If that's directed at me I think you missed my point. If you don't do any compensation for the compression driver mounted on a CD Horn/Waveguide it's not going to sound right. The off axis will still be CD but with no comp it will sound terrible. A CD Horn/Waveguide will still be a CD device no matter what signal you choose to pump into it.
Rob🙂
Last edited:
Well My point was in the more theoretical, rather than practical "How its achieved". CD, as toole discusses it, is a dream. It's an unachievable goal that we can come close to, but never fully achieve. That still means it is the goal that we should strive for. We happen to do this through waveguides and horns. There are a variety of CD capable devices, and so I think issues like the compensation you speak of is really assumed as a part of CD. CD isn't CD just because the device is CD. The System has to be CD. If the crossover doesn't integrate into the woofer correctly, of which this compensation would be a key part, then it wont be CD. Some might argue it simply isn't CD at that frequency anymore, but then you could call any dome Tweeter CD above a certain point. I think that misses the point, and CD really must be thought of as part of the system.
Compensation or EQ is an inherent part of properly implementing a CD waveguide device into a speaker system. If you didn't properly do this, I would argue then that the measurements will indicate this behavior very clearly, and it won't behave as CD anymore about that range.
I mean, to get into the specifics of CD devices as they stand in the real world, nearly none of them are meant to be listened to on axis. JBL and Geddes both argue that they should be aggressively toed in because they intentionally neglect the onaxis response in order to ensure that the power response, or sum of the off axis is smooth. EQing the the on axis to be flat would just mess up the cd behavior.
To that point, I would guess that people who have found CD based speakers to sound bright, may feel that way because of improper setup. Which isn't to say that the JBL's aren't known to sound a bit bright, even Geddes designs, when the crossover HF compensation is at the lowest amount, is known to sound a bit bright to many people. To me this isn't an issue of the overall design per say, it's a known issue with the way CD reproduces the HF's, but it can be easily compensated in the crossover.
In the case my my Abbey's, it was compensated for by adjusting the value of the HF roll-off capacitor that bypasses the series resistor. As some might recall from my build thread, I replaced all of the caps with higher quality units, including those, and intentionally replaced them with a value to offer greater HF rolloff. As can be seen in my in room measurements I recently took, as well as those from the flat, I do have a very flat response above 2khz, with a level that is slightly lower than the response below 1khz. Thats even in light of a terrible room I've recently moved everything into with horrible reflections and diffraction sources. Again, I think that is just a good sign of what a proper CD speaker can do.
Compensation or EQ is an inherent part of properly implementing a CD waveguide device into a speaker system. If you didn't properly do this, I would argue then that the measurements will indicate this behavior very clearly, and it won't behave as CD anymore about that range.
I mean, to get into the specifics of CD devices as they stand in the real world, nearly none of them are meant to be listened to on axis. JBL and Geddes both argue that they should be aggressively toed in because they intentionally neglect the onaxis response in order to ensure that the power response, or sum of the off axis is smooth. EQing the the on axis to be flat would just mess up the cd behavior.
To that point, I would guess that people who have found CD based speakers to sound bright, may feel that way because of improper setup. Which isn't to say that the JBL's aren't known to sound a bit bright, even Geddes designs, when the crossover HF compensation is at the lowest amount, is known to sound a bit bright to many people. To me this isn't an issue of the overall design per say, it's a known issue with the way CD reproduces the HF's, but it can be easily compensated in the crossover.
In the case my my Abbey's, it was compensated for by adjusting the value of the HF roll-off capacitor that bypasses the series resistor. As some might recall from my build thread, I replaced all of the caps with higher quality units, including those, and intentionally replaced them with a value to offer greater HF rolloff. As can be seen in my in room measurements I recently took, as well as those from the flat, I do have a very flat response above 2khz, with a level that is slightly lower than the response below 1khz. Thats even in light of a terrible room I've recently moved everything into with horrible reflections and diffraction sources. Again, I think that is just a good sign of what a proper CD speaker can do.
Science and religion are not similar. For one to make decisions based on supporting evidence and for one to base decisions are personal belief w/o evidence is different. If you want to say you know better, just show me some evidence to support your claim and you win. I'm open minded. No need to be hostile.
