Who else uses RAAL ribbons in a commercial speaker?
I remember some review at 6moons about some rather pricey European speakers... but I forget. The review was very good. I guess the ribbon is quite wideband.
Searching through the Internet I saw Ascend Acoustics... they seem to use RAAL... these small speakers seem interesting and rather affordable for such things.
https://ascendacoustics.com/collect...cts/sierra-2ex-v2-pair?variant=43264014024758
Excuse my ignorance on this... but from a physics point of view I'm curious.
How do you achieve a controlled radiation when the geometry of the driver is static?
Do you work with the driver's different radiation pattern over a frequency range?
Do you work with the interactions of different drivers as arranged in the speaker?
I assume you work in the frequency and time (phase) domains to do this?
The 8C has side ports that emit rear cancellation waves from the 8" woofer. This makes the polar pattern cardioid.
The 8" woofer above maybe 500Hz behaves directionally as you'd expect from an 8" woofer. But the side port prevents it from going omnidirectional below 500Hz. The 8" mid crosses at 100Hz to rear woofers that are omnidirectional.
This is the sort of wizardry that can ONLY be done with DSP and very careful attention to phase relationships between drivers. Not possible with passive. I'm not aware of any passive xover speakers with a cardioid radiation pattern for example.
Practically speaking what this means is that you can place them close to a rear wall without getting a rear wall low-midrange reflection (which is why most audiophiles seem to like having their speakers quite a bit out from a rear wall).
I'm guessing from my own work with OB designs that the side ports have undesirable effects on frequency response. So they fix those problems with DSP and still get the radiation pattern they want.
You can probe deeper about the design here: https://www.perplexity.ai/search/can-you-find-an-article-that-e-uq1dUzKTTa2Xc7yxiN0L8g
The 8" woofer above maybe 500Hz behaves directionally as you'd expect from an 8" woofer. But the side port prevents it from going omnidirectional below 500Hz. The 8" mid crosses at 100Hz to rear woofers that are omnidirectional.
This is the sort of wizardry that can ONLY be done with DSP and very careful attention to phase relationships between drivers. Not possible with passive. I'm not aware of any passive xover speakers with a cardioid radiation pattern for example.
Practically speaking what this means is that you can place them close to a rear wall without getting a rear wall low-midrange reflection (which is why most audiophiles seem to like having their speakers quite a bit out from a rear wall).
I'm guessing from my own work with OB designs that the side ports have undesirable effects on frequency response. So they fix those problems with DSP and still get the radiation pattern they want.
You can probe deeper about the design here: https://www.perplexity.ai/search/can-you-find-an-article-that-e-uq1dUzKTTa2Xc7yxiN0L8g
@perrymarshall -- Dave Gunness was mentioned in post #118. He is now with Fulcrum Acoustic. I don't know anything about this patent or his speaker but you might find it interesting.
audioxpress.com/article/patent-review-passive-cardioid-speaker
www.fulcrum-acoustic.com/audio-technology-insights-resources/passive-cardioid-technology/
audioxpress.com/article/patent-review-passive-cardioid-speaker
www.fulcrum-acoustic.com/audio-technology-insights-resources/passive-cardioid-technology/
This is well explained. While the 8C is amazing, you should check out what D&B does with their concert line arrays. They have dual 10" side firing drivers, Dual 14" forward facing woofers with some kind of mid range and compression driver for mid and highs. The side firing drivers cancel the rear waves for low frequencies and it's baffling to be behind these (which happens for me around 150 times a year due to my profession). It keeps so much noise off of the stage that you can't even tell that they're on sometimes but they're covering a stadium with even sound. To take the wizardry further, they do it with two amplifier channels. One channel drives the dual 14" woofers and the other channel drives side firing 10" woofers and the mid and high. The mid and high have passive components in them. Considering that it takes a good amount of DSP to get the delay on the side firing drivers correct, but keep the mid and highs undelayed then you realize how effective this box is. Also, if you look at the top 10 grossing tours of last year you will see that this line array was used on 7 or 8 of them.
https://www.dbaudio.com/global/en/products/all/series/sl-series/gsl12/
Given all the compromises in any loudspeaker there is no such thing as "optimum in every aspect”
dave
Last edited:
Well, here are a few things to think about...
1) Applying a small amount of delay to a tweeter can often make big difference in how the crossover filters are configured. Many people mischaracterize this as "bringing the tweeter and mid into time alignment", but often we find that perfect time alignment does not give the best performance. With a passive filter, the delay is fixed by the design of the tweeter and mid, and the physical geometry of the baffle. Changing the relative delay between tweeter and mid usually means making the baffle stepped, or tilting it back, or some other way of adding additional path length between the tweeter and the ear. With active filters, adding or decreasing delay can be done independent of the baffle shape. This allows us to build a speaker with a flat baffle, and then adjust delay later on to achieve optimum performance.
