They are both complete systems. The first is my own Ariel, as you can see designed back in 1994, and represents good but not exceptional direct-radiator performance.
The second is a Klipsch Chorus, very similar to the Klipsch Cornwall predecessor, and probably one of the better-measuring traditional high-efficiency Klipsch speakers. The time delays in the Klipschorn, La Scala, Belle Klipsch are all much worse, due to the long time delays and internal reflections of the folded bass horn. The Heresy, unlike the Cornwall and Chorus, is not designed for flat response, and is intentionally 2~5 dB hot in the mids and highs, with a JBL 4310/L100 type of balance.
The time graphs make the point that speakers that are described by audiophiles as "fast", "quick", or "detailed" may not have very good time performance. Rapid settling time is mostly correlated with low coloration, imaging that is free of "detenting" towards the cabinets, and good depth reproduction.
Altec, JBL, Electro-Voice, and Tannoy could have been working on time response since the early Seventies, but they chose not to, so the industry has been polarized between low-efficiency systems with fair-to-good time response and high-efficiency speakers with poor time response ever since. The impulse response published a while ago for the Summa was the best I've ever seen for any high-efficiency speaker, a huge accomplishment that the big players in the prosound business are doing their best to ignore - "Not Invented Here", as they say.
I'm dragging this thread off-topic, so I'll be moving back over to the usual place. This is Earl's thread, after all, and I appreciate the insight and information he's been posting here.
The second is a Klipsch Chorus, very similar to the Klipsch Cornwall predecessor, and probably one of the better-measuring traditional high-efficiency Klipsch speakers. The time delays in the Klipschorn, La Scala, Belle Klipsch are all much worse, due to the long time delays and internal reflections of the folded bass horn. The Heresy, unlike the Cornwall and Chorus, is not designed for flat response, and is intentionally 2~5 dB hot in the mids and highs, with a JBL 4310/L100 type of balance.
The time graphs make the point that speakers that are described by audiophiles as "fast", "quick", or "detailed" may not have very good time performance. Rapid settling time is mostly correlated with low coloration, imaging that is free of "detenting" towards the cabinets, and good depth reproduction.
Altec, JBL, Electro-Voice, and Tannoy could have been working on time response since the early Seventies, but they chose not to, so the industry has been polarized between low-efficiency systems with fair-to-good time response and high-efficiency speakers with poor time response ever since. The impulse response published a while ago for the Summa was the best I've ever seen for any high-efficiency speaker, a huge accomplishment that the big players in the prosound business are doing their best to ignore - "Not Invented Here", as they say.
I'm dragging this thread off-topic, so I'll be moving back over to the usual place. This is Earl's thread, after all, and I appreciate the insight and information he's been posting here.
If I remember correctly there is a Meyer sound 15" 2-way studio monitor (x-10?) with nice looking time domain measurements. It is VERY spendy though.
Chris
Chris
Lynn Olson said:
Verbal descriptions can be very misleading unless various recording are described in detail taking note of individual instruments. You can never tell what description is based on. But since you mention Klipsch, I have listed to their speakers many years ago, and yes, the impulse certainly looks like what they sound like.
BTW the imuplse for the Ariel seems inverted.
Attached is a two way design. The design was pretty much the same, but the tweeter alignment was different. The difference was very small such that I cannot remember.
But I really think that measurements are used to help determine how things can be improved. It would be necessary to look at various forms of data to determine how it might sound.
Attachments
Lynn Olson said:
I agree with you about Klipsch, they've never been known for particularly flat response, coherent summing or any of those kinds of attributes. The path lengths in their classic horn designs are all much too long and varied and having the midrange run through a folded horn is a problem too. But I think it is unfair to characterize most or even many high-efficiency speakers that way. In fact, Altec was probably the first to place focus on this aspect of loudspeaker design.
I think the phrase "time alignment" has different meanings to different people. Some think time alignment is physical alignment. Others think it means phase. Some realize those are two different things. Some consider a pretty wide range of phase/offset as being aligned, others less so.
I've talked to many audiophiles that believe "time aligned" speakers are those that have zero degrees phase shift across the entire audio band, so it looks like marketing spin has done a pretty good job on those folks. But I think most of us realize that is not possible.
Seems to me it is useful to focus on the things we can do something about, which is to strive for coherent summing over a range of frequencies and coverage angles. This has a bearing on the phase relationships between sound sources and their physical positons in relation to one another.
To me, the realizable goal to match phase through crossover points and to avoid abrupt shifts. Those are attainable goals, and I think most good builders strive for that. I do think all of the better designs achieve that goal on their forward axis, and the best ones do it over a wider coverage angle.
Wayne Parham said:
I agree - summing with good inter-driver phase angles is what matters, since that controls the vertical (and lateral) polar pattern, and is still important in MTM and coaxial designs. Good summing over a broad area is even more desirable, and is essential for stable imaging. (Home-theater Center speakers with side-by-side MTM layouts do not meet this requirement, and are grossly colored off-axis.)
