salas said:
The OS contour is conical at the mouth. All waveguides are conical at the mouth just as all wavefronts are spherical far enough from any finite sized source.
But in an OS there are not waves bouncing across the device as you show. The wave starts out flat and is slowely curved to spherical by the time it reaches the mouth. The foam plug is as you show.
I'm not sure that this topic is appropriate for "Beyond the Arial" and if we are going to continue we should move it to its own thread.
gedlee said:
It was all foam I wanted to sketch, not bouncing waves. But I sketch awfully... Yes I will ask about WGs in another thread. See ya!
Round horns verses rectangular horns
I think the part Norris is referring to is the vertical coverage angle, not the throat features. I don't mean to speak for him, but he and I and Duke LeJeune have been discussing this lately over on AudioRoundTable.com.
I've been making DI matched loudspeakers for well over a decade, and like the post Norris referenced shows, I modeled my three Pi and four Pi loudspeaker designs somewhat on the principal features of the 1983 Keele/Eargle/Smith paper. Different horns and different drivers, but the same basic design features are used.
Using an asymmetrical flare, the loudspeaker can provide a pattern that's wide but not very tall. Drivers positioned one above the other cause nulls to form above and below the speaker. If the layout places those nulls at the outside edge of the horn's vertical pattern, they tend to punctuate pattern cutoff. In a sense, the nulls actually become an asset because they help the horn with pattern control at the low end of its range. Since the nulls limit the uniform radiation angle anyway, it's nice to use them to your advantage.
The main point of the post Norris was referring to is vertical directivity, not waveguide or horn features. You can use a prolate spheroidal flare if you like the way the OS/PS throat is radiused. The point was to limit the vertical angle to the position of the off-axis nulls.
gedlee said:
I think the part Norris is referring to is the vertical coverage angle, not the throat features. I don't mean to speak for him, but he and I and Duke LeJeune have been discussing this lately over on AudioRoundTable.com.
I've been making DI matched loudspeakers for well over a decade, and like the post Norris referenced shows, I modeled my three Pi and four Pi loudspeaker designs somewhat on the principal features of the 1983 Keele/Eargle/Smith paper. Different horns and different drivers, but the same basic design features are used.
Using an asymmetrical flare, the loudspeaker can provide a pattern that's wide but not very tall. Drivers positioned one above the other cause nulls to form above and below the speaker. If the layout places those nulls at the outside edge of the horn's vertical pattern, they tend to punctuate pattern cutoff. In a sense, the nulls actually become an asset because they help the horn with pattern control at the low end of its range. Since the nulls limit the uniform radiation angle anyway, it's nice to use them to your advantage.
The main point of the post Norris was referring to is vertical directivity, not waveguide or horn features. You can use a prolate spheroidal flare if you like the way the OS/PS throat is radiused. The point was to limit the vertical angle to the position of the off-axis nulls.
Hi all. I've been offline the past few days, collaborating with a couple of others on a new horn optimized to match with a single 15" driver or a pair of 12" drivers (radiating areas are similar). You'll be hearing more about this as time goes on.
The project still has this form: high-efficiency ribbon, 6 ~ 8 kHz crossover, large-format compression driver and a low-diffraction, smooth-group-delay horn, 700 Hz ~ 1 kHz crossover, single 15" or a pair of 12" drivers, and a pair of 15" bass-fill drivers that are independently powered and EQ'ed for dipole/quasicardioid application.
Aside from the bass-fill drivers, the driver lineup is similar to large-format professional monitor speakers. Similar to professional monitors, efficiency in the 97~102 dB/metre range and generous headroom are design goals. Unlike most professional monitors, settling time for the impulse response is a high priority; thus the selection of drivers and attention to diffraction reduction techniques.
Mige0, thanks again for Post #3909. The last CSD shows the kind of bizarre behavior that real-world drivers can exhibit. This kind of wacky energy storage is easily overlooked if all you examine is the frequency response (you wouldn't see the long energy tail) or the impulse response (all you see is a lot of clutter). Although you report good subjective results, I get wary when the measurements look this questionable - they make me dig deeper and try to find how the coloration is concealing itself.
Salas, our current crop of politicians better watch out for the Erinyes. Sometimes I wonder if these things pursue them into the next life - they certainly seem to get away with it in this one. To all things there is a season.
To extend the roll call of thanks to all who are continuing to extend and expand the art today (as we all know, it sure ain't for the money) - Dr. Earl Geddes, Tom Danley, Jean-Michel Le Cleac'h, John Janowitz, Nick McKinney, Aleksandar Radisavljevic, and a whole bunch of others I can't think of right away. Audio may be out of fashion in the engineering schools, but it still matters to music lovers.
