@tomic601 asked:
"are you 3D printing the waveguide?"
No. I looked into it and concluded that was impractical. The piece is 22" in diameter.
Duke
|
There is an inevitable tradeoff relationship between box size, bass extension, and efficiency. And it’s a fairly brutal relationship.
If you keep the same bass response, every 3 dB increase in efficiency calls for a DOUBLING of box size.
So compared with a 1/2 cubic foot 85 dB mini-monitor that goes down to 40 Hz, a 97 dB efficient speaker with the same bass response would be SIXTEEN times the size (four doublings of box volume), or EIGHT cubic feet.
Now imo there are definitely some qualitative advantages to that higher efficiency, but in my experience the market for eight cubic foot speakers is rather limited.
Anyway my guess is that the higher costs and the box size penalties attached to high efficiency are the primary reason why low efficiency speakers dominate the marketplace.
Duke
|
@phusis wrote:
"What’s particularly interesting with high eff. larger speakers, to me, is that simplicity can be maintained when incorporating horns or waveguides fitted to compression drivers, and thus avoid some of the complexity issues that face low(er) eff. larger speaker designs. A 2-way high eff. design can be had with fairly large woofer/mids crossed to a horn/waveguide with very nice power response in the XO-region, though needing subs augmentation. I use such a configuration myself, now awaiting new horns from ~700Hz on up with a mouth area of some 2 x 3 feet for controlled directivity if needed, down to 500Hz for a smoother transition to the dual 15" drivers below."
This is almost exactly what I’m working on, and had hoped to introduce in 2020 but... stuff happened that year...
Anyway I designed a large-format Oblate Spheroid waveguide using Earl Geddes’ equations, like you targeting a 700 Hz crossover to twin 15" midwoofers. That 700 Hz figure is consistent with the findings of David Griesinger which Geddes subscribes to, and is very close to the 800 Hz crossover that Greg Timbers uses in the JBL M2. Imo the ability to cover the spectrum from there on up with a single driver is a major advantage over more "conventional" approaches, in addition to the other advantages of large drivers and high efficiency.
And of course the way around the bass extension/box size/efficiency tradeoff relationship is to hand off the bottom couple of octaves or so to subwoofers.
I’m rather surprised by how similar our approaches are. I knew we were barking up trees in the same forest, but didn’t realize it was the same tree!
Duke |
@phusis asked a bunch of good questions. My response:
The round waveguide has a pattern width of 75 degrees and will sit atop the midwoofer box, somewhat reminiscent of some of Avantgarde’s models. So the configuration is "HMM" instead of "MHM". (We heard PBN’s "M2!5" speaker which uses the "HMM" configuration, and even at fairly close range with eyes closed it was coherent.)
Passive crossovers, OTL and SET friendly impedance curves (nothing against active, but my target market is elsewhere). Multiple subwoofers south of 70 Hz or so. Obviously not cheap, but there will be some trickle-down to more affordable models.
* * *
Regarding dynamics:
I'm friends with a recording engineer who, for decades, has been measuring the dynamic compression characteristics of loudspeakers, both home audio and prosound. He has amassed data on over a hundred loudspeakers. He measures the compression of peaks, something that might be called "short-term power compression", as it happens vastly faster than long-term thermal compression. Earl Geddes was the first to bring this up to me, and Floyd Toole exchanged a few messages with me on the subject. He sees it too, and said that it's an area which has not been adequately researched.
Anyway my friend finds a strong correlation between efficiency and freedom from compression on peaks. I'm not going into the specifics because I consider them confidential, as he hopes to publish his findings some day, but in general high efficiency and large-diameter voice coils translate to freedom from compression on peaks.
Duke |
@mijostyn wrote:
"Duke, your friend has more work to do. There are so many factors involved that I doubt you can make a blanket statement that high efficiency speakers are ALL more dynamic than low efficiency speakers of various types given appropriate power. "
I don’t think I made a blanket statement. Here is what I actually wrote:
"My friend finds a STRONG CORRELATION between efficiency and freedom from compression on peaks... IN GENERAL high efficiency and large-diameter voice coils translate to freedom from compression on peaks."
Emphasis mine in both quotes.
