Time Domain measurements?

[Joachim Gerhard]

I recently talked to Peter Lyngdorf on this exact topic and he also referred to the incredible SPL levels inside a cabinet and he suggested that there is no good way of damping that enough so that it does not ricochet back through drivers. (I think there was talks on 120+ dB)
That's one of his reasonings for going dipole in the Steinway Lyngdorf speakers

Dipoles have other problems like low sensitivity and placement issues. Linkwitz last design sounded very good but did not reach remotely the soundpreasure to mimic a big orchestra.

 

[Absolute]

Did he really say that or are you misinterpreting?
How can you determine if a low-level low-Q resoance at say 200Hz is a cabinet resonance from a Spinorama?

You know the response of the drivers before you placed them in a cabinet and you have probably simulated the response with a cabinet, so unexpected differences is likely to be because of the cabinet.

Edit; Me as a user can't possibly know if the resonance in the spin is from the cabinet, port, driver or any other parts that can resonate. I can only see the resonance and decipher if it's in the audible range or not. The microphone doesn't care where the resonance comes from, and frankly, neither do my ears

Yes, he wrote that in a forum post either here at ASR or at AVS forum. Logically it makes sense. Resonances will be visible in the frequency response if audible because it's like a disturbance in the force. (He didn't write that ending, sadly.)

 

[haraldo]

Dipoles have other problems like low sensitivity and placement issues. Linkwitz last design sounded very good but did not reach remotely the soundpreasure to mimic a big orchestra.

I always kind of fancied transmission line designs, mainly for the woofers, but there are some stuff out there that use TL for small 3" fullrange drivers in a line source... (Adyton). These work really well!

Isn't a properly designed TL a way also for a woofer/mid to kill the backwave, so it doesn't backfire to the same extent through the driver?

 

[tuga]

You know the response of the drivers before you placed them in a cabinet and you have probably simulated the response with a cabinet, so unexpected differences is likely to be because of the cabinet.

So the answer is you can't.
Just look at ASR's measurements, they're proof that the Spinorama is manifestoy insuficient which is why you need independent driver and port measurements FR, THD, CSD, now IMD and soon cabinet.

To believe that anything audible will show in a Spinorama is pure delusion.

 

[tuga]

Yes, he wrote that in a forum post either here at ASR or at AVS forum. Logically it makes sense. Resonances will be visible in the frequency response if audible because it's like a disturbance in the force. (He didn't write that ending, sadly.)

A small peak in the FR, which can be almost undetectable with 1/6 octave smoothing, can ring for a bit. FR plots don't show what happens in the time-domain. Again, look at the recent IL10 thread.

Of course the easiest way is to decide that anything that doesn't show in a Spin doesn't matter...

 

[tuga]

I recently talked to Peter Lyngdorf on this exact topic and he also referred to the incredible SPL levels inside a cabinet and he suggested that there is no good way of damping that enough so that it does not ricochet back through drivers.

How about this:

 

[haraldo]

How about this:

Isn't this most of all a showcase from B&W?, the manufacturing cost is so high that it's probably not possible to sell. I never seen them anywhere for commercial sales.

As far as I know Richard Vandersteen also uses TL for all the drivers from midrange and upwards in most of his upscale products. Well, they are not exactly cheap too....

If you look to PMC, they only use TL for woofers, so that does not count, I assume....

 

[Absolute]

So the answer is you can't.
Just look at ASR's measurements, they're proof that the Spinorama is manifestoy insuficient which is why you need independent driver and port measurements FR, THD, CSD, now IMD and soon cabinet.

To believe that anything audible will show in a Spinorama is pure delusion.

A small peak in the FR, which can be almost undetectable with 1/6 octave smoothing, can ring for a bit. FR plots don't show what happens in the time-domain. Again, look at the recent IL10 thread.

Of course the easiest way is to decide that anything that doesn't show in a Spin doesn't matter...

Speculating about cause and effect is not proof, it's guessing. Important distinction.

I think the rationale is something like this; if you play a tone at 90 dB continuously and the cabinet is ringing at that frequency, you can't hear the ringing unless the overall spl increases due to the ringing. Fair enough, that makes sense.
If you stop the continuously tone and the cabinet keeps ringing for a little time, you might be able to hear it depending on the time (due to masking) and spl of the ringing. That also makes sense.