The project still has this form: high-efficiency ribbon, 6 ~ 8 kHz crossover, large-format compression driver and a low-diffraction, smooth-group-delay horn, 700 Hz ~ 1 kHz crossover, single 15" or a pair of 12" drivers, and a pair of 15" bass-fill drivers that are independently powered and EQ'ed for dipole/quasicardioid application.
Aside from the bass-fill drivers, the driver lineup is similar to large-format professional monitor speakers. Similar to professional monitors, efficiency in the 97~102 dB/metre range and generous headroom are design goals. Unlike most professional monitors, settling time for the impulse response is a high priority; thus the selection of drivers and attention to diffraction reduction techniques.
Mige0, thanks again for Post #3909. The last CSD shows the kind of bizarre behavior that real-world drivers can exhibit. This kind of wacky energy storage is easily overlooked if all you examine is the frequency response (you wouldn't see the long energy tail) or the impulse response (all you see is a lot of clutter). Although you report good subjective results, I get wary when the measurements look this questionable - they make me dig deeper and try to find how the coloration is concealing itself.
Salas, our current crop of politicians better watch out for the Erinyes. Sometimes I wonder if these things pursue them into the next life - they certainly seem to get away with it in this one. To all things there is a season.
To extend the roll call of thanks to all who are continuing to extend and expand the art today (as we all know, it sure ain't for the money) - Dr. Earl Geddes, Tom Danley, Jean-Michel Le Cleac'h, John Janowitz, Nick McKinney, Aleksandar Radisavljevic, and a whole bunch of others I can't think of right away. Audio may be out of fashion in the engineering schools, but it still matters to music lovers.
Etienne88 said:
Lynn Olson said:
Lynn Olson said:
Etienne, Lynn thanks for your positive and encouraging comments.
Yes, I wanted to make clear what we deal with - at least most of the time.
AND yes, I wanted to make clear that there is surprise waiting just around any corner – hence I put that picture (http://members.aon.at/kinotechnik/diyaudio/dipol/space/CSD_HowTo/Westra.gif) at the very end.
Sonic comparison in short - the "cluttered" speaker sounds way more alive and its NOT the clutter (in this case) that sounds good. Its freedom in movement and an envelope of rise and decay we are familiar with.
Overly dampened membranes (and VC) add something dead. A strange coloration that should not be possible at all.
ucla88 said:
john k... said:
Mark, JohnK, thanks for outlining the interchangeability of frequency response TOGETHER WITH phase response and impulse response in detail.
But mercy!
Still, if YOU can see what to expect in terms of slow decay (CSD) when looking at a frequency-phase-plot, veeeery view people will be able.
And frankly – me neither – as I am just too lazy to learn usage of magnifying glasses when there is a comfortable way to produce AND a intuitive way to read great (CSD) plots.
Check yourself:
BEFORE equalizing:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
AFTER equalizing:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Besides that I would like to come back to my starting point – its easy to do CSD and it provides additional data – not in the mathematical terms specified above – but in everyday live where its even hard to get decent FR from manufacturers.
Very often datasheets provide way too less resolution (FR), veeeery seldom you see phase response included – look at all the PRO like RCF, B&C, BMS, 18Sound, JBL, EV, etc
ucla88 said:
.
" pseudoscientifically validate misattribution" – bad word and I strongly disagree with "you have to know what the csd should look like" – its simply too obvious.
Can you provide any CSD plots that aren't - in contrary?
Sadly I don't have the time to put this together. It should be rather easy to do that by some free convolution software like SIR:
Imagine a speaker with a triangle close by behind a curtain.
The CSD of a triangle alone will look something like we saw from the B100 at the single resonance (plus harmonics) – without having any wider FR than that single tone (plus harmonics) everybody is familiar with..
Now – every time our speaker has to reproduce tones that correspond to the triangle - that triangle will be stimulated and will decay as it does normally.
In fact, veeery bad dampening of the triangle through the speaker, you agree?
😀
This will overlay the - lets say ideal – speaker behaviour, you agree?
Audition would reveal this speaker sounding slightly like a triangle, you agree?
The FR of the system speaker-triangle will show no great changes but the CSD will for sure.
Easy to imagine how the CSD of the system speaker-triangle would look like, no?
AND easy to imagine how to cure this very simple system, no?
So why refuse to accept CSD as a great tool?
I know why – its just "too easy"!
Greetings
Michael
mige0 said:
mige0 said:
If everything were properly captured on the recording and amplified in its proper proportion to the speaker, and the speaker were able to dynamically convert it without compression, then a highly damped speaker should not sound dead. Admittedly, lotta ifs there.
Now, are you saying that something is missing from the reproduction chain and that the speaker needs to fill it in? In other words, the speaker needs to re color the sound? So it's just a matter of what added color reconstitutes a realistic illusion?
Sheldon
One thing not so widely covered by this thread is the melting of sound not by compression but by power loss due to VC heating (and subsequent effects with passive XO).