The ability to reproduce a square wave (so-called linear-phase) is another matter entirely, and is not equivalent to a rapid decay-to-zero, which more about good driver design, good horn or waveguide design, and careful minimization of diffraction, reflection, and resonance. This is where a lot of of time-domain measurements are needed during the development process.
A little OT regarding diffraction on a XT25 tweeter(front plate has been modified with rounded edge near the diaghragm)
I have just plugged the tweeter mounting screw holes
It has been that before but fore maintenance I had removed the vax/modelleling clay, or what ever you call it
Well, I thought if it really did matter that much
I filled the holes on one tweeter and listened ... what a surprice
Now left and right speaker didnt play together at all, and the one which hadnt been modified yet provoked a nasty pressure in my ears
With the screw holes plugged the sound is much sweeter, smooth and coherent, so no I dont think enough can be done about removing diffractions
I have just plugged the tweeter mounting screw holes
It has been that before but fore maintenance I had removed the vax/modelleling clay, or what ever you call it
Well, I thought if it really did matter that much
I filled the holes on one tweeter and listened ... what a surprice
Now left and right speaker didnt play together at all, and the one which hadnt been modified yet provoked a nasty pressure in my ears
With the screw holes plugged the sound is much sweeter, smooth and coherent, so no I dont think enough can be done about removing diffractions
I'm afraid that 18sound don't make waveguides, that hasn't got a round unobstructed opening, I've used quite a few of these.
Lynn Olson said:
I would argue that it is of extreme importance to have vertical nulls spaced fairly widely. If a design puts too much emphasis horizontal directivity at the expense of vertical, you can have a situation where the good pattern is reduced to a thin layer centered on the forward axis. It doesn't matter how good the horizontal pattern is if the vertical null angle is very thin. Sitting in your easy chair might sound great, but slump just a little and you're in a null.
So to me, if the null angle is too small, the rest of the design doesn't really matter. It's not that vertical directivity is the most important thing, but a really thin angle is a deal breaker. What purpose is there in having a speaker with constant directivity in the horizontal plane if the vertical nulls are only 10 degrees from the forward axis?
I would also suggest that since there is no useful benefit for having HF at large vertical angles, it makes sense to choose a horn that limits the pattern as much as possible outside the null angle. The nulls mark the edge of coherent summing in the crossover region, so any sound outside that angle is undesirable. For that reason and for the reduction of ceiling and floor reflections, limited vertical angles are a good thing, in my opinion.
You said, "Home-theater Center speakers with side-by-side MTM layouts do not meet this requirement, and are grossly colored off-axis." I agree, and measurements bear that out. What you are saying, is sound off-axis outside the null angle is grossly colored. This applies side-to-side on horizontally stacked drivers or up and down on vertically stacked ones.
That's exactly what I'm talking about. Outside the nulls, the average amplitude in the crossover region dips, and there are also sharp notches at specific frequencies and angles. The sound output beyond this angle is undesirable. You don't want to place any listeners outside this angle, nor do you want any reflected energies sourced outside this angle. So in my opinion, there are two things that come from this: As much as possible, increase the null angle and as much as possible, reduce sound outside the null angle.
Whether you view this in the frequency domain or the time domain, it's really the same thing. Move up vertically, and the HF arrives first, down and the LF arrives first. So in the time domain, you see the impulse response vary considerably depending on the vertical angle. It's better if the null angle is larger. In the frequency domain, amplitude response is smooth within the null angle because summing is coherent. Outside the null angle, response becomes choppy because summing is incoherent. Again, it's better if the null angle is larger.
I cannot see any good reason to direct sound outside the null angle. I think it is useful to match the horn's vertical pattern to the null angle, which tends to maximize the usefulness of the horn's pattern, putting its energy where you want it instead of where you don't.
Hello Soongsc,
You remark is interesting in leading to the question:
for a multiways loudspeakers system what can be consider as the good polarity of the pulse reponse?
I think that the pulse response as shown by Lynn possess a good polarity.
This is related to the audibility of the absolute polarity and to the non linearity of our human hearing. For what I read and what I experienced myself, due to the internal ear distortion the good polarity is the one that lead to the least compression of the part of the pulse response corresponding to the frequency lower than few kilohertz. In other words polarity of frequency over few kilohertz is unimportant.
With the pulse response of the Ariel as given by Lynn the part of the pulse reponse corresponding to the low frequency is clearly rising when the part (the narrow peak) corresponding to the highest frequencies go down.
A good solution for having during listening the polarity right (it changes from record to record) is to use an inverting polarity switch. On records having very asymetrical waveshape (saxophone...) the good polarity generally lead to a higher perceived level in the bass.
