Re: Re: Re: To Lynn..
http://www.shakti-innovations.com/images/Hoffman-pic-1.jpg
I prefer to nail them to a broomstick, bang it to get a rhythm and sing along...Tie me kangaroo down, sport...
True, but I prefer the 10% THD method.EdwardWest said:
Sadly, 'tis true.terry j said:
Don't need to. Buy one of these, get a free mastering engineer.terry j said:
http://www.shakti-innovations.com/images/Hoffman-pic-1.jpg
We could make a fortune selling 'Acoustic Adulterators' just from what he have laying around here.terry j said:
I prefer to nail them to a broomstick, bang it to get a rhythm and sing along...Tie me kangaroo down, sport...
We do need a good larf periodically. And he does say nice things about Nelson's gear.terry j said:
Back to drawing board, no AutoCad here.
Guys ... Was snooping around B&O's site and the ellipse was mentioned. I remembered how easy it is to draw.
http://www.cleavebooks.co.uk/scol/callipse.htm
Looks as if it would make an excellent OB shape.
Options: cut them off after rounded ends or make thin minor axis and extend all the way around. Bendable plywood is relatively easy to obtain. I have a 1/8" 4x8 sheet ($22) coming for a look see. Note: non-sudsy ammonia will make wood bend like crazy. Clears the sinus passages at the same time.
A first thought was its concentric nature. Offset speakers are usually better.
So I guess a flat center panel could be built to more easily mount speaker(s) and use different ellipses on each side.
Definitely has to be a step above curved edges, or feathers glued to sharp ply edges.
Zene
Guys ... Was snooping around B&O's site and the ellipse was mentioned. I remembered how easy it is to draw.
http://www.cleavebooks.co.uk/scol/callipse.htm
Looks as if it would make an excellent OB shape.
Options: cut them off after rounded ends or make thin minor axis and extend all the way around. Bendable plywood is relatively easy to obtain. I have a 1/8" 4x8 sheet ($22) coming for a look see. Note: non-sudsy ammonia will make wood bend like crazy. Clears the sinus passages at the same time.
A first thought was its concentric nature. Offset speakers are usually better.
So I guess a flat center panel could be built to more easily mount speaker(s) and use different ellipses on each side.
Definitely has to be a step above curved edges, or feathers glued to sharp ply edges.
Zene
Zene,
My first explorations in EDGE were with ellipse forms. Much better for ripples in FR than any of the hard edged shapes, except one.
And even that was not exactly hard edged and has since morphed into essentially a Buck Rodgers Space Ship outline. Of course, anything that looks this foolish and sims so well, must be built. So, I am letting those who can tell the difference between wood and fingers, when encountering spinning saw blades, build them. Much trading going on cross border of course. But, the builders are first rate and if I can ever get them plans they can read, saw dust should fly.
Bud
My first explorations in EDGE were with ellipse forms. Much better for ripples in FR than any of the hard edged shapes, except one.
And even that was not exactly hard edged and has since morphed into essentially a Buck Rodgers Space Ship outline. Of course, anything that looks this foolish and sims so well, must be built. So, I am letting those who can tell the difference between wood and fingers, when encountering spinning saw blades, build them. Much trading going on cross border of course. But, the builders are first rate and if I can ever get them plans they can read, saw dust should fly.
Bud
Bud ... Thought I would be off in my explanation. I mean elliptical as you look down on panel and would be curved away as you look at its face. A squished stand up tube if you like that better. Further Budweiser thoughts was to have smaller ellipse at top and larger at bottom. The configurations are endless.
Zene
Zene
Dave,
Yes, of course! I especially liked the one you sent to Scott, from sketch up I believe? Very mysterious and arcane looking. Helps to offset the basic silliness involved in how it really looks.
Zene,
Sort of like a warriors shield?. I would very much like to see a sketch of what you are envisioning, and trying to pound into my rather dense brain, with words.
My base level of mentation (as if) is that straight edged geometrical shapes are somehow insulting to the audio pressure wave and are possibly the real reason for such worry over how disruptive the diffracted energy is.
I am thus instinctively attracted to rounded surfaces and edges. I did try to model a sort of French Curve derived shape in Edge, but without any noticeable success, whereas the Rocket Ship shape Dave will shortly make appear, was useful right away. Even though, in my usual witless fashion I did the initial sim with 100 meters as the microphone distance.