Dan
5 part list?
So now 5 things on the list:
1. Controlled directivity (note: have been doing this for over 12 years)
2. Frequency balance based on Bark frequency intervals and not that silly 1/3 octave nonsense. (note: there is no other piece of electronics that sounds worse than a 1/3 octave analog EQ, always remove and discard!)
3. Phase coherence throughout the pass band with phase to remain always in the same polarity
4. Controlled environment. Reading this thread so many times the room is mentioned and so it belongs here. Indoors, I never have heard other than a speaker/room combination save for headphones. Every pro recording and playback studio considers the room an extremely important aspect. This clearly belongs.
I would like to ad this one:
5.Matched pairs of stereo speakers. This has not been mentioned however it is well established the formation of the stereo image is closely related the the ability of the playback system to create a null in the sound field. The depth of that null is exactly related to the matching of the two speakers of the stereo pair.
Correct this list or add to it, please!
Am getting really tired of who is best- is it possible to work on the top 5 or 7 or 9 things and come to some agreement on the items that belong on the list and how to measure or correlate them? Save what is number one for another time, possibly? I really believe the list is more useful as a whole than "identify the winner- Pink Floyd."
I know there was a request for studies: Here is a favorite paper, Passive crossover networks for noncoincident drivers by Siegfried H Linkwitz March 1977
As for other things about speaker design, slow sweep testing of the raw drivers and finished assemblies has shown a high percentage of the products in the market bark, belch, fart, rattle, burp, hiss, spit, screech, chug, especially buzz, and so on. Tested one pricey sub that went hopping across the floor at one frequency. To me this seems like a big problem. Another very expensive and highly touted driver made sounds of two skeletons copulating on a tin roof at one frequency. The manufacturer said that noise could not be heard with music. Hummmm....
=SUM
So now 5 things on the list:
1. Controlled directivity (note: have been doing this for over 12 years)
2. Frequency balance based on Bark frequency intervals and not that silly 1/3 octave nonsense. (note: there is no other piece of electronics that sounds worse than a 1/3 octave analog EQ, always remove and discard!)
3. Phase coherence throughout the pass band with phase to remain always in the same polarity
4. Controlled environment. Reading this thread so many times the room is mentioned and so it belongs here. Indoors, I never have heard other than a speaker/room combination save for headphones. Every pro recording and playback studio considers the room an extremely important aspect. This clearly belongs.
I would like to ad this one:
5.Matched pairs of stereo speakers. This has not been mentioned however it is well established the formation of the stereo image is closely related the the ability of the playback system to create a null in the sound field. The depth of that null is exactly related to the matching of the two speakers of the stereo pair.
Correct this list or add to it, please!
Am getting really tired of who is best- is it possible to work on the top 5 or 7 or 9 things and come to some agreement on the items that belong on the list and how to measure or correlate them? Save what is number one for another time, possibly? I really believe the list is more useful as a whole than "identify the winner- Pink Floyd."
I know there was a request for studies: Here is a favorite paper, Passive crossover networks for noncoincident drivers by Siegfried H Linkwitz March 1977
As for other things about speaker design, slow sweep testing of the raw drivers and finished assemblies has shown a high percentage of the products in the market bark, belch, fart, rattle, burp, hiss, spit, screech, chug, especially buzz, and so on. Tested one pricey sub that went hopping across the floor at one frequency. To me this seems like a big problem. Another very expensive and highly touted driver made sounds of two skeletons copulating on a tin roof at one frequency. The manufacturer said that noise could not be heard with music. Hummmm....
=SUM
Did you have difficulties with the thread I posted before this, or did you understand what I was getting at? (..because it's inter-related with what I'm about to go through.)
As far as this tangent, the real heart of the matter is over-rationalization - usually leaping to unfounded conclusions, sometimes resulting in absolutes. And while Earl may think this is "ugly", he has himself *many* times cautioned against this.
Ex.
Toole et al produce excellent research *largely* showing that for *most* listener's - uniform (relatively wide) horizontal dispersion is the most critical criteria for achieving good sound within the context domestic audio with regard to loudspeakers.
