Kef LS60 wireless review by ErinsAudioCorner

[Absolute]

@Muddywaters Excellent stuff, thank you for sharing! Nice room.

It looks to my naked eye that there's compression in the lower midrange as well, but it would be easier to spot if you were to show this normalized to the green one. I forgot how you do that in REW, but perhaps someone remember or can do it for you if you upload the mdat data?

 

[Soniclife]

Did some REW testing to get an idea of when limiting kicks in and how it impacts the frequency response in my room with my LS60/HSU subs combo. I'm running the LS60's 25 inches off the front wall.HSU ULS 15MK2 subs are behind and to teach side of the speakers backed into the corners (GIK double stacked TriTraps) They are spaced 8.6 ft and 8y.2 ft from my ears (measuring position). Room is 14.3 x 16.6 x 8 ft. Speakers are along the 14.3 ft width. KEF App settings: Wall Mode - Off Treble - 0;; Phase - On; Extension - Extra;; Balance - Center; Sub Out - Two (obviously) and Stereo;; Hi Pass - 100Hz (tried some at 120Hz); Low Pass - 135Hz (have run at 100Hz too green trace is 100/100); Sub Gain - 0dB

Room is treated with ceiling mounted GIK 244's, Double stacked TriTraps front corners. One TriTrap back upper left corner another currently standing on the floor butted up to an Ikea cube stack left side 2/3 back. Front side walls have SlatFusors and back right side corner a couple more 24x24's. Back wall has a 5x5 IKEA cube shelving unit 70% full of LP's with two 12x48 Monster Bass Traps mounted at rear wall ceiling interface above the shelving and two 12x48 Monsters with Scatter Plates vertical next to the sides of the back Ikea shelving.

View attachment 338098

Measurements
It's obvious the limiter is kicking in somewhere between the 90-93dB individual speaker test level. Cutting back the tweeter level. The waterfalls show it very clearly too.



View attachment 338100
Waterfalls

View attachment 338105
View attachment 338106
View attachment 338107
View attachment 338110
I'm very much a neophyte running and interpreting REW but the I think the FR isn't too bad excluding the room mode dip in the 100-126Hz range and lower hump. I may play some more and try dropping the Hi Pass to 80Hz see if it fills in some of the dip at all. Hopefully no limiting too.

The limiting is clearly apparent, less so listening to my untrained ear, but comes across as a softening of the top end sorry for my lack of techie terminology. I'm concerned a little that when I finally tackle running Dirac on this system the headroom used up may limit the volume more than acceptable for me during boisterous sessions.

I heard these last week, just before Erin's review came out, and was puzzled by what I heard, they seemed lifeless at times, and initially I thought they were in mono, I had to listen carefully to tell they weren't. I think what you have measured explains it very well, I was listening in the low 80s a weighted long term at about 3m, so probably running out of headroom at the top end. I thought the bass was really good.

 

[jackocleebrown]

Did some REW testing to get an idea of when limiting kicks in and how it impacts the frequency response in my room with my LS60/HSU subs combo. I'm running the LS60's 25 inches off the front wall.HSU ULS 15MK2 subs are behind and to teach side of the speakers backed into the corners (GIK double stacked TriTraps) They are spaced 8.6 ft and 8y.2 ft from my ears (measuring position). Room is 14.3 x 16.6 x 8 ft. Speakers are along the 14.3 ft width. KEF App settings: Wall Mode - Off Treble - 0;; Phase - On; Extension - Extra;; Balance - Center; Sub Out - Two (obviously) and Stereo;; Hi Pass - 100Hz (tried some at 120Hz); Low Pass - 135Hz (have run at 100Hz too green trace is 100/100); Sub Gain - 0dB

Room is treated with ceiling mounted GIK 244's, Double stacked TriTraps front corners. One TriTrap back upper left corner another currently standing on the floor butted up to an Ikea cube stack left side 2/3 back. Front side walls have SlatFusors and back right side corner a couple more 24x24's. Back wall has a 5x5 IKEA cube shelving unit 70% full of LP's with two 12x48 Monster Bass Traps mounted at rear wall ceiling interface above the shelving and two 12x48 Monsters with Scatter Plates vertical next to the sides of the back Ikea shelving.

View attachment 338098

Measurements
It's obvious the limiter is kicking in somewhere between the 90-93dB individual speaker test level. Cutting back the tweeter level. The waterfalls show it very clearly too.



View attachment 338100
Waterfalls

View attachment 338105
View attachment 338106
View attachment 338107
View attachment 338110
I'm very much a neophyte running and interpreting REW but the I think the FR isn't too bad excluding the room mode dip in the 100-126Hz range and lower hump. I may play some more and try dropping the Hi Pass to 80Hz see if it fills in some of the dip at all. Hopefully no limiting too.

The limiting is clearly apparent, less so listening to my untrained ear, but comes across as a softening of the top end sorry for my lack of techie terminology. I'm concerned a little that when I finally tackle running Dirac on this system the headroom used up may limit the volume more than acceptable for me during boisterous sessions.

What you see here is thermal protection kicking in on the tweeter. This occurs because a test sweep has extremely high average spectrum at high frequencies. On real music this isn't the case and this protection will almost never be triggered.