Some folks here were kind enough to very gently point that out, as opposed to how I would have been eviscerated on some other forums. However, even with closer mic distances, the sim was still very encouraging.
Bud
Yes, of course! I especially liked the one you sent to Scott, from sketch up I believe? Very mysterious and arcane looking. Helps to offset the basic silliness involved in how it really looks.
Zene,
Sort of like a warriors shield?. I would very much like to see a sketch of what you are envisioning, and trying to pound into my rather dense brain, with words.
My base level of mentation (as if) is that straight edged geometrical shapes are somehow insulting to the audio pressure wave and are possibly the real reason for such worry over how disruptive the diffracted energy is.
I am thus instinctively attracted to rounded surfaces and edges. I did try to model a sort of French Curve derived shape in Edge, but without any noticeable success, whereas the Rocket Ship shape Dave will shortly make appear, was useful right away. Even though, in my usual witless fashion I did the initial sim with 100 meters as the microphone distance.
Some folks here were kind enough to very gently point that out, as opposed to how I would have been eviscerated on some other forums. However, even with closer mic distances, the sim was still very encouraging.
Bud
Hi all, thanks for excellent contributions, much appreciated as always. Bud, thanks for the phone call, sometimes I need a little cheering up, and you did the trick, raising my spirits once again. Thank you!
Been thinking this morning about my rather flip dismissal of the mid-horn variation of the speaker - this was a little too casual. There are things about the proposed design that I don't like - the cone drivers and the ribbon tweeter are close to the edges of their working range, and the vocal region (which I consider falling in the telephone bandwidth of 500 Hz ~ 3 kHz) is split by the 1.6 kHz crossover.
There is a different variation possible for horn enthusiasts: a big mid-horn covering the 500 Hz to 5 kHz range, an efficient ribbon tweeter above that (RAAL naturally), and either the delta-trio of 12" drivers or a vertical pair of 15" drivers covering the 80 ~ 500 Hz dipole range (with its own room/dipole EQ and amplification).
If I were doing this, the Azurahorn AH-340 with a 1.4" or 2" throat would probably be my first choice, with a big Altec multicell in second place. Potential 1.4" to 2" exit compression drivers span the range from unobtanium like Vitavox, Vitavox, Westrex, etc. to modern drivers from Radian, JBL, TAD, and who knows, Cogent.
This is a completely different-sounding speaker than what I've discussing so far, but it would be good one. The horn would be working in its most comfortable range, same for the woofers, and the same for the ribbon tweeter. Nothing is being "pushed" to the edge of its range, and the vocal range is covered by one driver, the horn/CD system.
The gotcha, and it's a huge one, is the extremely demanding requirement on the horn/CD system. It's in the center of the spectrum, covering a range where the ear is at its most acute in terms of audibility of peaks, hidden resonances, and distortion, and not by coincidence, that's the voice range as well. I don't think any small-format (1" exit) compression driver would be up to the task - this is a job for the big boys, serious theatre-size compression drivers with 3-to-4 inch diaphragms and 1.4~2-inch exits, and a matching high-performance horn (adapters not welcome).
The reason I picked this frequency range is that horns are troublesome to design below 500 Hz, with resonance-inducing, bandwidth-limiting tricks like folds, corners, and badly-sized mouth areas that don't properly interface with the floor boundary.
Above 5 kHz, HOM's really start to dominate the frequency response, rapidity of decay in the time response, and polar pattern, and response above 10 kHz is usually problematic at best, with any size of horn. My instinct is that horns & waveguides are at their best in the center of the spectrum, and only covering a moderate frequency range of 3 octaves to one decade of frequency response, with other technologies covering the bass and HF ranges.
Been thinking this morning about my rather flip dismissal of the mid-horn variation of the speaker - this was a little too casual. There are things about the proposed design that I don't like - the cone drivers and the ribbon tweeter are close to the edges of their working range, and the vocal region (which I consider falling in the telephone bandwidth of 500 Hz ~ 3 kHz) is split by the 1.6 kHz crossover.
There is a different variation possible for horn enthusiasts: a big mid-horn covering the 500 Hz to 5 kHz range, an efficient ribbon tweeter above that (RAAL naturally), and either the delta-trio of 12" drivers or a vertical pair of 15" drivers covering the 80 ~ 500 Hz dipole range (with its own room/dipole EQ and amplification).