Does that mean that it's an absolute? NO. Rather it shows that it's far more probably than not, leaving room for exceptions that perhaps weren't included in various tests. Even then, the fact that it IS an *average* - is its own limit, there will always be "outliers".
Of course if you have "polarized" your view of that research, viewing the results as anywhere close to absolute - THEN you are moving *beyond* the fundamental context/limits of that research. At that time your "reasoning" is no longer within the bounds of *reason*. Often when this occurs ADDITIONAL "leaps" in logic occur as well, as often does the need to *share* with others.
These "additional leaps" are usually extrapolations to topics similar in nature, yet further removed from the research's "bounds".
Ex.
Loudspeaker measurements: specifically measuring the horizontal polar response of loudspeakers to produce better loudspeakers - while similar to, is NOT the subject of the research.
Polar measurements may, and likely will, help with this process, but they are NOT required. There are other methods available to achieve particular polar patterns, again: simulations, the use of manufacturer's data, common sense with crossover slopes and driver diameters when compared to freq., etc..
Now the degree of *accuracy* can certainly be improved in the design process by using off-axis measurements (in addition to on-axis), but it most certainly isn't a *requirement*.
That is to say: polar measurements are NOT the most important process in the design of a loudspeaker.
That "additional leap" was (B below):
A. uniform "wide" horizontal polar response is the most important criteria of a loudspeaker, therefor..
B. polar measurements to design those loudspeakers are the most important measurements in the design process.
Again neither A nor B is correct. A is more *probable*. B is simply incorrect.
As far as Science and Religion? When I used the term "faith-based", indeed I was comparing various claims as "religious" in nature. But what I was NOT doing was comparing Science to Religion. Instead I was comparing the misuse of Science to Religion. 😉
Throughout history, inventors just went out and did it without sufficient supporting evidence. As they did it, the supporting evidence gradually became available becaus they had faith in their judgement. This is not to say that all inventors are successful, but there is just a different spirit in them that drives them. You think we are trying to convince you?
Last edited:
This is the first time I have seen anyone mention this one. This has been on my SoundEasy wish list to show the Bark band tics for about a year now.
Ha ha. The idea that the group can make a great speaker... (it's all just opinions, the same with ideas). I'm not saying this is the same that with a special technical brainstorm or group work. Unless, there is someone that knows better in charge... (note: I'm not saying the same as when you read the post for the first time, from it, it seems a group would make a better work of it, or faster). I have some experience in team work and unless it's like a good team in place, it almost never works.😎
"The dog that barks doesn't bite", that's the problem, there's a lot of people that know better and keep quiet.
Scott, can you show an example of a speaker that was only modeled by common program available to the DIY guy that turned out a speaker that actually measured as it modeled?
The end result is what's most important. How you get there is up to you. If you can get there without measuring, by all means. Oh, please share how.
Soongsc, when you design a speaker that you believe sounds great and has a poor polar response, do a DBT in a room not specifically design for that speaker to sound vs. ones that do, and post result of it being preferred, I'll gladly eat my words about proven design. If that fails, try it again in a room specifically designed for that speaker. I'd be glad to see you pull that off as well. I sincerely mean that and I hope you can. Discovery is always a cool thing. Best of luck.
Until then, I'll stick with what's known as I'm not that smart.
Dan
The end result is what's most important. How you get there is up to you. If you can get there without measuring, by all means. Oh, please share how.
Soongsc, when you design a speaker that you believe sounds great and has a poor polar response, do a DBT in a room not specifically design for that speaker to sound vs. ones that do, and post result of it being preferred, I'll gladly eat my words about proven design. If that fails, try it again in a room specifically designed for that speaker. I'd be glad to see you pull that off as well. I sincerely mean that and I hope you can. Discovery is always a cool thing. Best of luck.
Until then, I'll stick with what's known as I'm not that smart.
Dan
Last edited:
Scott, you are arguing against a research you haven't familiarized yourself with?
