KEF LS50 vs. Revel Performa M105

[shkumar4963]

1 kHz is just slightly higher than an octave above the open A string on a violin. Lots of instruments play higher than that. Clarinet, Flute, oboe, piano, trumpet, even viola (which I play), not to mention harp, triangles, cymbals, and xylophones. Most speakers run into trouble on overtones, not fundamentals, with the exception of deep bass extension.

Deleted

Sent from my SAMSUNG-SM-G930A using Tapatalk

 

[ichigo]

Now regarding this brightness, is it limited to the KEF LS50 or all coaxials in general? Because it seems like Floyd Toole/Revel are implying when they tested some Kef speakers, they had sounded bright/edgy despite having good looking spinorama measurements, and attributed that to coaxials in general having high IMD distortion.

 

[Beave]

I haven't seen any posts or statements from Toole/Revel saying or implying that the KEF LS50s are bright/edgy. I have seen some statements saying they didn't do all that well in blind testing despite having good spinorama measurements, and that was attributed to what people are calling IMD.

I say "what people are calling IMD" because my understanding is that the LS50s had an audible flaw where the listeners could hear the midwoofer driver's movement affecting the output of the tweeter (because the midwoofer is the waveguide for the tweeter).

This is certainly a concern when running the LS50s full range, and when playing bass heavy, and when playing at loud levels that result in large movements of the midwoofer (the tweeter's waveguide).

But I suspect that's a use case that doesn't happen all that much in real world situations. I'd venture many people use them with subwoofers and/or don't play them loudly and/or don't play them loudly with bass heavy content but without subwoofers.

In other words, when highpassed and used with a subwoofer, or when not played so loudly, or when not played with bass heavy content, the LS50s midwoofer doesn't need to move as much, so it stays a more consistent waveguide to the tweeter.

 

[Beave]

That also means that any of the 3-way KEFs (such as the R300, R500, newer R3, R5, R7, etc) don't have this problem (or have it to such a small degree as to be inconsequential), because they have a dedicated woofer to cover the lower range, leaving the midrange free to cover a range that doesn't require it to move much at all.

 

[KenM10759]

I have LS50's, as well as many other KEF speakers with various iterations of the Uni-Q driver. Most are used in conjunction with a subwoofer. I've never experienced any situation where movement of the midrange driver affected the tweeter, but theoretically I'm sure it could happen in situations described above.

In all my use of them, the benefits of the coherent timing far outweighs the potential negative effects possible in those unusual conditions.

I did find the LS50's to sound better for my ears when I eliminated toe-in. When I first put them on my desk I had them pointed at my head. Soundstage was too pinpointed, and they did seem a touch bright. Straightening them out solved everything, and it's a wonderful world again.

 

[Steve81]

I say "what people are calling IMD" because my understanding is that the LS50s had an audible flaw where the listeners could hear the midwoofer driver's movement affecting the output of the tweeter (because the midwoofer is the waveguide for the tweeter).

This is certainly a concern when running the LS50s full range, and when playing bass heavy, and when playing at loud levels that result in large movements of the midwoofer (the tweeter's waveguide).

There's that, and the fact that the LS50's midwoofer is relatively limited in both Sd (~75cm^2) and Xmax (~3mm) anyway. Under that kind of use, it's not hard to imagine the LS50's running out of gas quickly, particularly around 100-125Hz, where the port contribution is small and the displacement requirements are non-trivial. In comparison, the Scan woofers in the BMRs have roughly 4x the rated linear displacement (Sd of 145cm^2, Xmax of 6.5mm), which gives it a substantial advantage in that respect.

 

[PENG]

I can't wait to compare the LS50 and BMR side by side, and plot some graphs too, but that has to wait, hopefully before snow melts.

 

[ichigo]

I haven't seen any posts or statements from Toole/Revel saying or implying that the KEF LS50s are bright/edgy. I have seen some statements saying they didn't do all that well in blind testing despite having good spinorama measurements, and that was attributed to what people are calling IMD.