If I were doing this, the Azurahorn AH-340 with a 1.4" or 2" throat would probably be my first choice, with a big Altec multicell in second place. Potential 1.4" to 2" exit compression drivers span the range from unobtanium like Vitavox, Vitavox, Westrex, etc. to modern drivers from Radian, JBL, TAD, and who knows, Cogent.
This is a completely different-sounding speaker than what I've discussing so far, but it would be good one. The horn would be working in its most comfortable range, same for the woofers, and the same for the ribbon tweeter. Nothing is being "pushed" to the edge of its range, and the vocal range is covered by one driver, the horn/CD system.
The gotcha, and it's a huge one, is the extremely demanding requirement on the horn/CD system. It's in the center of the spectrum, covering a range where the ear is at its most acute in terms of audibility of peaks, hidden resonances, and distortion, and not by coincidence, that's the voice range as well. I don't think any small-format (1" exit) compression driver would be up to the task - this is a job for the big boys, serious theatre-size compression drivers with 3-to-4 inch diaphragms and 1.4~2-inch exits, and a matching high-performance horn (adapters not welcome).
The reason I picked this frequency range is that horns are troublesome to design below 500 Hz, with resonance-inducing, bandwidth-limiting tricks like folds, corners, and badly-sized mouth areas that don't properly interface with the floor boundary.
Above 5 kHz, HOM's really start to dominate the frequency response, rapidity of decay in the time response, and polar pattern, and response above 10 kHz is usually problematic at best, with any size of horn. My instinct is that horns & waveguides are at their best in the center of the spectrum, and only covering a moderate frequency range of 3 octaves to one decade of frequency response, with other technologies covering the bass and HF ranges.
Bud ... This is as crude as I can draw, and as good as I can draw.
If you look at the ellipse calculator I sent, you see how quickly many variations can be drawn.
Lynn ... I think the ellipse could be added to the mouth of a horn for a quick listen. Cardboard is a great tool. In fact the ellipse itself could be permanently made from cardboard, after all it's just wood, and wood sounds good.
Zene
If you look at the ellipse calculator I sent, you see how quickly many variations can be drawn.
Lynn ... I think the ellipse could be added to the mouth of a horn for a quick listen. Cardboard is a great tool. In fact the ellipse itself could be permanently made from cardboard, after all it's just wood, and wood sounds good.
An externally hosted image should be here but it was not working when we last tested it.
Zene
I'd like to draw everyone's attention to a comment made by Dr. Geddes in the "Geddes on Waveguides" thread.
This is worthy of some thought, along with the previous comment in the thread about the success of using foam in the throat region of a horn to smooth out the transition from the compression driver exit angle to the start of the throat of the horn. (Unless the CD and horn/WG are designed to be used to together, these angles frequently doesn't quite match, thus generating strong HOM's right at the beginning of the expansion.)
Although I don't expact strong HOM's on the flat plane of the baffle, they won't be zero, either. Thus the suggestions to cover all of the rear baffle surface with 3/8" of wool felt, and use full or partial bullnose round-overs on the front surface. Anywhere there's a sharp transition to the expanding wavefront, diffraction will be generated and re-emitted at the boundary edge.
Picking up the previous comments I made above, the topology I've been discussing all along (not the horn version) does have issues with drivers working close to the edge of their ranges. The intuitive answer to gaining more HF extension (by using a smaller driver) runs into the serious drawback of hitting a brick wall at a given SPL level.
Cones are constant-acceleration devices (the acceleration is the same regardless of frequency), and beyond a certain G level, the cone starts to break up - badly. Exotic materials have a somewhat higher ultimate breakup level, but the almost inevitable trade-off is much more severe and chaotic break-up when it occurs - the "brick wall" is much harder and the audibility of break-up is more of a shriek than the gradual onset of mush (with softer materials like paper). Not only that, the choice of different cone materials only affects the onset of breakup by a few dB (keeping the cone size constant) - don't expect a 10 or 15 dB improvement from a "miracle" cone material.
By comparison, large cones don't need to accelerate as much as smaller cones, simply because the emissive area is larger. This means more headroom without resorting to exotic cone materials. The tradeoff is what size is "too large" for the voice coil and spider to control - certainly, we shouldn't be expecting much midrange from a 18" cone. But the cone sizes of 8 inches to 15 inches are wide open - the smaller cone will excel in the small-signal condition, and the bigger one will excel when substantial peak SPL's are required. This is the sense of "headroom" that horns are known for - another name would be "Dynamic Fidelity", something you get from studio monitors but almost never from audiophile speakers.
