linkwitz lxmini

[abdo123]

Generally that is true as generally the speakers being tested have rising directivity. If the speakers were predominantly constant directivity the result would be different. The LX521 is too bright without the high end shelving filter, just as most constant directivity designs need a little slope to the on axis to avoid sounding bright. Genelec S360, Gedlee Summa etc.

It's not ineffective but it is fraught with danger without a lot of listening testing, measuring, tweaking and repeating. When a speaker has diffraction problems that create off axis issues, sometimes the only option is to drop the on axis in that area to compensate and make the speaker sound better. There are elements of room behaviour that can behave in a similar way. Addressing these can improve the sound even if the direct sound is not so flat anymore. Drawing a curve and forcing a measurement to it without knowing why is a bad idea, but it's not the only way.

Could you please substantiate any of your claims?

 

[fluid]

Could you please substantiate any of your claims?

Certainly but providing proof to every single statement would most certainly derail this thread. Is there anything in particular that would help without needing 20 posts of explanation?

 

[abdo123]

Certainly but providing proof to every single statement would most certainly derail this thread. Is there anything in particular that would help without needing 20 posts of explanation?

Research that suggests we should prioritise steady state (in-room) measurements or the predicted in-room response.

None of the speaker designs you mentioned have an on-axis dip by the way, the S360A, since it’s fresh in memory, has a very aggressive tilt (-10 dB at 10Khz) in predicted in-room response.

This is unprecedented in any Genelecs designs, and nonetheless they kept the on-axis flat. Whatever shelving behaviour the measurements show is too benign and can be explained with anything from temperature differences to unit to unit variation. The measurements that Genelec advertises for the speaker are +/- 1.5 dB flat.

 

[fluid]

Research that suggests we should prioritise steady state (in-room) measurements or the predicted in-room response.

If that's what you got from what I wrote then you misunderstood.

None of the speaker designs you mentioned have an on-axis dip by the way, the S360A, since it’s fresh in memory, has a very aggressive tilt (-10 dB at 10Khz) in predicted in-room response.

I didn't say they did, I said that they have a gentle downslope to the high frequency on axis response instead of keeping it flat, and they do. You have taken two statements and added them together.

The second separate statement was that there are circumstances where an on axis depression could be preferable to a power response peak, a situation where it is a matter of balancing the priorities.

You can read Lipshitz and Vanderkooy for some research on that topic, from the abstract "In particular, it appears that neither the direct sound alone nor the reverberant sound alone but rather some combination of the two determines the perceived spectral balance."

https://www.aes.org/e-lib/browse.cfm?elib=11454

 

[abdo123]

If that's what you got from what I wrote then you misunderstood.

reading it again you're right, I seem to have misunderstood, but at the same time there doesn't seem to be a clear message from what you said either other than 'exceptions may or may not exist'. Exceptions will always exist, that's hardly valuable for any sort of discussion.

I said that they have a gentle downslope to the high frequency on axis response instead of keeping it flat, and they do.

The Geddess loudspeakers do, the Dutch & Dutch 8C doesn't. Neither does the Revel M16 or the KEF LS50 Meta. So it's hard to draw conclusions that way. By the way I wouldn't consider the S360A a constant directivity speaker, there is a gentle slope to its directivity.

You can read Lipshitz and Vanderkooy for some research on that topic, from the abstract "In particular, it appears that neither the direct sound alone nor the reverberant sound alone but rather some combination of the two determines the perceived spectral balance."

https://www.aes.org/e-lib/browse.cfm?elib=11454

Thank you for sharing i'm going to look at it, but seems like we're back to flat on-axis with smooth directivity?

 

[fluid]

reading it again you're right, I seem to have misunderstood, but at the same time there doesn't seem to be a clear message from what you said either other than 'exceptions may or may not exist'. Exceptions will always exist, that's hardly valuable for any sort of discussion.

Maybe you need to keep reading as I said nothing like that.