In my opinion "strong correlation" and "in general" are not blanket assertions, whereas "all" would have been.
Duke
|
@erik_squires , thank you very much for that information on Siegfried Linkwitz's tone burst tests!
I was aware of the article in Stereophile years ago which "debunked" short-term thermal compression, but the methodology in the test was flawed because it looked at the average compression over time, rather than the rapid-onset compression that Linkwitz's test reveals.
"Bigger voice coil, more ventilation, lower power dissipation result in lower dynamic compression."
That's my understanding as well, but JBL went a step further in their M2 studio monitor: They use an alloy in the woofer's voice coil whose resistance stays essentially constant as it heats up. I'm not sure whether they did this for the compression drivers' dual voice coils as well. Anyway that seems to me like a brilliant idea which would be especially welcome in high-end audio speakers where efficiencies are lower and therefore voice coils are smaller.
For the record, my own priorities are much more focused on speaker/room interaction, and the types of drivers which do what I want in that area just happen to be fairly high efficiency.
Duke
|
@erik_squires wrote:
"
I should point out that we should not attribute thermal compression to
what might also be bad acoustics. Very reflective environments will have
similar audible results, in at least as similar as you can type about
them. A lot of bad / compressed treble complaints I’ve seen on audiogon
were addressed with better room treatments. Was it excess reflection, or
better treble/bass balance, or did the improvement in sound quality
lead to turning down the knob, therefore reducing tweeter power
dissipation? Really hard to say unless we are measuring. I sure could
not explain in words how to hear a difference. :)"
Your observation makes total sense to me.
Dynamic contrast can be viewed as a "signal-to-noise-ratio" thing, and to the extent that undesirable/excess reflections raise the effective in-room "noise floor", they reduce the system's dynamic contrast. I suspect this may be more common and/or often of greater audible significance than the short-term thermal compression revealed by those tone-burst tests.
This "signal to noise" ratio thing has implications for the sense of envelopment as well: The further down in level we can still detect the reverberation tails on the recording, the stronger the perception of the recording venue's hall ambience. (This isn't the only thing that matters for "envelopment" to take place, but imo it's one of them.)
The ear/brain system classifies reflections as such based on their spectral content, so imo it makes sense for absorption to be broadband, such that the spectral balance of the reflections is largely preserved (assuming they were spectrally correct to begin with). If the spectral content of the reflections is skewed too much, they are no longer classified by the ear/brain system as "signal", and so they become "noise".
Duke |
@erik_squires wrote:
"As studies in learning and acoustics have shown, filtering out noise is energy consuming. Your brain works harder in an acoustically messy environment and I absolutely feel it."
Totally agree.
In home audio, "your brain works harder" = listening fatigue.
[public service announcement] In an acoustically messy environment like the back of a classroom (or even worse the back of a lecture hall), "your brain works harder" = you are straining to use ALL of your CPU power just to understand the individual words, and that takes not only more energy but also more TIME. So by the time you understand one word, the lecturer has already moved on to the next word, and you have neither the spare CPU power nor the TIME to comprehend complex concepts so that you can store them in your long-term memory. This is one of the reasons why the kids in the back of the classroom get tired within fifteen minutes and are by far the ones most likely to flunk. So even if they are shy introverts, tell your kid and grandkids to sit in the front if at all possible! [/p.s.a.]
Duke |
@erik_squires wrote:
"Another area where attribution is difficult, the dynamic nature of high efficiency systems.
"Is it that, or is it the controlled dispersion?
"I'm
not saying low compression speakers aren't good. I'm saying that some
of what we may attribute to "fast" or high impact speakers is really
just better room integration."
I totally agree.
I would not be surprised to learn that, in practice, room acoustics and radiation patterns usually play as big if not bigger role in real-world dynamic contrast as thermal compression.
Duke |
Mijostyn wrote:
"A speaker that can hit 110dB without compression is going to be more
dynamic than a speaker that can only get to 100 dB even if it is less
efficient."
Agreed.
Mijostyn again: "Another issue is trying to run 15" woofers up to 700Hz then crossing to a
horn."
I understand your skepticsm.