When we sweep we measure more frequencies in short time and any ringing will have a Q that extends beyond one exact frequency. If that ringing is strong enough, it will make the frequencies before and after the peak of the Q rise a little resulting in a frequency response bump over a certain area. If it does, it's audible. If the ringing is not strong enough to increase the spl in any of the surrounding frequencies during that sweep, it's simply being masked by all the other louder frequencies being played.
How much it needs to alter the SPL to become audible will depend on the Q, frequency range etc.
This certainly makes sense if there's other stuff playing. It stops to make sense if we stop playing other masking frequencies.
Do we agree this far?

All distortions are frequency response (stuff that gets invented by speaker/electronics) and all time-related noise is also frequency response. We hear frequency response, so everything that makes noise at a loud enough spl will be heard if the circumstance allows for it.
Does the spinorama tell us everything we need to know whether or not some particular noise might end up audible in all circumstances?
Highly doubt that.

 

[Matias]

I always kind of fancied transmission line designs, mainly for the woofers, but there are some stuff out there that use TL for small 3" fullrange drivers in a line source... (Adyton). These work really well!

Isn't a properly designed TL a way also for a woofer/mid to kill the backwave, so it doesn't backfire to the same extent through the driver?

View attachment 73424

What is the manufacturer and model of those speakers please? Just curious.

 

[tuga]

Isn't this most of all a showcase from B&W?, the manufacturing cost is so high that it's probably not possible to sell. I never seen them anywhere for commercial sales.

As far as I know Richard Vandersteen also uses TL for all the drivers from midrange and upwards in most of his upscale products. Well, they are not exactly cheap too....

If you look to PMC, they only use TL for woofers, so that does not count, I assume....

£57,750 per pair. A lot of money but not really expensive as high-end goes – roughly the same as a pair of BeoLab 90, TAD Ref1 or Vandersteen 7 mk2, cheaper than a pair of Vivid G1 or Wilson Alexx.

 

[haraldo]

What is the manufacturer and model of those speakers please? Just curious.

That is an Adyton iMagic 1.5, made by Geir Fredriksen (Norway) ... there is some small 3" fullranges in a TL, these fullranges go completely without any components in series, there is a Ribbon that only has a capacitor in series. There are two 8" woofers set back-to-back also in a TL. You can't buy these anymore.

There is a friend of mine that has these speakers and also an Audio Physic Rhea II subwoofer. I like them a lot

 

[haraldo]

£57,750 per pair. A lot of money but not really expensive as high-end goes – roughly the same as a pair of BeoLab 90, TAD Ref1 or Vandersteen 7 mk2, cheaper than a pair of Vivid G1 or Wilson Alexx.

I just recently listened to speakers costing way more than that, so it's pretty steep but not crazy insane from a hi-end perspective, but you have to look at this towards competition .... there are some pretty nice stuff in the € 25.000 upwards range
Edit: and € 25.000 is not exactly pocket change

 

[tuga]

I think the rationale is something like this; if you play a tone at 90 dB continuously and the cabinet is ringing at that frequency, you can't hear the ringing unless the overall spl increases due to the ringing. Fair enough, that makes sense.

I must have not made myself clear but I wasn't referring specifically to cabinet resonances.

All distortions are frequency response (stuff that gets invented by speaker/electronics) and all time-related noise is also frequency response.

But if what makes a resonance is the fact that it rings as the remaing spectrum decays uniformly then what was a mostly flat frequency response on axis at full level now shows peak which becomes more prominent (from the remaing spectrum).

Resonances may or may not be noticeable depending on the frequency, the musical programme and the listener.




As for cabinet contribution, Hi-Fi World uses CSD plots to observe what they call "box boof":

www.hi-fiworld.co.uk/index.php/loudspeakers/69-tests/95-decay-spectrum-waterfall.html

Wheter it's effective or not I cannot say.

 

[Joachim Gerhard]

I always kind of fancied transmission line designs, mainly for the woofers, but there are some stuff out there that use TL for small 3" fullrange drivers in a line source... (Adyton). These work really well!

Isn't a properly designed TL a way also for a woofer/mid to kill the backwave, so it doesn't backfire to the same extent through the driver?

View attachment 73424

I know a lot about Adyton. In the 90th I owned Adyton 50% with Rolf Olstad. I still have a Temper and Cordis 1.5 that I restored like new.

 

[Joachim Gerhard]

How about this:

It works. One of the great designs of our day, especially in a active version with an Audionet active crossover.