This, by the way, is why I keep dragging the design back to 12 and 15-inch drivers and large-area ribbon tweeters. I want a reasonable portion of the Dynamic Fidelity of horns without all the endless design hassles - mostly centered around good time-domain performance, the area where horns are weakest. The graph below shows the time performance of the Ariel - compare it to any horn or waveguide system and see which has the quickest return-to-zero decay performance.
The Ariel wasn't even optimized for time performance, since it has a 360-degree phase rotation at the crossover frequency of 3.8 kHz, and the tweeter is acoustically ahead of the Vifa midbass drivers. Even so, it's pretty much returned to zero in 700 microseconds - if the echo at 500 microseconds was chased out, it would return to zero even faster.
The goal for the new system - with the widerange driver, delta-trio bass array, and double-high ribbon tweeter - is time performance as good or better than the Ariel, 10~15 dB more headroom, and oh yes, it would be nice if it sounded good too. For the first go-round, I'll be using MLSSA, and for long-term optimization, teaching myself how to use more modern software (like SoundEasy) on the new PC with the M-Audio 192/24 soundcard.
gedlee said:
This is worthy of some thought, along with the previous comment in the thread about the success of using foam in the throat region of a horn to smooth out the transition from the compression driver exit angle to the start of the throat of the horn. (Unless the CD and horn/WG are designed to be used to together, these angles frequently doesn't quite match, thus generating strong HOM's right at the beginning of the expansion.)
Although I don't expact strong HOM's on the flat plane of the baffle, they won't be zero, either. Thus the suggestions to cover all of the rear baffle surface with 3/8" of wool felt, and use full or partial bullnose round-overs on the front surface. Anywhere there's a sharp transition to the expanding wavefront, diffraction will be generated and re-emitted at the boundary edge.
Picking up the previous comments I made above, the topology I've been discussing all along (not the horn version) does have issues with drivers working close to the edge of their ranges. The intuitive answer to gaining more HF extension (by using a smaller driver) runs into the serious drawback of hitting a brick wall at a given SPL level.
Cones are constant-acceleration devices (the acceleration is the same regardless of frequency), and beyond a certain G level, the cone starts to break up - badly. Exotic materials have a somewhat higher ultimate breakup level, but the almost inevitable trade-off is much more severe and chaotic break-up when it occurs - the "brick wall" is much harder and the audibility of break-up is more of a shriek than the gradual onset of mush (with softer materials like paper). Not only that, the choice of different cone materials only affects the onset of breakup by a few dB (keeping the cone size constant) - don't expect a 10 or 15 dB improvement from a "miracle" cone material.
By comparison, large cones don't need to accelerate as much as smaller cones, simply because the emissive area is larger. This means more headroom without resorting to exotic cone materials. The tradeoff is what size is "too large" for the voice coil and spider to control - certainly, we shouldn't be expecting much midrange from a 18" cone. But the cone sizes of 8 inches to 15 inches are wide open - the smaller cone will excel in the small-signal condition, and the bigger one will excel when substantial peak SPL's are required. This is the sense of "headroom" that horns are known for - another name would be "Dynamic Fidelity", something you get from studio monitors but almost never from audiophile speakers.
This, by the way, is why I keep dragging the design back to 12 and 15-inch drivers and large-area ribbon tweeters. I want a reasonable portion of the Dynamic Fidelity of horns without all the endless design hassles - mostly centered around good time-domain performance, the area where horns are weakest. The graph below shows the time performance of the Ariel - compare it to any horn or waveguide system and see which has the quickest return-to-zero decay performance.
The Ariel wasn't even optimized for time performance, since it has a 360-degree phase rotation at the crossover frequency of 3.8 kHz, and the tweeter is acoustically ahead of the Vifa midbass drivers. Even so, it's pretty much returned to zero in 700 microseconds - if the echo at 500 microseconds was chased out, it would return to zero even faster.
The goal for the new system - with the widerange driver, delta-trio bass array, and double-high ribbon tweeter - is time performance as good or better than the Ariel, 10~15 dB more headroom, and oh yes, it would be nice if it sounded good too. For the first go-round, I'll be using MLSSA, and for long-term optimization, teaching myself how to use more modern software (like SoundEasy) on the new PC with the M-Audio 192/24 soundcard.
Attachments
Like This...