The Geddess loudspeakers do, the Dutch & Dutch 8C doesn't. Neither does the Revel M16 or the KEF LS50 Meta. So it's hard to draw conclusions that way. By the way I wouldn't consider the S360A a constant directivity speaker, there is a gentle slope to its directivity.

The 8c has about 3dB DI rise and a subtle tilt from low mid on. The M16 has a tilt down, the LS50 Meta has the 500 to 1000Hz higher than 10K. The right place to tilt the response varies. They all have smooth directivity but do not have ruler flat on axis responses.

Thank you for sharing i'm going to look at it, but seems like we're back to flat on-axis with smooth directivity?

Not at all, the directivity slope of the speaker and the overall pattern will determine what sounds right. There is no universal single right answer for everything.

 

[Tangband]

This thread makes me wanna build something similar to this, but maybe with other drivers. The Peerless tc9FD18 is far cheaper and have less distortion at 1,5 kHz than the seas unit. A seas er18rnx is also probably better ( I have two of those from the HYBRID building ) .

Playing around with EDGE in the loudspeaker program Basta ! , there is also clearly some possible diffraction at 8-12 kHz in frequency response if no baffle is used for the 3 inch unit.

The cardioid response at 700 Hz is a very interesting thing that I want to hear how it works.

LXmini Clone

As proof of the concept, I decided to make a clone of the LXmini speaker. The results are excellent both audible and measurable. The drivers I used are Mivoc swm 68 and TC9FD18-08 which in my opinion is a far better driver than Seas FU10RB (H1600-04). I ordered the Seas L16RN-SL (H1480) waiting...

www.audiosciencereview.com

 

[MKW]

So, now we have the first complete measurements of the LXmini and the results are very interesting - to keep the post from becoming page long, just a few notes....

In two posts (see here and here) I tried to analyze the LXmini a bit based on S. Linkwitz (SL) original measurements. Some things are confirmed by Erin's measurements, others differ strongly.


Comparison on-axis measurement SL and Erin's Klippel
Both measurements use the center of the 4'' full-range driver as the measurement axis. The measurement of SL was probably made at 1m distance, the Klippel measurement calculated at 2m distance. Because of the small distances of the drivers, the different measuring distances has nearly no influence on the shape of the on-axis FR.
Therefore, the differences in frequency response are astonishing:
View attachment 222054
I don't know if SL made any drastic changes after his published measurements, but the 3dB differences in the 1.8-4kHz range are dramatic sound-wise.
If a PEQ is set in the MiniDSP in the range around 10-16kHz, it should be checked - it seems that the break-up frequency of the driver is not hit cleanly.

The Youtube comments on Erin's video say that some find the speaker harsh in certain frequency ranges.
This could have something to do with the somewhat too present 2-5kHz range that Erin's on-axis measurements show - in SL's measurements, this range was lowered 3dB.
This is reflected in the performance of the ER, SP and PIR.

PIR:
View attachment 222056


"The response in the vertical plane should be similar, because of closeness of the acoustically small sources"
This quote from SL with regard to the LXmini is true on the whole, but rather not in detail.
Horizontally and vertically, you can observe a kind of dipolar radiation behavior ("hypercardioid") in the range of 1-8kHz. However, the vertical radiation has additional side lobes (see pink ellipses). Mainly due to the upward radiating driver. The hypercardioid radiation and the side lobes could be the explanation for the huge sound stage that many hear.
View attachment 222077
This also results in a lot of sound energy being emitted in the vertical range 30-60° (see purple ellipse) - which can be easily seen in the unusual slope of the "ceiling bounce" in the early reflections diagram.
View attachment 222078
The result is a significant increase in the sound power output in the range 400-1000Hz - can be seen in the above PIR graph. Whether this has an effect on the sound is difficult to say (speculation based on the measurements: possibly the LS could sometimes sound a little tinny, nasal or not).
The "problem" of the increased sound power output only becomes really visible when the vertical measurements are included and can easily be overlooked if only the horizontal FR are considered.

