setups, how the consumer will ever be able to choose the right loudspeaker for his specific room and how problems within the transition region can be corrected. Earl's white paper will tell us that, right?
Best, Markus
Yes, that's true. Too bad I didn't understand that well at the start of my last build. Not that I necessarily understand it well now, but I am sure that's a good statement. I do know that trying to get deep bass from a dipole is an uphill battle. The deeper you go, the steeper the climb.😉 None the less, I do enjoy the mid/bass from these speakers, but I think they create as many problems as they fix.
Dan
You mean -6 dB in 45 degrees? dB/oct refers to a falloff in frequency and there shouldn't be more than 3 dB at 10 kHz, but with angle it should fall off much faster. This is exactly the point - that people need to get used to thinking in terms of spatial location changes as well as frequency response changes.
Yes there are more modes the higher you go. My understanding is that there are so many they are of little consequence to the response, not that people mask them.😕 I saw a graph of this somewhere on this site once. I believe it was in the Geddes multisub approach thread in the subwoofer forum.
Dan
Last edited:
My paper assumes that pattern control below a certain frequency is a different problem, and it is. At low enough frequencies its a room modal problem, NOT a loudspeaker directivity problem (small rooms of course) and even above the modal region its still not a significant audibility issue since we are not nearly as sensitive to room effects, VER and the like, at 200-500 Hz, than we are above 1 kHz. So the paper assume that polar control, in a small room, begins to become important above about 500 Hz. Argue with this if you like, but thats my position.
Yes, the DENSITY of modes goes up with the cube of the frequency - so the rate is very fast. The situation is like pixels in a picture, once the density is high enough more does not matter.
I guess you caught me thinking in 2-D there. I was more thinking about mastering tricks and what I seem to find on albums a lot. Where they will use a wide Q EQ to lift the upper mids. Since it is a wide Q not many people notice this as an EQ move.
I forget where you said it. But I was thinking that an ideal speaker would just be as small as possible and omni-directional. You said something about me being half right and that the off axis should follow a certain rolloff pattern. I guess all will be answered in the paper and I am just slowing you down by guessing/asking questions before you are done.
Guessing is good, means that you are thinking, as long as you don't "own" your guesses in the sense that if they are not correct that you are unwilling to let them go. It is very common herefor people to make guesses and then go to extremes to support them just to avoid being "wrong". Thats no way to proceed on anything. Guessing is great, its not being open minded enough to know when its a guess and when its not that is a problem.
I wonder how long people have been speculating and theorizing about this stuff. I was looking for something entirely different and found this.
Patent US2020166
Patent US2020166
what is the purpose of such thorough measurements? speaker is already built, drivers are chosen. there is not much that can be done to improve it, unless you are ready to tear this speaker apart 🙂
Wow - I could not disagree more!
Basically, to me, drivers are a commodity, I almost don't care whose they are as long as they meet some minimum requirements. But then the crossovers become the key thing as they are used to correct the drivers "issues" and they all have them, and in general they are all similar. But doing a crossover based on a single axial response is not going to get you a great sounding speaker. In fact it may well be the worst thing that you could do, its certainly NOT the best location on which to base a crossover.
Hence, doing the right crossover REQUIRES a complete set of data. Nothing less is going to allow you to "get it right".
I will agree with one thing about this comment. Once you select the systems archituecture you are pretty much limited in what you can achieve. For example, if you choose to use pistons in a two way system, then you can be guaranteed to NOT be able to get a smooth polar response. It simply cannot be done. So choosing the architecture first is essential - then the actual drivers don't make a whole lot of difference.
I have a lot more trouble finding a good purpose with that statement than I do finding a purpose in this thread.🙄 W/o measurements, we wouldn't have a clue as to what's wrong with this design. We wouldn't even know IF something was wrong with the design. In simple on axis sims it all looks good for the investment. If you were to follow this design lead, all you'd have to base the design principles quality on is words about how good these make me feel. Many people may find their way down the wrong path based on just that. I've learned more by actually doing something than I ever did by reading anything w/o logic or proof. I'm not the type to take a manufacturer or reviewer at their word w/o proof or w/o at least deep trust that they understand the principles involved and their not just trying to get my money. I actually think these types that buy stuff based on how it makes others feel are just silly.😱 Silly types can be happy people though, so whatever floats your boat. Ignorance can indeed be bliss. If the DIY community is watching and trying to learn something, these measurements are useful to all. We all make mistakes, it's important to learn something from them. If others can learn from my mistakes and the discussion ensuing as much as I have, they should be wise enough to be grateful they didn't have to make them. It's a lot of work!
I guess that's the purpose and you are welcome.😀
Dan
Last edited:
Bad example? Data from Genelec suggest that it can be done, e.g. Genelec 8020:
An externally hosted image should be here but it was not working when we last tested it.
Best, Markus
No doubt that measurements are very important.
I meant crossover overengineering for this particular speaker. Once you adjusted treble level, blue graph in first post, IMO that's it, it's the best sound you'll get from this speaker. What possibly can be done in crossover (full ranger + woofer) to improve it?
Nevertheless, it's very interesting and useful thread about measuring techniques.
I meant crossover overengineering for this particular speaker. Once you adjusted treble level, blue graph in first post, IMO that's it, it's the best sound you'll get from this speaker. What possibly can be done in crossover (full ranger + woofer) to improve it?
Nevertheless, it's very interesting and useful thread about measuring techniques.
- Status
- Not open for further replies.