2) With a passive filter, we can use series or parallel notch filters to shape the response of a driver, but from a practical standpoint, we usually are limited to one notch filter per driver. So we usually pick the most problematic peak, and apply the notch. In the case of a mid driver, there are often several peaks above the passband which are near enough to the crossover to be troublesome. Our typical solution is to apply a wide notch as a compromise to try to take care of the several peaks. With an active filter, we can use as many PEQ notch filters as there are peaks. We can also use gain-adding PEQs to fill in low spots, a tactic which is hard to employ with passive filters. All of this needs to be done in a judicious manner of course, and it is easy to get carried away. But if done carefully, the active filter can be much more effective at shaping the response of a driver.
3) Passive filters in the 100 - 250 Hz region are rare. When we run the simulations, we understand why... the sizes of the inductors become quite large, and quite expensive. But if we are working with a 15" woofer and a 7" mid driver, this region might be exactly where we want our crossover.
4) In a 3-way, the passive low-pass filter on the woofer will often interact with the woofer impedance. This can create an unwanted boost in the upper bass.
My findings: passive filter bass boost
Purifi blog post: side effects of passive crossover
With an active filter, the bass response can be tailored and adjusted to fit the needs of the design. In my opinion, the very best quality bass comes from a sealed box woofer with a Linkwitz Transform to extend the bass. This technique is much easier to implement with an active filter.
So these are some tools that would be available to you with an active design.
No nothing wild. I have generally found that 2nd order slopes work best between woofer and mid, and 2nd or 4th order works best between mid and tweeter... not all the time, but most of the time. I use FIR DSP filters, so I get the same magnitude/phase relationship that I would get with a passive filter.
Errr, isn’t that one of the main reasons for using DSP ? If you use top flight software based DSP you can have perfect time and phase correction. Try doing that with a passive network.
Were Kroma speakers being demonstrated at the show? I heard some recently at a UK show being driven by an amp I’d never even heard of before, Orpheus Lab, and the sound quality was mesmerisingly good.
This person (me, Art Welter) never tried to sway your opinion of what you heard.
Regardless of what you have been told, speakers used for live sound reinforcement can use passive crossovers with or without digital processing.
In small systems using only two or three drivers, using passive crossovers and a single amp may be more cost effective.
In large systems, the cost savings of using a single processor, whether analog or digital, rather than dozens of high order passive crossovers is huge.
Given a choice between two equally good sounding systems, most will choose the more cost effective solution, if it also fits their visual and ergonomic desires.
I never implied knowledge about a technology changed your experience.
An individual's knowledge of how specific technologies were implemented can help inform their opinions about an experience.
I find it interesting you would think that I would think knowledge about how a system works is required to for it to "sound good".
A system can sound good or bad regardless of the technology it uses.
When I got started building sound systems in the 1970s "active" simply meant the crossover was placed before the amplifier. The crossover itself could have been a passive line level or a line level analog electronic circuit.
"Good sounding" digital circuits didn't become affordable for a long time after that.
As Thomas Gray wrote in 1768 when the only "digital processor" was an abucus:
"And happiness too swiftly flies.
Thought would destroy their paradise.
No more; where ignorance is bliss,
'Tis folly to be wise."
Opinions and recollections of the "digital era" we live in now are interesting.
I provided sound for a Stomu Yamashta concert back around 1977, it was over three decades later I learned his 1971 records were the first digital recordings ever released.
Providing and mixing sound in the active Minneapolis music scene, I'd assumed the first digital recordings were made with the system at Herb Pilhofer's Sound 80 studio in 1978.
At any rate, blaming a technology that has matured quite a bit over the last 54 years for bad sound at a HIFI trade show you attended where most of the speakers "do not sound good" makes no sense to me, even though you didn't happen to hear a good sounding system using DSP at that event.
Art
Last edited:
The adjustable delay would be a plus. Usually, I'm sloping the baffle, or recently, I'm using a wave-guide type tweeter, that's close to aligned with my 5.25" woofer. (2-way.) When that's not the case, I use a higher order filter on the tweeter than the mid, or possibly a sharp notch just below the x-over point, or a damped 3rd order filter. Somehow, I make it work.
I try not to use mids , or woofers, that are not smooth. Typical bump at 5k is OK, but a peak or dip at 1.2k..??? I'd probably not consider it. I ran into that multiple peak issue with a RS270P-4. It was a bear. That design is on hold.
Totally agree on #4. I cross to a sub, so I think some of that can be reduced with a higher x-over filter. Like 120hz where the roll-off starts sooner.I have (in Sims) added a big LCR to reduce the problem. Not practical though for the most part. Too big, and too expensive.