 

[Joachim Gerhard]

What is the manufacturer and model of those speakers please? Just curious.

Modern Adyton. You can hear that at Oslo HiFi Center. In my book it is so so.

 

[Joachim Gerhard]

I must have not made myself clear but I wasn't referring specifically to cabinet resonances.



But if what makes a resonance is the fact that it rings as the remaing spectrum decays uniformly then what was a mostly flat frequency response on axis at full level now shows peak which becomes more prominent (from the remaing spectrum).

Resonances may or may not be noticeable depending on the frequency, the musical programme and the listener.




As for cabinet contribution, Hi-Fi World uses CSD plots to observe what they call "box boof":

www.hi-fiworld.co.uk/index.php/loudspeakers/69-tests/95-decay-spectrum-waterfall.html

Wheter it's effective or not I cannot say.

A waterfall is hard to Interpret. I use a Blackman windowed transient response with 4 cosinus sines to look at the details. An ETC of that gives some clues.

 

[Joachim Gerhard]

Do a high resolution impedance sweep. There you see everything. Driver issues like surround resonances, mechanical cabinet resonances, acoustic resonances... you Name it.

 

[haraldo]

Modern Adyton. You can hear that at Oslo HiFi Center. In my book it is so so.

I have heard most the Adyton gear and talked a lot to the speaker designer (Geir Fredriksen) lately. Adyton is one of the most sold hi-end brand around here and it has unreal dynamics and resolution. I have, however, never 100% got to hold with the line sources... I think generally there is something unreal about the large size of the soundstage... Although I have heard one single implementation where this unnatural soundstage don't seem to be there, One of the employees at Oslo hi-fi center spent something like a year on integration and in this one case it works really well! .... but it's still not fully my piece of cake. Except for the iMagic 1.5 that I showed earlier, but it's not being sold anymore.

The one and only procduct available now is the Adyton Gracili below, featuring only 2.5" fullrange drivers (looks like fountek) and a ribbon.
These speakers use Valchromat panels for the cabinet and aluminium for the front, it is claimed that Valchromat has better behaviour for cabinets than mdf or hdf.

Geir Fredriksen is a very nice guy, very approachable and willing to share about mostly anything....

The speakers roll off naturally around 100 to 120 Hz (depending on model) There is also no crossover components in series with the main drivers; There must be two extremely capable subwoofers to match. If the subwoofer integration is not 1000% perfect it all falls apart. Fredriksen state that Adyton iMagic speakers are real phase coherent, but I never saw measurements proving it;

Those that never heard a linesource like this always seem to be shocked.... However, I prefer a speaker as close to point source as possible

Below some demo setup and Geir himself. I am, however, not so sure about this thing, the large REL stack can not be crossed high-enough to mate with the main speakers, so there is a small sub in bewteen that fill in between the subwoofer stack and main speakers .... strange setup to me

Adyton electronics is not sold anymore... The Cordis amps were really nice

There is also a spin-off, Audio Insight, the main product line is power conditioners, I got one of these myself, an IsoVolt 3K; I think it made a significant difference to me, taking away lots of "hash", makes music more smooth and nice and organic (Yes, it is NOT my imagination)

 

[KSTR]

Do a high resolution impedance sweep. There you see everything. Driver issues like surround resonances, mechanical cabinet resonances, acoustic resonances... you Name it.

+1. This way of analyzing a speaker (preferably the raw driver already mounted in the enclosure) is much overlooked and sadly, most published plots (manufacturers and magazines) have too low resolution and too much smoothing.

Actually, one can use the exact same techniques used for acoustic measurements and post-processing for electric measuremens as well. Obtain a high resolution impedace impulse response via log-sweep and convolution, and display it as mag+phase bode plot, waterfalls or S-transforms, wavelet analysis, disortion analysis (the IR contains the distortion components as well).
With modern computer-aided design tools, if you're about to design a passive XO and response shaping, you have to do that impedance IR anyway, as you have to do the SPL measurement.

I found that the type of drive (source impedance) plays a role, for the fine-print, in that it determines the excursion vs. frequency behavior, adding some skew to excursion vs frequency during the test. With high-Z drive and higher test level the driver may sky-rocket beyond its Xmax at any strong resonances, and when dialing down the level to avoid that the other frequencies are tested at very low levels compromsing signal-to-noise ratio.