This was sparkling a little between the lines of John_E_J postings but sadly there was no "hard data" given and no further discussion developed.
Not many manufacturers do provide that though.
Comparing a 1.5" VC speaker and a 3" VC speaker can improve from 4,5dB to 1,5 dB = 3dB less power losses at 100W input.
Quite something and easily making 1dB refinement in equalising and driver matching questionable.
(Don't get me wrong – it IS reasonable IMO to do equalizing down to– or even below 1dB)
Interesting that some PRO drivers use aluminium wires.
I found different temperature coefficients for aluminium and copper respectively. Anybody who knows which ones apply here?
I would be interested to hear what your sonic findings are in this respect.
Greetings
Michael
This was sparkling a little between the lines of John_E_J postings but sadly there was no "hard data" given and no further discussion developed.
Not many manufacturers do provide that though.
Comparing a 1.5" VC speaker and a 3" VC speaker can improve from 4,5dB to 1,5 dB = 3dB less power losses at 100W input.
Quite something and easily making 1dB refinement in equalising and driver matching questionable.
(Don't get me wrong – it IS reasonable IMO to do equalizing down to– or even below 1dB)
Interesting that some PRO drivers use aluminium wires.
I found different temperature coefficients for aluminium and copper respectively. Anybody who knows which ones apply here?
I would be interested to hear what your sonic findings are in this respect.
Greetings
Michael
Sheldon said:
No, nothing to fill in by the speaker.
There is just something that MAKES sound lacklustre.
Greetings
Michael
Material of speaker cone is one reason if it sounds dead.
cone shape may be another reason depending on how it effects acoustic transients.
cone shape may be another reason depending on how it effects acoustic transients.
soongsc said:
Since all cones are going to resonate and we can't get rid of it. Perhaps it is more important that a cone has good sounding "noise". Maybe, it is the quality or shape of the noise spectrum, in addition to being low in level that is important.
This is out in left field but I wonder if the noise of a cone driver in breakup mode can be likened to the positive effect of adding dither in digital filters. And I will support the wild idea from Chaos theory. Since when the cone is in breakup mode it is a very chaotic and non solid state system. I speculate that it can be more easily influenced by a low level signal than if it were a completely well behaved, non resonant and perfectly damped system. Could it be then that overly damped systems cannot respond well enough to very low level signals?
Now we're touching on coloration, a topic of endless depth. Good.
In principle, we'd like to build loudspeakers with the lowest coloration possible, so the music (which of course is filled with color) shines through. Unfortunately, the optical analogy breaks down, since unlike lenses, we cannot build loudspeakers with no audible coloration.
I've never heard a loudspeaker that doesn't have "loudspeaker" written all over it, nor have I ever heard an audio system that was indistinguishable from the real thing. Never. It always sounds like some flavor of hifi - sometime delectable, many times not.
Now, if you're lucky enough to find any system indistinguishable from the real thing, that's terrific! Your design task is then much easier - just aim in the direction that works for you, and keep on doing it.
That's not to say we shouldn't try to get coloration to the lowest perceptible level possible - that remains an excellent goal. But, considering there's always going to be a residue, what does the residue sound like? Is it disagreeable, or reasonably consonant with musical values? Are there methods to conceal it from (obvious) perception? Do the concealment methods continue to work with more extended listening, and with a wide variety of sources?
Going further, I find coloration from loudspeakers comes in two flavors: one adds colors that aren't there, such as resonances, diffraction, reflection, cabinet standing-wave modes, etc. etc. This is most audible with pink-noise, somewhat less so with large-scale symphonic and choral works, and least audible with sparse instrumentation and a lot of noodling around in the studio. If it's gross enough, though, it'll be obvious from the first moments the hifi is turned on - honky, harsh, boomy, shrill, you get the idea.
The other flavor of coloration is something we'd don't expect: it removes colors from the original performance. I've heard this often enough, and have had to contend with it in amplifiers and loudspeakers. I sometimes think of it as an overall "gray" coloration that washes out the vividness of the performances - think of a second-generation Dolby B cassette recording. Murky, opaque, dull, congested. There's no frequency-response tilting or anything obvious like that, but flip the switch and listen to the mastertape, and the degradation is obvious.
It's usual to point the finger at wretched electronics or lousy tape recordings, but loudspeakers can do this too. They can wash out the tone colors, smudge fine detail, and lose much of the expressiveness of a performance. Lots of IM distortion is the first place to look, but there can be subtler types of modulation noise that can be troublesome - noises from the spider, velocity effects in vents, and I suspect some types of damping materials have odd colorations in a subtractive sense. Weird things like this can be hard to measure, and are more easily heard by direct substitution of the driver with a different model.