Toole/Olive research showed a preference with high correlation for loudspeakers with:
Flat direct response/first arrival
smooth/smoothly sloping first reflection
smooth/smoothly sloping power response.
(also sound producers are worst judges of sound quality, no correlation with distorsion and so on)
if you take out bass response, this is the biggest issue influencing sound quality that has been tested, I know of no other research showing anything else.
Now what do you think, how to measure a loudspeaker that has the above listed qualities? Not with polar response, ideally taken from all-around the loudspeaker? Or do you think that those qualities are not worth as much and we should all prioritise those qualities that you deem worthy?
No not group think. Combined experience. Each individual taken from "the group" as interpreted by the individual reader from "the group" of individual posts.
Pretty much no one agrees on anything on DIYaudio (uhhuh
) which is why I am pressing for this list of "the top [N] most important items and not even putting them in order. This does somewhat answer the question "when, what, how, and why." I learn a lot by reading post to broaden perspective. Not to come to a decision based on group think.Sorry that was not clear. Hope this helps.
=SUM
Why not group together 1, 2 and 4 and make it one big 'loudspeaker on-and-off-axis response + resulting room interaction'? This includes everything from freq. balance to freely chosen directivity pattern/room absorbtion.
Also think that #3 can be omitted. If the phase between different non-coincident drivers are not 'correct', it will show in the polar measurement, no?
Mentioning requirement for matched pair is great, stereo doesn't create good illusion without it. Also worth mentioning a correctly steered/dedicated center channel, should have been there from the 1930s.
To complete the list, there should be a metric for dynamics. How to measure it is up to debate. I haven't seen power compression, as is measured today, correlated to the feeling of 'dynamic sound' - only as a way to gauge how loud a system gets in dB. I think dr. Geddes tried to come up with a metric that correlates short term termal mass to 'dynamic sound' - what happened to it?
Hmm, you seem to think it's about what I "deem worthy"..
My statements were logic based, not based on my own preference (..which, by the way, is in fact to take measurements off-axis - both vertical and horizontal)
So if you looking for what I deem worthy, yes - off-axis measurements are important.. but on-axis (..or at least the 0 degree axis relative to the listener), is much more important.
BTW.. what type of "target" dispersion would you characterize for Revel and Infinity loudspeakers?
"Get there without measuring"? Hmm, when did I make that claim? 🙄
Most of the designs measure as modeled.. or are you looking for some un-realistic degree of accuracy? Of course most designers don't publish their model, usually preferring to just display the measurements. However I believe that Zaph does, so if you looking at model vs. measurement, then by all means please peruse his designs and make your own judgment.
By the way I'm finding this "prove this to me" tiring. I've given it a "go" for your first "prove this otherwise".. and if you think I've failed then so be it.. 😱 It was entertaining (briefly). 🙂
Somehow I got the wrong impression that you thought polar response is not important?
Yes, direct arrival sound is very important, but why compromise the first reflection and power response? All 3 should be corrected and polar response is a good way to show it.
As for weighting between the 3, there should be a complex algoritm developed by now to show a single loudspeaker preference rating. Sadly, no matter how good the metric is, it doesn't come from an independent research facility, whicth affects credibility.
As for desired target dispersion, I truly beleive it depends on the listening room and recording technique - witch sadly is a moving target in music industry, dry recordings need wider dispersion and vice versa.
Dan, I'm not here to try and convince anyone. Nor do I claim to know the best. But what you are suggesting here is that a room specifically designed for a speaker will sound better. This I agree with, and is why I said the polar response is useful for determining how to mate with the room. However, to say the polar response is the most important measurement, is very misleading, and sounds like a saying marketing personnel would use.
PS. I have never been statisfied with any speaker for more than 3 months. One or two pairs of small speakers that I demonstrated was very interesting, I will post it a bit later.
Last edited:
Can you elaborate on that, for professional (factory development) or personal (relax/home use) reasons? Thanks.