Best regards from Sainte-Lucie-de-Tallano, Corsica
Jean-Michel Le Cléac'h
You remark is interesting in leading to the question:
for a multiways loudspeakers system what can be consider as the good polarity of the pulse reponse?
I think that the pulse response as shown by Lynn possess a good polarity.
This is related to the audibility of the absolute polarity and to the non linearity of our human hearing. For what I read and what I experienced myself, due to the internal ear distortion the good polarity is the one that lead to the least compression of the part of the pulse response corresponding to the frequency lower than few kilohertz. In other words polarity of frequency over few kilohertz is unimportant.
With the pulse response of the Ariel as given by Lynn the part of the pulse reponse corresponding to the low frequency is clearly rising when the part (the narrow peak) corresponding to the highest frequencies go down.
A good solution for having during listening the polarity right (it changes from record to record) is to use an inverting polarity switch. On records having very asymetrical waveshape (saxophone...) the good polarity generally lead to a higher perceived level in the bass.
Best regards from Sainte-Lucie-de-Tallano, Corsica
Jean-Michel Le Cléac'h
soongsc said:
I've been reading about the Gedlee approach to speakers and find it very interesting. I have not found anyone talking about how many ESP or Nathan/Abbey speakers they are using in their Home Theaters. I am also curious about placement in this application.
I am contemplating buying Gedlees book for Home Theater as it looks like an interesting read. Does the book describe how to use these speakers?
Cheers,
Robert
I am contemplating buying Gedlees book for Home Theater as it looks like an interesting read. Does the book describe how to use these speakers?
Cheers,
Robert
The book predates the ESP and Summa designs so it is devoid of any discussion on this topic. I am rewritting the book and adding some talk of speaker placement, but this project is falling behind. There are a few free chapters on my web site.
I use, and recommend three identical speakers across the front, or only two and the use of a phantom center. I would never recommend the use of a different center channel than the LR speakers as this never sounds right.
I DO NOT recommend any of my speakers for the suround channels as they are too directional. Surrounds should have a wide directivity to fill the space with sound coming from multiple directions. SInce they are already delayed early reflections are not an issue. I am going to do some suround designs one of these days, but for now I just recommend some good classical two-ways as the suround speakers.
Just be for-warned that I have blown all of the tweeters in my surounds as they simply can't keep up with the Summas.
I use, and recommend three identical speakers across the front, or only two and the use of a phantom center. I would never recommend the use of a different center channel than the LR speakers as this never sounds right.
I DO NOT recommend any of my speakers for the suround channels as they are too directional. Surrounds should have a wide directivity to fill the space with sound coming from multiple directions. SInce they are already delayed early reflections are not an issue. I am going to do some suround designs one of these days, but for now I just recommend some good classical two-ways as the suround speakers.
Just be for-warned that I have blown all of the tweeters in my surounds as they simply can't keep up with the Summas.
"Just be for-warned that I have blown all of the tweeters in my surounds as they simply can't keep up with the Summas."
How about B&C 8CXT coaxials, facing upward and outward to reflect the sound off of the ceiling and walls?
How about B&C 8CXT coaxials, facing upward and outward to reflect the sound off of the ceiling and walls?
Almost exactly my idea. Two of these at the ends of a tube - out of phase. No pressure at all inside the tube, lots of power. No sound on the nuetral axis (per Thomas Holman) Easy to make with sonotube. I already have the drivers, now all I need is some free time.
"No pressure at all inside the tube, lots of power."
But no bass, either, and double the driver $; curious how you'd address the former.
But no bass, either, and double the driver $; curious how you'd address the former.
And this would resemble a classic dipole design used 90deg off-axis, conceptually. But, depending on the length of the tube and no internal damping there would issues with standing waves, starting from the point where the tube length is half a wavelength.
- Klaus
- Klaus
gedlee said:
I couldn't agree more. The sound quality (and horizontal dispersion) of commercial home theater Center speakers is nothing short of a scandal. They can't do dialog, they can't do music, all they do is sit beneath the TV set and look cool (in the minds of the HT crowd).
A ridiculously simple test is playing a classic 1950's mono recording. This was something even a Zenith AM-FM table radio could do with reasonable fidelity. Play that same mono recording through a modern MTM-layout Center speaker and the sound will be murky, opaque, and the lyrics will be nearly unintelligible. If a mono recording doesn't sound good, what happens to the dialog on the DVD soundtrack, or music that is mixed to the center?
" But, depending on the length of the tube and no internal damping there would issues with standing waves,"
He didn't say he wouldn't use damping.
He didn't say he wouldn't use damping.
noah katz said:
I don't worry about bass in the surrounds. The surrounds are a fill speaker - thats all. They don't need highs or lows, they just need to not burn out🙂
Yes, I would dampen the tube which would be very effective since its a pure velocity field.