 

[palm]

Likely indeed, and not a bad idea with such a speaker with built in preamp and source selection increasing the risk of accidental high volume I guess.
But I see the traces are called LR, how was the measurement done?

 

[Puddingbuks]

Oops, never mind.

 

[Absolute]

What you see here is thermal protection kicking in on the tweeter. This occurs because a test sweep has extremely high average spectrum at high frequencies. On real music this isn't the case and this protection will almost never be triggered.

Good to see your involvement!
Perhaps we can get your thoughts on what to expect in terms of real-life spl performance?

Kali Audio previously gave a wonderful spec with their speakers to avoid guesswork where unrealistic test signals might give less meaningful results for real-life performance. They gave a maximum continuous dB spec + unclipped 20 dB peaks IIRC, say 85 + 20 dB.

I can't recall more specifics than that, but perhaps something like this could ease the mind some of us might have in regards to whether or not the smaller speakers in your line-up will suffice in terms of spl?

 

[thewas]

Too bad Kef isn’t reading this and replying.

@Kef: I know you are reading this. So please respond.

? The senior engineer of KEF replied 2 posts above yours.

 

[Puddingbuks]

? The senior engineer of KEF replied 2 posts above yours.

I stand corrected.

 

[jackocleebrown]

Good to see your involvement!
Perhaps we can get your thoughts on what to expect in terms of real-life spl performance?

Kali Audio previously gave a wonderful spec with their speakers to avoid guesswork where unrealistic test signals might give less meaningful results for real-life performance. They gave a maximum continuous dB spec + unclipped 20 dB peaks IIRC, say 85 + 20 dB.

I can't recall more specifics than that, but perhaps something like this could ease the mind some of us might have in regards to whether or not the smaller speakers in your line-up will suffice in terms of spl?

Thanks for the suggestion. I agree it's something we can improve. Our spec sheet max SPL figure is based on 1m 1speaker instantaneous max peak SPL on pink noise. I think that what you suggest is better.

A couple of extra notes.

On active speakers we do as much limiting as possible with average limiters using slow release times. This is internationally so we don't reduce the dynamic range of the music being played. It does tend to mean that sine max SPL looks artificially worse than real world max output. The benefit is lower distortion on sine and less compression on music at max replay level.

Second point is that we have a discrepancy with the max SPL data from Erin. We think it should be a good bit higher. At least 3dB above LS50W.2.

Finally, the compression curves approach to showing max SPL isn't great in my opinion because you can't see the real output at the same time. That's why, for example, some small speakers without much bass look great, yet the LS60, which has built in driver excursion protection, would be easily beating them for SPL and distortion at the same playback level. I favour multiple FR curves overlayed on the same chart.

 

[AOR]

To add to the info from @jackocleebrown 's posts above:

The limiter suite in the LS60 is pretty complex, with different limiters acting on different drivers and different portions of the frequency spectrum, all with different time-domain behaviour. Therefore, measurement methodologies that assume time-invariance, like sine sweeps, can throw up weird results that are difficult to interpret without prior knowledge of the architecture and design of the limiters.

Without going into too much detail, there are 4 different limiters: An LF excursion limiter, a broadband output level limiter, an MF thermal limiter, and an HF thermal limiter. These have been designed and tuned to be as transparent as possible with a music signal. For signals with a different crest factor and/or frequency content to music, such as a sine sweep, these limiters would activate earlier (at a lower level) than with the music signal. This can be seen in post #200. From our testing during the design of the LS60, we haven't found any music that would actually trigger the thermal limiters in the real, non-measurement, world.

 

[Muddywaters]

Likely indeed, and not a bad idea with such a speaker with built in preamp and source selection increasing the risk of accidental high volume I guess.
But I see the traces are called LR, how was the measurement done?

I measured using my Umik1. For each set of sweeps I used the TEST LEVELS button on the REW Measurement screen to check L then R speaker output individually. I believe REW outputs a pink noise signal? I noted the volume result shown for the each test eg 85dB, 90dB, 93dB and 94DB displayed by REW, also noting the corresponding volume setting on my ADI 2 Pro. Initially I hoped to determine some safe volume settings and where limiting might kick in. (I realized it was a hopeless endeavor due to widely varying source levels.) I the ran L, R and L+R. I’m displaying the L+R test results from some of the sweeps.

I initially jumped from the initial 85 dB level increasing my ADI 2 volume level 10 dB from 25.5 to 15.5 dB thinking I’d get a 95 dB but didn’t. The limiter was kicking in I guess. Hope this explains my methodology.

 

[Muddywaters]

@Muddywaters Excellent stuff, thank you for sharing! Nice room.

It looks to my naked eye that there's compression in the lower midrange as well, but it would be easier to spot if you were to show this normalized to the green one. I forgot how you do that in REW, but perhaps someone remember or can do it for you if you upload the mdat data?

I need to read how to properly upload the mdat file but will do so if appropriate for this review thread?

 

[Muddywaters]

What you see here is thermal protection kicking in on the tweeter. This occurs because a test sweep has extremely high average spectrum at high frequencies. On real music this isn't the case and this protection will almost never be triggered.

