GRIMM Audio LS1c & SB1 DSP Speaker Review

[Purité Audio]

The 8C is undoubtedly a top speaker and has its advantages. But the wide, controlled dispersion characteristic also has its charm, if you like it. Finally, don‘t forget that the LS1 is now a fairly old design. But as you can see, it's still anything but bad. Rather the opposite. It would be particularly interesting to compare the individual variants to get an idea of whether there are any significant differences. Perhaps aa LS1a can hardly be distinguished from the other versions even in the double-blind test?

We had the original version and when they were released the Berylium/ motional feedback/ hi-matt expensive enclosure version well I didn’t have both versions at the same time but I wasn’t shocked by the difference when the expensive pair arrived.
Keith

 

[Heinrich]

The GGNTKT M3 also has a much narrower in room slope which almost always gets forgotten in assessments. It is almost nearing flat on axis in terms of PIR

Like me, you can listen for many years through loudspeakers that radiate quite narrowly. For example, the “econowave” 12 inch bass/midrange plus a 90x40(60) horn, which has long been popular in DIY circles. A brief switch to very wide dispersion real bookshelf speakers then made a big difference. It was brighter, but also much softer in sound. After moving the loudspeakers to a room with a long reverberation time, it was not a great success. The details were largely retained, but the crispness was no longer there.

12 inch plus horn, then you can also ask how this is designed. The directivity first increases continuously, and then from about 1kHz the directivity is determined by the horn. Usually it increases significantly in the vertical direction - similar to ribbons, only not so blatant. Real Constant Drirectivity specimen don't have that, and they sound different. The grass on the other side of the fence is always the greener. A wide field.

With my previous DIY wide baffle speakers I used tem in an arrangement like the following. The bass felt lean, but also quite satisfying because is was clear, more presenting, less pressurizing - as in real life ;-) That was before the spinorama, though.

 

[Kennyknetter]

If it was half the price and if the sub went half an octave lower and could muster about 15db more output, I'd be impressed.

 

[AudioKC]

Again, they are not the same speaker. D&D has half the beam width of the LS1c:

Those of us who like spatial presentations, won't warm up to it as much as the LS1c.

My understanding of the beam and it's logically tight coupling with horizontal / vertical sound pressure that: is more important how differ sound pressure in tonality on different angles from 0 degree angle, than to have incredible wide beam.

Judging these two picture, yes, LS1C+SB1 is "wide" but non linear in that wide area, which might lead (I did not experience these speakers, so I cannot tell) to funky reflections and WOW effect (as for example in many Elacs).

For some it's something that they are looking for, for others, it's speaker producing sounds waves in the room that never been recorded.

In the end, it's becoming preference (as I understand it). and not an objective choice.

My preference is: wide enough beam with good Horizontal SPL that leads to less errors in room and allows to EQ (correct for the room) speakers.

There is a chance that this mid range Horizontal SPL overlapping in many angels is not audible by human ear, and indeed LS1 is just incredible speaker in any room. So, I would wish to experience such expensive speaker

 

[OlsonSystems]

Not an inquiry, but did you also measure vertical towards up and a little bit to the sides, horizontally?

There's so much theoretical talk about this speaker, if it is beautiful, and how bass shy it is in all its excellence on paper.

But is such a wide dispersion really desirable? It realizes +/-100° (-6dB) roughly, while a KEF R series would do +/-50°. And if I read it right, the LS1 was ousted by the Kii, cardioid etc/, same developer?

It may also raise the question if the more discriminative test listening in mono is as easy as it sounds ;-) Such a wide speaker ;-) would feel better to the ears due to many reflections, less sharp, less empty in comparison to again a KEF R series.

Not the least, to put a speaker into a well filled ;-) bookshelf might have a similar effect as the wide baffle. But that is an antiquarian idea ;-) me thinks.

ps:
- may it be worthwhile to exchange the ceiling bounce with the floor bounce and vice versa; could be done via a software update
- the wiggles in the nearfield bass response in higher frequencies may originate in the residuals from the midrange speaker
- bass response, if meant to be anechoic, is perfect for room integration; distortion taken in percentage will drop by a factor of two or more in-room, and bass mangement will ask for extra subs distributed around the room anyway, as always, and then equalize to individual taste

I did not perform any off axis measurements of my DIY speaker. I did the typical measurements on axis including reversing the phase and adjusting the delay to get the best null for the crossover. With the 4th order LR crossover driver overlap is not going to be a big problem. With a flip of the switch in the DSP an 8th order LR could be substituted in.

Don't read too much into my measurements of my DIY mockup. A few months after I made these measurements I found that my 1982 JVC measurement mic response had developed some new bumps in the response and the calibration curve I created was flawed. Off axis measurements of my mockup would not represent the LS1 as my drivers were offset horizontally from center and the vertical spacing of drivers is larger due to cabinet walls in the way.

I found the bass of my DIY speaker to be very satisfying and just what you would expect from any quality 8" driver in a sealed box.