The only reason I bring this up is that I've gone pretty far down the road to a low-coloration loudspeaker before, and found when I was finished that something was missing - not the coloration, but a sense of liveliness, sparkle, and excitement. It is entirely possible to build a loudspeaker with low subjective coloration by all the usual criteria (subjective and objective), but end up with a speaker that frankly doesn't play music all that well - it sounds flat, dull, and unexciting - but switch to a set of top-of-the-line headphones, and then the music comes alive again. If speakers make all of your recordings sound flat and dull, no matter how they measure, something is very wrong.
In principle, we'd like to build loudspeakers with the lowest coloration possible, so the music (which of course is filled with color) shines through. Unfortunately, the optical analogy breaks down, since unlike lenses, we cannot build loudspeakers with no audible coloration.
I've never heard a loudspeaker that doesn't have "loudspeaker" written all over it, nor have I ever heard an audio system that was indistinguishable from the real thing. Never. It always sounds like some flavor of hifi - sometime delectable, many times not.
Now, if you're lucky enough to find any system indistinguishable from the real thing, that's terrific! Your design task is then much easier - just aim in the direction that works for you, and keep on doing it.
That's not to say we shouldn't try to get coloration to the lowest perceptible level possible - that remains an excellent goal. But, considering there's always going to be a residue, what does the residue sound like? Is it disagreeable, or reasonably consonant with musical values? Are there methods to conceal it from (obvious) perception? Do the concealment methods continue to work with more extended listening, and with a wide variety of sources?
Going further, I find coloration from loudspeakers comes in two flavors: one adds colors that aren't there, such as resonances, diffraction, reflection, cabinet standing-wave modes, etc. etc. This is most audible with pink-noise, somewhat less so with large-scale symphonic and choral works, and least audible with sparse instrumentation and a lot of noodling around in the studio. If it's gross enough, though, it'll be obvious from the first moments the hifi is turned on - honky, harsh, boomy, shrill, you get the idea.
The other flavor of coloration is something we'd don't expect: it removes colors from the original performance. I've heard this often enough, and have had to contend with it in amplifiers and loudspeakers. I sometimes think of it as an overall "gray" coloration that washes out the vividness of the performances - think of a second-generation Dolby B cassette recording. Murky, opaque, dull, congested. There's no frequency-response tilting or anything obvious like that, but flip the switch and listen to the mastertape, and the degradation is obvious.
It's usual to point the finger at wretched electronics or lousy tape recordings, but loudspeakers can do this too. They can wash out the tone colors, smudge fine detail, and lose much of the expressiveness of a performance. Lots of IM distortion is the first place to look, but there can be subtler types of modulation noise that can be troublesome - noises from the spider, velocity effects in vents, and I suspect some types of damping materials have odd colorations in a subtractive sense. Weird things like this can be hard to measure, and are more easily heard by direct substitution of the driver with a different model.
The only reason I bring this up is that I've gone pretty far down the road to a low-coloration loudspeaker before, and found when I was finished that something was missing - not the coloration, but a sense of liveliness, sparkle, and excitement. It is entirely possible to build a loudspeaker with low subjective coloration by all the usual criteria (subjective and objective), but end up with a speaker that frankly doesn't play music all that well - it sounds flat, dull, and unexciting - but switch to a set of top-of-the-line headphones, and then the music comes alive again. If speakers make all of your recordings sound flat and dull, no matter how they measure, something is very wrong.
Lynn Olson said:
Such loudspeakers you can meet in studio control rooms more than anywhere else.
Also, listening to good headphones is always a great sanity check. I always do that when I change something anywhere in the system, but especially when voicing a loudspeaker.
Lynn Olson said:
We must escape mass to reach real sound. A big plasma ball modulated by signal, hanging on air in the focal point of the perfect waveguide.
Somewhere in the future maybe.
I think the wires to the speaker have to be as nearly straight as possible. If there are too many bends and too many strange devices for the electrons to negotiate, very little life is left in the electricity by the time it reaches the speaker. If you were to disconnect the speaker, the energy would just sort of dribble from the ends of the wires instead of spurting as it should.
As a corollary of sorts, it is a good idea to put shorting plugs on unused inputs. This keeps the juice from leaking out the front end as well.
As a corollary of sorts, it is a good idea to put shorting plugs on unused inputs. This keeps the juice from leaking out the front end as well.
Especially with digital sources. The 0's are smooth enough to get through bent and twisted speaker wire, but the 1's tend to hang up on any kinks. Note that one end of a "1" is pointier than the other - this is why speaker cables are directional. Simple, really...
Russell Dawkins said:
avoid adding too much electron spin with big coils in passiv XO's or connect VC the other way around at least...
😉
mikey_audiogeek said:
Moreover, Data are like American race car drivers, who can only drive around in circle and will get into trouble in turns and corners. So, you need to dumb down the driving requirements and make sure that Data only need to drive straight. Otherwise, Data may be lost.