Thanks for your reply.
What exactly is an FIR filter?
Finite Impulse Response vs. Infinite Response?
I can see how the analysis of the impulse response of a system to an input can yield insight into linear distortions and a corresponding filter can be crafted to ensure no changes from input to output.... in the digital analysis realm, specifically with an analog signal that exhibits both time and frequency domain characteristics, I understand how with numerical analysis techniques you can effect both time (phase) and frequency modifications to the signal.
But I still don't know why it is called "Finite" Impulse Response filters.
I've spent an hour reading in to the math of it... and I see a lot of circular explanations on this... It seems like people are throwing big words around without really understanding the mathematics of it! And it confuses me ( heck, I've even studied, and used, numeric analysis! ).
Under what circumstances would you use an FIR vs IR filter, for example?
Finite Impulse Response vs. Infinite Response?
I can see how the analysis of the impulse response of a system to an input can yield insight into linear distortions and a corresponding filter can be crafted to ensure no changes from input to output.... in the digital analysis realm, specifically with an analog signal that exhibits both time and frequency domain characteristics, I understand how with numerical analysis techniques you can effect both time (phase) and frequency modifications to the signal.
But I still don't know why it is called "Finite" Impulse Response filters.
I've spent an hour reading in to the math of it... and I see a lot of circular explanations on this... It seems like people are throwing big words around without really understanding the mathematics of it! And it confuses me ( heck, I've even studied, and used, numeric analysis! ).
Under what circumstances would you use an FIR vs IR filter, for example?
@Bmsluite; as Andrew Jones seems to have impressed you, it is perhaps worth mentioning that he is quite open about the fact that he would much prefer to design active speakers but due to resistance in the marketplace he is prevented from doing so for commercial reasons. Something to cogitate on . . .
Thanks so much for the write-up. I have to say that there are no drivers I've wanted to like more than I've wanted to like Accutons, especially after they came out with their fully front mounted tweeters, mids and woofers. And yet... I've never heard an Accuton based speaker that sang to me. At best they sounded like the most sterile headphones I've heard. On the other hand, given the prices I should feel lucky I haven't really liked them! 🤣
I'm rebuilding my passive speakers as actives... sure hope I don't end up like the DSP based speakers you heard at the show!
I'm rebuilding my passive speakers as actives... sure hope I don't end up like the DSP based speakers you heard at the show!
@tonyEE - I don't understand all of it either. Here's what I know: The difference is in how quickly the energy of any given sample stops having an effect on the samples after it. All of these filters work in one way or another by taking the current sample and what has gone before. Kind of like a capacitor does. With IIR there's no end to when a sample stops affecting future samples. I mean, it eventually gets infinitely small effect but it's not deterministic. With a Finite Impulse Response filter you are guaranteed that the current sample will only affect the next n samples.
I also believe that mathematically FIR is more intense and that FIR allows for phase linearization, which IIR does not.
Someone else will jump in to correct me if I'm wrong.
I also believe that mathematically FIR is more intense and that FIR allows for phase linearization, which IIR does not.
Someone else will jump in to correct me if I'm wrong.
Last edited:
Smarter people can answer this better than I, but I have designed one speaker with FIR crossover filters. Essentially, the phase doesn't change when you implement the high or low pass filter. You can use any slope order that you want and the end result is that the phase is generally at zero when you measure them together. There is no phase wrap around the crossover frequencies.
This speaker is the best sounding that I have ever made. It's much more accurate and has more detail than the others even though they measure just as good.
I wish I could explain this. I've read on it and it makes sense when I read it, but I'm no expert here.
I should have remembered... I have a center channel which was built with a Hypex 3-way plate amp and I honestly can't tell it apart from my passive L and R speakers. All 3 may be too laid back but they definitely do not have any mid to treble harshness. I think @AllenB may be onto something when he surmises designers are using DSP to hammer bad ideas into place. In my next DSP adventures I'll attempt phase unrolling.... hope that doesn't ruin the sound! 🤣
If you wanted to make a speaker with "perfect" transient response, which requires flat phase response, you would use a FIR filter.
With IIR or passive analog filters, each "pole" introduces 90 degrees of phase shift, which some designers prefer to avoid.
I won't attempt to explain the math, but simply put, with FIR filters, their output phase, and the speaker's phase can be flattened regardless of the crossover slope used.
That uses a lot of high speed processing compared to an IIR filter with similar slopes.
Processing power used to cost a lot of $$, but those days are over.
It's almost funny looking how much more we paid for processing in the early 1980s compared to a chip with 4000 times the speed now.
See Post in thread 'An exercise in converting a speaker to time-phase coherent'
https://www.diyaudio.com/community/...er-to-time-phase-coherent.345669/post-7942981