To add to the info from @jackocleebrown 's posts above:

The limiter suite in the LS60 is pretty complex, with different limiters acting on different drivers and different portions of the frequency spectrum, all with different time-domain behaviour. Therefore, measurement methodologies that assume time-invariance, like sine sweeps, can throw up weird results that are difficult to interpret without prior knowledge of the architecture and design of the limiters.

Without going into too much detail, there are 4 different limiters: An LF excursion limiter, a broadband output level limiter, an MF thermal limiter, and an HF thermal limiter. These have been designed and tuned to be as transparent as possible with a music signal. For signals with a different crest factor and/or frequency content to music, such as a sine sweep, these limiters would activate earlier (at a lower level) than with the music signal. This can be seen in post #200. From our testing during the design of the LS60, we haven't found any music that would actually trigger the thermal limiters in the real, non-measurement, world.

Thank you both for your replies. I knew that there were likely some confounding factors.

I want to state unequivocally these speakers are wonderful in my room at a wide variety of volumes on a huge diverse range of genres of music. They absolutely disappear. Only occasionally with some hard panned material do I hear/locate them. They provide a broad, deep (depending on source, of course), prescence. My goal in testing was trying ascertain if they limited or compressed at some volume levels that I may reach at times. Now I’m unsure how I can figure it out definitively.

I have, on a few tracks, felt they have run out of output and a couple times and thought I heard some softening of the very upper mids and treble range. Certainly this was at extreme levels that I would rarely be using for any length of time. Thus I embarked on the sweep testing.

For context I’m an oldster soon to be 68 and my hearing is limited to 12-13kHz so no claims of hearing oddities in the upper limits

Again thank you @jackocleebrown for your participation here at ASR and in this thread. @AOR I'm curious about your in depth knowledge of the design? Thank you for sharing. I haven’t burned out any drivers on speakers for 12 years and didn’t want to start again. Seems they have multiple layers of protection on these. I’ve only had one other pair of powered speakers, some Kali, that were placeholders for me as I built the room. I never witnessed the protection LED light up on them. They’d get rough/harsh sounding if I got to exuberant with the volume. These LS60’s haven’t. [As a side note: We had 40” of the Gulf of Mexico in here 9/28/22 from Hurricane Ian so everything is “new” to me versus the past 14 years.]

A question for you or others following along. Using the REW check levels prior to running the sweeps I believe the level check is a pink noise signal. As I mentioned up thread in post #211 I went from the 85bB setting (25.5 on the ADI 2) to -15.5 on the ADI anticipating I’d see a 95dB result instead getting 93 dB. I backed down in half dB steps from there on the ADI 2 testing with the pink noise. At -17dB the measured output was 92dB so back up to -16.5 brought me to the 93dB that I ran the sweep level at. It took going up in level to -11.5 on the ADI to get 94dB from both L and R. Thus I’m guessing the “broadband output limiter”was coming into play?

*(I need to edit my posts reflecting the correct ADI volume setting for 93dB was -16.5 not -15.5)

 

[sweetchaos]

Erin posted an update 1hr ago to address the LS60W.

Start at 12min mark:

- YouTube

Enjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.

m.youtube.com

UPDATE:
The previously posted compression testing (what Erin calls the "Response Linearity" test) is done with "2.7 second logarithmic sine sweep":

This will trigger any active speaker's protection circuits (just like Kef LS60W) and therefore the high frequency is typically where you'll see the most compression (as shown in above graph).

Here is the new graphs that Erin provided in the video above, which he'll start doing for active speakers from now on...
This is multi-tone, similar to pink noise, which lasts 30 seconds.
This test is more indicative of real world music, since it doesn't overstress the tweeter (like we see in the response linearity graph above).

Both graphs above are from Erin's written review.

 

[abdo123]

Glad to see engagement from KEF here and Erin’s follow up.

It did seem like the speaker has ridiculously low output for its size before these things have been clarified.

 

[Puddingbuks]

Erin posted an update 1hr ago to address the LS60W.

Start at 12min mark:

- YouTube

Enjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.

m.youtube.com

Thanks!

Exact time link: https://m.youtube.com/live/cvXrzMWqRq0?t=728s

Nice to see Erin an Kef sorting things out!

 

[sweetchaos]

So is Erin's new multi-tone compression test the same as @Nuyes compression tests that he typically posts?
Here's the graph from Nuyes's Buchardt Anniversary 10 review (also an active speaker):

Just trying to understand if we can compare apples to apples.

Erin has the full explanation of his method for this test, on his written review of the LS60W.

 

[Muddywaters]

@Absolute here is the mdat file oops the file is too large! I'll have to go back and remove some the L,R and a few of the extra L+R files. Just load up these four measurements.

 

[Sancus]

Tweeter compression to me is never important, as was stated above real music just doesn't stress it at all.

The MTON compression test still shows the same amount of compression in bass and midrange as the sweep did, pretty much.

 

[Muddywaters]

@Absolute and other interested parties here is the mdat file of the four sweeps. Dang, it says this one is too large. Sorry for my fumbling. Do I have to upload each measurement separately?