As I have been building very wide dispersion omni-directional spheres and very nearly omni-directional bipole speakers of late, so I thought of this LS1 as a relatively narrow dispersion speaker.

Bipole (drivers front and back) 20 drivers on sphere 6 drivers on sphere

The LS1 with 29 mm dome, with a waveguide, and an 8" woofer on a wide baffle will present as narrow dispersion relative to say an LS3/5a with a 19 mm tweeter, no waveguide, with a 4" woofer on a very narrow baffle. My narrow dispersion speakers are a pair of JBL 28P with the large waveguide.

It comes down to what you want. A very narrow controlled dispersion, cardioid pattern, in a padded control room adds nothing from the listening room to the playback content and presents a window into the performance space equivalent to opening up a big hole in the wall to a performance happening in the adjacent room with all of it's spatial information unmodified. This is perfect for mixing.

The opposite end of the presentation spectrum is the omni-directional speaker in a lively room ideally playing a performance with little or now ambient information so the presentation is that of the instruments playing at the speaker locations or on a line between them in the acoustic space presented by your listening room.

Most speakers are in between these extremes and most have wild swings in dispersion up and down the scale.

To sample the whole menu leads to accumulation of about ten pairs of speakers to cover the range as I have, or look at the variable dispersion speakers like the B&O Beolab 90, so you can switch modes, front firing, cardioid, omni-directional, with the remote control. I achieved the variable dispersion trick with my 20 driver spheres separating the drivers in the three groups and selecting different phase, delay and EQ with the DSP as the Beolab speaker does.

 

[Vuki]

Yes, that may be. How do audio.de measure?

However they measure, they got Genelec 8361a's frequency response quite similar to Amir's NFS.

KEF R3 also...

 

[Aperiodic]

Thanks for the test Amir! Speaker tests, the most interesting tests you do. At least that's what I think and I suspect many others share that opinion.
____

I leave the price aside. I think the looks were pretty cool BUT why not build two subwoofers/bass boxes that take up the entire space under the speaker, against the speaker stand down to the floor? The space is there so why not fill it with as much bass box volume as possible? Same surface finish on the bass boxes as the speaker. Okay then it will be more or less, regarding looks, like a conventional three-way speaker but still.

OR they could offer different subwoofer modules to (for instance) scale the bass extension and/or output to the room as needed or desired.

 

[Vuki]

They stated to me that their evaluation was non anechoic. Indeed their only measurement shows bass response as "simulated."

You know that NFS measurement is also simulated anechoic...or non anechoic and simulated.

 

[amirm]

You know that NFS measurement is also simulated anechoic...or non anechoic and simulated.

No, it is not simulated. Real measurements are used to compute the anechoic results. Simulation means you used artificial input data which is not the case here. It is no different than using Chirp signal to compute frequency domain graphs. You don't call that a simulation, do you?

 

[Vuki]

No, it is not simulated. Real measurements are used to compute the anechoic results. Simulation means you used artificial input data which is not the case here. It is no different than using Chirp signal to compute frequency domain graphs. You don't call that a simulation, do you?

It's simulated - they use nf measurement to simulate mid/farfield measurement. I'm sure Grimm also used real measurement to simulate anechoic response. That is the usual practice when there is no gigantic anechoic chamber available.
That's what was shown in their white papers.

 

[MattHooper]

I will make a modest prediction : this will be @MattHooper next piece of gear And thank you Amir for an excellent review (and yes, Oh that waterfall plot!!)

Ha! Thanks for the shout out.

However, I’m out of the audio gear purchasing game at this point. Good with what I have.

I’m just doing the audiophile equivalent of scratching a phantom limb in drooling over these speakers.

 

[MattHooper]

Matt,
The Waveform speakers that we have discussed previously had a waterfall plot much like the Grimm's, no resonances. And I always thought that that was a big reason why they sounded so good.
We have been there before.......
Steve

This comment, perhaps might’ve been aimed at me since I am the other Waveform fan around here.

We certainly agree the waveforms were stellar. I would say my current Joseph speakers sound subjectively more clean and pure.

Sometimes I’d like to have my Mach MCs around though.

 

[Thomas_A]

It's simulated - they use nf measurement to simulate mid/farfield measurement. I'm sure Grimm also used real measurement to simulate anechoic response. That is the usual practice when there is no gigantic anechoic chamber available.
That's what was shown in their white papers.

The correct word would be "predicted", based on data.

 

[Heinrich]

It's simulated - they use nf measurement to simulate mid/farfield measurement.

Measured: "as is"
Simulated: "as if"

On NFS see also: https://www.audiosciencereview.com/...g-how-the-klippel-nfs-works.13139/post-393407

It is not that advanced, acually. Science often deals with approximations. Thing is, there are real mathematical methods, they are well based on pretty critical thinking, how to do that in a trustworthy fashion.

The NFS sports such an approximation. The speaker's radiation is described by a well defined set of math/ functions, that then can be applied to any situation. The speaker is modeled in regard to its output, not entirely, but for the most part as needed. By all means, that is measurement. To define a model (aka 'concept') of something, and then apply some values to the model's parameters.