I never tried to categorize, but most serious listening were done in dedicated listening rooms. It's just each speaker seem to have their strengths and weaknesses. Normally if I start to listen to one type of music, it might sound great, switch to another, it might sound terrible. Once you find something that doesn't sound right, you catch it immediately. Sometimes others hear things that I may have missed, but once I know the defficiency is there, I catch it in many different recording. Once I did a design that had one driver inverted to get a flat frequency response, but when I started to listen, the instruments where most sound were coming from the tweeter could never seem right when the bass sounded right. After that, I always go to designs where all drivers are connected in the same polarity. Listening to each system was different. One system could sound dry, another a bit harsh, some have image shifts depending on what note an instrument is played, etc.
Once I was asked by a friend to have a look at his newly furnished listening room. Not big, but seemed quite adequate for the size of his speakers. His main dissatisfaction was the sound was on the try side, and wondered whether he should add some diffusors. I looked at his speakers, which had diaphragms made of soft material, which based on my experience, would always produce dry sound. So I took a pair of prototype full range active speakers (designed for a target consumer price about 1/6th of the speaker he had) I had in my car, and we also sat down to listen, switching between both systems using the same source. The full range active speaker used metal diaphragm, and did not sound dry at all.
After that, the friend kept bugging me for three months for a pair of those active speakers.
After that, the friend kept bugging me for three months for a pair of those active speakers.
Floppy cones have seldom worked out well for me so the "result" of many test is heavy cones The 6.5" cone weighs 28 grams. I have some very good sounding soft cone speakers but add much power over 1 watt average and the sound is not so good. This heavy cone 6.5 does not do that rather, it just plays louder up to peaks of 110dB.
For daniel's question- polar pattern does not infer in phase response. Also polar pattern can be very nice yet the delay through the transducer can change quite a lot through the energy band. Think of an (2nd order) all pass filter- flat frequency response and shifting phase 180 degrees. Translate that to a transducer and you have great polar response and lousy phase response.
Also note I said "controlled directivity" and not whatever it is that is called constant directivity these days. I do not wish to repeat the earlier part of this thread about the "now" and the "then" definitions for beamwidth. Every direct radiator shows lobes in the radiation pattern. With #1 the key point is to keep the main lobe as a constant angle of radiation and constant axis versus frequency. How this relates to what the side lobes do is not what I was after. There could be a lot more to this. I was most interested in and concentrate on the main lobe of radiation and "attempt" to suppress side lobes.
As for #4 above- thought the room and speaker always play together design of the room is usually taken separate from the speaker. But the room is very important was this key point and not that the room should be designed by the speaker design and the speaker designed by the room. In practice the two are taken together at the "calibration" stage for pro studio installs. Any system not calibrated in the room as the final step really misses far to much.
To elaborate the importance of calibration, the manufactured product test the same in the test setup between many different matched pairs. Stick that speaker in a room and no it is not the same as the "factory curve" so requires calibration in the room every time to achieve great results.
Thanks for the input- please more!
=SUM
For daniel's question- polar pattern does not infer in phase response. Also polar pattern can be very nice yet the delay through the transducer can change quite a lot through the energy band. Think of an (2nd order) all pass filter- flat frequency response and shifting phase 180 degrees. Translate that to a transducer and you have great polar response and lousy phase response.
Also note I said "controlled directivity" and not whatever it is that is called constant directivity these days. I do not wish to repeat the earlier part of this thread about the "now" and the "then" definitions for beamwidth. Every direct radiator shows lobes in the radiation pattern. With #1 the key point is to keep the main lobe as a constant angle of radiation and constant axis versus frequency. How this relates to what the side lobes do is not what I was after. There could be a lot more to this. I was most interested in and concentrate on the main lobe of radiation and "attempt" to suppress side lobes.
As for #4 above- thought the room and speaker always play together design of the room is usually taken separate from the speaker. But the room is very important was this key point and not that the room should be designed by the speaker design and the speaker designed by the room. In practice the two are taken together at the "calibration" stage for pro studio installs. Any system not calibrated in the room as the final step really misses far to much.
To elaborate the importance of calibration, the manufactured product test the same in the test setup between many different matched pairs. Stick that speaker in a room and no it is not the same as the "factory curve" so requires calibration in the room every time to achieve great results.
Thanks for the input- please more!
=SUM
- Status
- Not open for further replies.