Intuitively it sure seems wrong because it's almost never done in home audio. Actually, running a 15" woofer to 700 Hz is like running a 5" woofer to 2.1 kHz: For the right kind of 15" woofer, it's a piece of cake. (The 15" midwoofer I'm using is plus or minus 1 dB to about 1.7 kHz with no filtering, then it has a 3 dB peak at 2 kHz. Its effective motor-strength-to-moving-mass ratio surpasses every small high-end midwoofer I know of, and falls in the ballpark of 5" cone midranges.)
Some of the finest studio monitors in the world, the classic Augspurgers and the magnificent JBL M2, use the 15" woofer + horn format. There are several brief YouTube videos about the M2 which are imo worth watching.
Mijostyn: "Two very dissimilar drivers crossed right in the meat of the
midrange."
The big woofers and horn-loaded compression drivers are visually dissimilar, but ACOUSTICALLY they are far more similar than most cone midwoof/dome tweet combinations in the crossover region. Let me explain:
What we hear is a combination of the direct sound and the reverberant sound, the latter being dominated by the speaker's off-axis response. Ideally the off-axis response tracks the on-axis response very closely. However if there is a directivity mis-match in the crossover region, it is impossible for the on-axis response to match the off-axis response through the crossover region.
A directivity mis-match in the crossover region is almost inevitable for a cone midwoof/dome tweet combination, because the cone's radiation pattern will be narrower than the smaller dome's radiation pattern. There are two ways around this: One is to widen the midwoofer's pattern by using (hopefully well-behaved) cone breakup, and the other is to use a horn or waveguide of some sort to deliberately narrow the tweeter's radiation pattern so that it matches the midwoofer's.
The latter is what I do, only at a lower frequency than most midwoof/tweet combinations.
Crossover frequencies are a juggling of tradeoffs. Briefly, for a combination of psychoacoustic and practical reasons, imo 700 Hz makes sense. It arguably makes better psychoacoustic sense than just about any higher frequency does.
So like I said I understand your skepticism, but I've put some thought into my (often unorthodox) design decisions.
Duke
|
@ctsooner wrote:
"it's simply not true to make a statement that a higher sensitivity
speaker is more costly to make..."
Based on sixteen years of manufacturing fairly high efficiency speakers, I disagree.
The enclosure is usually the most expensive component and high efficiency calls for large, typically labor-intensive enclosures. As enclosure size goes up so does the enclosure cost, even moreso if the larger enclosure is also more complex. And enclosure size has two hidden costs: Shipping cost and opportunity cost. The latter is due to the sheer amount of space the speaker takes up in a showroom.
Many other costs also are typically (though not always) higher with high efficiency, but enclosure cost is always higher for the larger enclosure assuming an equivalent enclosure build technique, and dramatically so for dramatic size differences, and even moreso as the enclosure complexity increases. Sort of like the cost of building a house goes up as the size and number of rooms (complexity) go up.
Duke
|
@atmasphere wrote:
"Harmonic orders above the 2nd decrease rather slowly as compared to a
circuit that has a ’cubic non-linearity’ (produces the 3rd as the
primary distortion component). An amplifier that has this quality has
its distortion decreasing much faster as the order of the harmonic is
increased! This is important since the ear uses higher ordered harmonics
to sense sound pressure (and assigns the quality of harshness and
brightness to them). IOW, an amplifier with a cubic non-linearity will
sound more detailed (because distortion masks low level detail) and
**smoother** because the higher ordered harmonics are at a lower level.
"In
terms of circuit design an amplifier with this characteristic must be
fully differential and balanced from input to output. In this way even
orders are cancelled with each stage in the amp (instead of being
compounded), leaving the 3rd as the primary distortion component, at
about 1/10th what you would get with a single-ended circuit, assuming
that neither employs any feedback."
VERY INTERESTING!!
I think this explains part of the difference I hear between your amps and good single-ended triode amps, in particular:
"
An amplifier that has this quality has
its distortion decreasing much faster as the order of the harmonic is
increased!" And,
"...even
orders are cancelled with each stage in the amp (instead of being
compounded), leaving the 3rd as the primary distortion component, at
about 1/10th what you would get with a single-ended circuit...".
Thank you for putting this down in writing. I copied it to a file so I can find it again.
Duke
|