Would my next SUV fit into the garage? Lets conceptualize ;-) an SUV, make a model, o/k with that is has length to it. Lets make a model of the garage, and again it has length. Measure 'length' of both ... You wouldn't tell that such was simulation. Before any (!!) measurement you need that concept/model, otherwise a measurement doesn't make, literally, any sense. Let that sink in, maybe? (Makes you probably guess my profession.)

... once you have the measurement, you may generalize the outcome to a variety of situations. That is simulation. Would it be feasible to invite the SUV for lunch, as it may not fit into the chair to my right side? Different situation, but reasonably similar concept of things? In speakers we do it with the "predicted in-room response". I personally don't think of "PIR" as being successful.

That's the point here on topic: You need to equalize the response of the LS1 anyway, at least in the bass, to what you may find at home in that particular situation. I really do not understand why there is so much fuzz about it. As if us never bought a speaker, get it home and optimize to our needs, to find vast deviation from the prediction. Well?

 

[MattHooper]

That bloody bass cone used to take off big-time in the upper kHz region. I appreciate a speaker like this has a proper high order crossover, but this resonance is still there and the hope always is that if the cone is excited by hf from the tweeter, that it'll be so low down in volume we won't hear it. My amateurish take is still that I'd rather this tin-can resonance isn't there at all and there are some speaker drivers out there that are very clean at the top of their range

I can understand what you mean when you see the measurements of the SEAS woofers as they all seem to have a resonance peak.
But boy oh boy, when they are paired with the proper crossover to me, they produce some of the cleanest sound I’ve ever heard.
(and the waterfall plot from the speaker would seem in in line with this as well)

 

[amirm]

It's simulated - they use nf measurement to simulate mid/farfield measurement.

Nope. Do you think an FFT is a simulation of a real time analog signal? That is also a computed result based on time domain samples and subject to variation by picking its length.

To the extent FFT finds the basis sine waves, NFS finds the (3-D) Henkel functions to describe the sound field. Neither it, nor FFT computation fall in any kind of simulation domain. Just because complex computations are needed to convert sample data to graphs that we like to see (understand), doesn't mean the result is simulated.

Now, if you took some hand created data and fed it to NFS, sure, that would be a simulation. But that is not the case. NFS in my measurements uses 1000 to 2000 sweeps to compute its final results. The data presented is simply the 3-D representation of that 2-D data.

 

[amirm]

On NFS see also: https://www.audiosciencereview.com/...g-how-the-klippel-nfs-works.13139/post-393407

That is the math behind the computation of the field. What is fed to it per above is real measurements. As such, it doesn't remotely make it a simulation.

As I also explained, we use CHIRP log sweep signals which are in time domain to create frequency response measurements. That also involves computation. But just because it does, it doesn't mean at all that it is a simulation. It simply means that we use signal processing to our advantage. In the case of CHIRP signal, we gain SNR. In the case of NFS, we get to use near field data to compute far field. And by performing two parallel scans, we can compute reflection-free response. Computation does NOT mean simulation.

COMSOL is an example of a simulation (software) where you feed it all the physical properties of what you have, and simulate what the outcome might be (using CFT, FEA). It is very difficult system to use and prone to errors if you don't know what are you doing. In sharp contrast, Klippel NFS needs no information whatsoever about the source. Nor does it know the source is a simple driver or a planar speaker. It gets a set of frequency response sweeps and goes to town computing the sound field from it.

Really folks, please learn the basics here before disputing them.

 

[amirm]

... once you have the measurement, you may generalize the outcome to a variety of situations. That is simulation. In speakers we do it with the "predicted in-room response". I personally don't think of "PIR" as being successful.

PIR is an entirely different concept and as I have explained, has nothing to do with design of NFS. It is a software feature where it applies cookbook formula as specified (and originally developed by Harman) to make some kind of prediction of the response in a perfect room.

 

[kyuu]

I personally don't think of "PIR" as being successful.

That's a strange statement. What about it isn't successful? From what I've seen, the PIR matches the actual measures in-room response of the speaker astonishingly well, above Schroeder of course. I've seen numerous examples from people posting their measurements, and even Dirac's measurements of my speakers (which of course are an average over an area) were very close to their PIR.

 

[Heinrich]

PIR is an entirely different concept and as I have explained, has nothing to do with design of NFS. It is a software feature where it applies cookbook formula as specified (and originally developed by Harman) to make some kind of prediction of the response in a perfect room.

Exactly that was my point. When discussing 'measurement' in audio people almost always forget about the underlying concept. What it is good for, how it describes what a speaker "is", and how that thing, given its properties, might interact with the room, and then how that would again interact with the listener.

On me not loving the PIR:

That's a strange statement. What about it isn't successful? From what I've seen, the PIR matches the actual measures ...

I don't think the match is that good. Not the least, I might argue that some people don't measure the actual in-room response according to the standard. But that's an entirely differnt discussion.

Anyway, who in the world would buy a 40k speaker and *not* equalize at least the bass to taste? Why the fuzz about the measurements?