I say "what people are calling IMD" because my understanding is that the LS50s had an audible flaw where the listeners could hear the midwoofer driver's movement affecting the output of the tweeter (because the midwoofer is the waveguide for the tweeter).

This is certainly a concern when running the LS50s full range, and when playing bass heavy, and when playing at loud levels that result in large movements of the midwoofer (the tweeter's waveguide).

But I suspect that's a use case that doesn't happen all that much in real world situations. I'd venture many people use them with subwoofers and/or don't play them loudly and/or don't play them loudly with bass heavy content but without subwoofers.

In other words, when highpassed and used with a subwoofer, or when not played so loudly, or when not played with bass heavy content, the LS50s midwoofer doesn't need to move as much, so it stays a more consistent waveguide to the tweeter.

Seems like all the high end coaxials I've seen recently have high IMD distortion, such as the Genelec 8351, which is a 3-way and is no by no means a small speaker.

 

[tyhjaarpa]

I

Now regarding this brightness, is it limited to the KEF LS50 or all coaxials in general? Because it seems like Floyd Toole/Revel are implying when they tested some Kef speakers, they had sounded bright/edgy despite having good looking spinorama measurements, and attributed that to coaxials in general having high IMD distortion.

I have not experienced that with any Kef speakers. I own R500 and R200c and have had those for years. Have been running them with and without subs. I have also listened to LS50 without sub on moderate sound levels and have not experienced this. Have also heard their Q, Reference and new R series and again I haven't experienced this. So quite safely I can say that if you are not pushing the speakers you should be fine. By moderate sound levels I mean 75-80db and for normal home use that should be enough.

 

[KenM10759]

Seems like all the high end coaxials I've seen recently have high IMD distortion, such as the Genelec 8351, which is a 3-way and is no by no means a small speaker.

Could you please cite (post links to) any of the specific reviews and/or measurements of these specific "high end coaxials" you've seen recently with high IMD distortion?

There are well-done reviews with scientific measurements done by SoundStage Networks on KEF LS50, R500, R100, and Reference 1 and 3. None of these reviews reveal any high levels of any kind of distortion. In fact, KEF has worked hard to reduce distortion and is one of the very few speaker manufacturers that publishes that spec, in both speaker specifications and on white papers they publish regarding many individual speakers.

Scroll down to "K" and check my statement for yourself.

https://www.soundstagenetwork.com/index.php?option=com_content&view=article&id=16&Itemid=140

 

[PENG]

Seems like all the high end coaxials I've seen recently have high IMD distortion, such as the Genelec 8351, which is a 3-way and is no by no means a small speaker.

Where can I find the IMD bench test data for the LS50? I can't hear any objectionable distortions from mine even without the sub, may be because I rarely listen to average spl of more than 75 dB.

 

[KenM10759]

The most info I've found on the LS50 is the white paper they published.

https://www.google.com/url?sa=t&source=web&rct=j&url=https://www.kefdirect.com/pub/media/wysiwyg/documents/ls50/ls50_white_paper.pdf&ved=2ahUKEwjYhMaM8PneAhUomeAKHZyJBWsQFjAAegQIBhAB&usg=AOvVaw3pexdVbUmRZ1keFfxKwfWs

 

[PENG]

The most info I've found on the LS50 is the white paper they published.

https://www.google.com/url?sa=t&source=web&rct=j&url=https://www.kefdirect.com/pub/media/wysiwyg/documents/ls50/ls50_white_paper.pdf&ved=2ahUKEwjYhMaM8PneAhUomeAKHZyJBWsQFjAAegQIBhAB&usg=AOvVaw3pexdVbUmRZ1keFfxKwfWs

Thank you, but I read that white paper long time ago and don't recall they published any IMD measurements.