It is not surprising that such an LS does not tolerate high SPL levels (sealed speaker and 4'' mid-high) and is confirmed by high multi-tone distortion - but high SPL is not the goal of the LS.

@ctrl "(sealed speaker and 4'' mid-high)"...
Although it doesn't affect the validity of your good point, bass driver is not sealed - review build instructions. Bass driver/enclosure is effectively an infinite transmission-line.

 

[fluid]

@ctrl "(sealed speaker and 4'' mid-high)"...
Although it doesn't affect the validity of your good point, bass driver is not sealed - review build instructions. Bass driver/enclosure is effectively an infinite transmission-line.

I don't have the build instructions for the LXmini but this is SL's own words on his website.

"The woofer sits at the top of a sealed pipe, which is so heavily stuffed with absorbent material that the two poles of the woofer highpass become real and roll-off is at less than 12 dB/oct. The stuffing attenuates higher frequency pipe resonances and stiffens the restoring force to the driver. The driver cone undergoes large displacements, the electro-mechanical structure is non-linear and the large amount of mechanical damping is thought to stabilize the response."

 

[MKW]

I've had LXmini pair years now. B&W sold, Tekton DI's sold, Martin Logan e-stats sold, Elac's moved to the garage (too good of a value), GE Triton 3+ moved to HT system (too good of a value). About a year ago, I had the opportunity to hear YG Carmel, they sounded wonderful but like listening to speakers (obvious where the sound is coming from - no 'magic'). Earlier this week I had opportunity to demo another supremely great speaker (at the custom-builder's demo room), again an amazingly capable speaker with beautiful cabinet and highest-quality drivers and crossovers. Again, sounds like listening to speakers - no 'magic'. What surprised me the most was that I could hear the cabinet edge diffraction. Its like hearing amp noise - once you hear diffraction, you can't un-hear it.

There are several aspects of LXmini that stand-out, here's my take of the pro's:
- Because there are no baffles, they disappear immediately. No need to close your eyes, or wait for it (like some speakers that disappear; although, most don't).
- Because they use DSP for crossover and filtering, you can change their sound signature, tonal attitude, etc. whenever and as frequently as you desire (for different music genres, different playback volumes, or just because). This benefit CANNOT be over-valued. Personnaly, I wouldn't have another serious pair of speakers without this ability.
- Because of cardioid/omni dispersion, they image like crazy without spurious sidewall or font-wall reflections. Easy placement, no further into the room than typical speakers and sound great everywhere in the room - goodbye ML e-stats, take your sweet spot with you.

Cons:
- They do not play particularly loud. They have an intimate, natural, delicate, emotional sound that puts you on the stage with the performer breathing on you. If you want to be in the front row head-banging with speakers blaring in your face - get Tektons.
- Their 'form follows function' construction materials and looks might not appeal to everyone. Like the BMW GS, they aren't designed to look classy - they're designed to perform.
- Requires 4 channels of amplification (to retain DSP tunability). Thankfully, good multichannel amps are widely available due to the proliferation of HT amps. I use a pair of 3-way Hypex plate DSP/amps to crossover/PEQ/drive LXmini and stereo-pair of OB-subs.
- They do not play full-range (20-20k). Sub(s) required to fill-in below 45Hz (this very typical of all small form-factor speakers). I'd recommend open-baffle or alternately REL (for accuracy - needed to match LXmini). I use Linkwitz free design of Phoenix (alt) subs, each w-frame with two 12" Peerless XLS drivers for excellent results.

 

[MKW]

I don't have the build instructions for the LXmini but this is SL's own words on his website.