Lynn,
Your goal is achievable
This impulse response is of my current 3-way system - an active, tri-amplified system. I have chosen a 15" Lambda (93 dB), a 6.5" PHL (95 dB) and a 3/4" Hiquophon (89 dB). As you can see, the time domain performance is comparable to the Ariel's with substantially more headroom. In spite of the size of the drivers (against conventional HiFi wisdom) they are essentially silent in 500 microseconds. The system also possesses excellent detail resolution, while sounding incredibly lifelike. It excels with instruments like guitar, harpsichord and piano - all of which are very unforgiving of time-domain short-falls.
But most important of all... They are very easy and incredibly enjoyable to listen to.
Edward
Lynn,
Your goal is achievable
This impulse response is of my current 3-way system - an active, tri-amplified system. I have chosen a 15" Lambda (93 dB), a 6.5" PHL (95 dB) and a 3/4" Hiquophon (89 dB). As you can see, the time domain performance is comparable to the Ariel's with substantially more headroom. In spite of the size of the drivers (against conventional HiFi wisdom) they are essentially silent in 500 microseconds. The system also possesses excellent detail resolution, while sounding incredibly lifelike. It excels with instruments like guitar, harpsichord and piano - all of which are very unforgiving of time-domain short-falls.But most important of all... They are very easy and incredibly enjoyable to listen to.
Edward
salas said:
I use the PHL 1120 - a true midrange, playing from 350 to 3500 Hz.
I often play the system pretty loud, but I sit about 2.5 meters from the speakers, so the sustained SPL requirements are pretty reasonable. Often I listen at an average level of 85 - 90 dB, but there have been a few sessions where the sustained level were near 100 dB. For a system which is capable of this level of output and is flat from 20Hz to 20KHz, music becomes very hypnotic to me. I have been developing this system for about two-years, and as it gets better and better I find I listen to more music and work on it less.
Edward
PHL is good. I have chosen the 1220 slower brother of yours, bcs I needed a slow cut under 200Hz, and its lower Fs helped. I did a trick with the impedance under 100Hz you see. 3 drivers blending there. Aperiodic mid-woofer 1220 and double 6X9s shielded. But no usual peaks.
Attachments
Re: Like This...
Nice time response - again, I doubt very much any horn or waveguide could match that, at least from the data I've seen published. This takes us to electrostatic territory with much better dynamics thrown in. Curious about the crossover points you've chosen - I imagine the 3/4" HF driver is crossed fairly high, to take advantage of the PHL's extension.
Curious about the pre-ringing over the first 300 microseconds - it's not easy for a tweeter to do that, it looks more like a sound card artifact due the built-in anti-alias lowpass filter - set to 20 or 40 kHz, maybe? Although my MLSSA card is ancient, dating back to 1991, I run it at 120 kHz to avoid lowpass filter artifacts and see what the tweeter is really doing. If the pre-ringing is what I think it is, your speaker is actually better than what's shown above.
I find the rapid decay in the time response corresponds to depth portrayal, or spatial fidelity - not imaging or reverb per se, but an accurate impression of the actual size of the performing venue. Weirdly enough, you can actually sense this before any music starts to play, or likewise with applause, which is extremely revealing of transient distortion and midrange/HF coloration.
EdwardWest said:
Your goal is achievable
But most important of all... They are very easy and incredibly enjoyable to listen to.
Edward
Nice time response - again, I doubt very much any horn or waveguide could match that, at least from the data I've seen published. This takes us to electrostatic territory with much better dynamics thrown in. Curious about the crossover points you've chosen - I imagine the 3/4" HF driver is crossed fairly high, to take advantage of the PHL's extension.
Curious about the pre-ringing over the first 300 microseconds - it's not easy for a tweeter to do that, it looks more like a sound card artifact due the built-in anti-alias lowpass filter - set to 20 or 40 kHz, maybe? Although my MLSSA card is ancient, dating back to 1991, I run it at 120 kHz to avoid lowpass filter artifacts and see what the tweeter is really doing. If the pre-ringing is what I think it is, your speaker is actually better than what's shown above.
I find the rapid decay in the time response corresponds to depth portrayal, or spatial fidelity - not imaging or reverb per se, but an accurate impression of the actual size of the performing venue. Weirdly enough, you can actually sense this before any music starts to play, or likewise with applause, which is extremely revealing of transient distortion and midrange/HF coloration.