 

[Beave]

The NRC measurements requested by Soundstage don't include any IMD measurements.

Even if they did, it would only be a tiny snapshot. There are infinite possibilities when it comes to IMD measurements, and probably only a tiny fraction would capture an issue that is audible.

(Even the harmonic distortion measurements shown in Soundstage reviews should be taken with a grain of salt, as they don't correlate well at all to audibility.)

 

[Beave]

What's more, the issue with the two-way KEFs isn't true IMD, as I alluded to in my earlier post.
A simple on-axis IMD test wouldn't capture the issue.

 

[zieglj01]

I would like to put the LS50 against my M105 -- however, time will tell

 

[aarons915]

I went back and forth a bit with the Revel guys a few months back about the whole IMD distortion topic, I actually used REW to measure IMD in my setup, which was with the Q150 at the time. With an 80Hz crossover and an 80Hz tone, the most I could get was just under 4% IMD distortion and this was with a volume about 15-20db louder than I ever listen to, I estimate about 90-95db in room. Like another poster said, at normal listening levels with a crossover in place, IMD distortion isn't a factor.

 

[Beave]

A tone at 80Hz is probably right in the band of the port output, so there wouldn't be much cone movement.

The problem is at its worst when the cone movement is at its maximum, which would be just below the range of port contribution (say, 30Hz), or just above the range of port contribution (say, 120Hz).

(See port contribution in Figure 3 here KEF LS50).

But, even then, as I pointed out before, this isn't IMD in the traditional sense. An on-axis measurement won't really capture it well.

A better way to capture it might be to give the midwoofer a DC voltage to get it at its max forward excursion, then measure the treble response off-axis curves. Then use a negative DC voltage to get the midwoofer at its opposite max excursion, then again measure the treble response off-axis curves.

Doing so will show a worst case of how the treble response changes as the tweeter's waveguide (the midwoofer) moves in and out.

 

[aarons915]

I'm sure that test will show quite a bit of IMD distortion but I'm looking for more real world conditions that I would actually experience. The port is tuned right around 52Hz I believe so there is still considerable cone movement at 80Hz but you're right that 100-120Hz might be a better frequency to test, based on the NRC measurements, distortion peaks at 100Hz. What about the 2nd tone to check for though? I did 2500Hz in the Q150 because it was the crossover frequency, my reasoning was that it should pick up IMD distortion in the midwoofer and tweeter but I don't know if that's right. I wouldn't mind doing the test again with my LS50's.

 

[Beave]

The NRC graph for THD on the LS50 peaks at 100Hz, but that's due to the nature of the way the data is presented. The THD is shown relative to the output of the speaker - and the speaker's output starts dropping around 100Hz. Relative to the speaker output, the THD is quite a bit higher at 50Hz than it is at 100Hz. However, keep in mind these measurements don't include the output of the rear port. So the bass wouldn't drop off quite as rapidly, but the THD measurement isn't including any port distortions. So it's hard to make much of the THD measurements with a rear-ported speaker (not to mention that, as I've said before, the audibility of the THD is far more complex than the measurement would lead one to believe).

Anyway, as I think I've stated, as far as I know there is no way to capture this coax issue with a single IMD measurement. That's why I'm reluctant to call it a problem of IMD. It's more like intermodulation dispersion than intermodulation distortion.

Generally IMD measurements are done on a single driver. Doing IMD on a multi-way speaker, and setting one tone well in the passband of one driver and the other tone well into the passband of a different driver should give you no IMD at all. In the case of a standard two-way tweeter over woofer, a tone of 100Hz and a tone of 4kHz would put the first tone in the woofer range and the second in the tweeter range, so there should be practically no IMD (assuming infinite slope crossovers). But a tone of 100Hz and a tone of 400Hz *would* show some IMD because both tones are played by the same driver. Likewise 4kHz and 10kHz.

I think you're chasing ghosts trying to capture this complex problem with a simple measurement.