"The woofer sits at the top of a sealed pipe, which is so heavily stuffed with absorbent material that the two poles of the woofer highpass become real and roll-off is at less than 12 dB/oct. The stuffing attenuates higher frequency pipe resonances and stiffens the restoring force to the driver. The driver cone undergoes large displacements, the electro-mechanical structure is non-linear and the large amount of mechanical damping is thought to stabilize the response."

That quote is wrong on this point (I don't know why cause I didn't write it) - read the build plans or ask in Orion owners forum.

 

[fluid]

That quote is wrong on this point (I don't know why cause I didn't write it)

I know you didn't write it, the designer of the speaker himself did.

- read the build plans or ask in Orion owners forum.

As I stated I can't read the plans, they are not public domain so if there is something in them that you think contradicts the above it would be more useful to share it. I don't need to read the plans to know that the bass section is a driver mounted to a piece of wood glued to a plumbing coupling attached to a sealed pipe with a lot of stuffing in it, as a former Orion and LX521 owner/builder I read every post on the OPLUG for many years.

It is what it is, the measurements show what happens, what does the semantics of sealed vs infinite transmission line offer to anyone as explanation?

 

[MKW]

Agreed - carry on. Not my perogative to argue or convince. We are flashes of light in the darkness - no time to waste.

 

[Tangband]

This is an interesting thread. This is apparently not a speaker that will measure especially good in a klippel test. The 4 inch:er radiates backwards and are filling up the stereosystem flaws. This is not a speaker that measures good with only one speaker - its made for stereo listening in a real room that have reflective walls.

Some writers in this thread suggests ( between the lines ) that the direct sound from the speaker and the power response at the listening position is equally important . Its not - the direct sound from the speaker are more important than most of the people believe in this forum, and this specific loudspeaker is clearly also made to be listening to, in stereo in a real room.

The precedence effect in the brain always gonna favour the first 5 ms sound coming from the speakers, and the brain attenuates all other sounds coming after those 5ms as much as - 10 dB , very unlike a microphone.

 

[jmf11]

I built my LX-Minis in 2016. I was sold by Siegried pitch about stereo imaging and design for easier integration in normal rooms. I have not been disappointed up to now. There are not so many DIY speakers with Cardioid feature in the place.

When I discovered the hype around the Kii3 and D&D 8c, I looked at discussions in depth. Then tought: Ok I would love have those, but did I really tried to optimize my LX-Mini usage => no... except few Rew measurements.

What do I like in the Kii3 and D&D 8c ?

• Cardioid: LX-Mini OK
• DSP to adjsut to room: LX-Mini OK
• Active: LX-Mini OK
• Bass extension: can add a subwoofer (3 designs for LX-Minis)

By design, the off axis response is OK and they repond well to equalization: good

Cost: About 300-400€ for the speakers and maybe 150€ for the 2 TAS3251 based amplifiers (I was first running with a stm32F7 doing the DSP and 2x FX-Audio D802 digital input amps). Challenge: get 95% near of D&D 8c above 40 Hz for 5% of the cost (and some elbow grease) ;-)

Last, and very specific: they are light, and very easy to move around. Quite a advantage in the room shared with familly, for someting that is not used a so big fraction of the time.

JMF

 

[Lingwendil]

Did this ever lead to measurements from Amir?

I'm getting ready to build a pair of Vifa/Tymphany TC9FD based LXmini similar speakers, and was curious how this went.

 

[jmf11]

Klipper measurements of LX-minis have been done by Erin: https://www.erinsaudiocorner.com/loudspeakers/linkwitz_lx_mini/

 

[HarmonicTHD]

Klipper measurements of LX-minis have been done by Erin: https://www.erinsaudiocorner.com/loudspeakers/linkwitz_lx_mini/

Yep. Not really something to write home about. But of course, if one values the DIY aspect.

 

[Lingwendil]

I saw Erin's review and measurements- those combined with cdennlers work on them made me think that the FR driver is an issue here- with the Vifa being a good candidate to improve the design. Referring to this- https://sites.google.com/site/cdenneler/home/lxc