KEF LS50 thoughts

[Steve81]

Holy sh!t, did I get anything right?

You got the basic outline of things right, so you're not totally senile

I went sniffing around looking for the info and quickly became consumed with gear lust for gear I don't need, can't use etc. Thanks.

Who doesn't need a 21" subwoofer???

 

[Dennis Murphy]

I looked at that thread, the specific graph of the LS50 you posted in #24 was definitely the off axis only.

But I'm not sure why anyone is disputing that a 5dB rise from 1.5khz measured across the listening window and beyond, in an anechoic chamber, would be anything but a speaker induced resonance. Even a few of Stereophile's relative dB response graphs show the 5 dB rise - the waterfalls favor resolution in the time domain.

A serious resonance should show up on the waterfall plot. They certainly do on mine. I think the jury is still out on this one.

 

[Beave]

As I stated in this (or the other parallel thread?), my guess now is that JA's waterfall plot (and others I've seen) for the LS50 are clean because they are based on the on-axis signal from the speaker. The peak in discussion (or resonance, if you want to call it that) develops off-axis, but isn't present on-axis.

 

[PENG]

I looked at that thread, the specific graph of the LS50 you posted in #24 was definitely the off axis only.

I copied/pasted the wrong graph by mistake.

But I'm not sure why anyone is disputing that a 5dB rise from 1.5khz measured across the listening window and beyond, in an anechoic chamber, would be anything but a speaker induced resonance. Even a few of Stereophile's relative dB response graphs show the 5 dB rise - the waterfalls favor resolution in the time domain.

Who's disputing that? For me, I simply asked for the evidence. If you picked compared the lowest dip to the highest pick in the graph you get barely 5 dB, the overall FR graph seems to show +/- 2 dB or less. Resonance or not, that's actually among the better ones in the NRC's graphs inventory. Again, not disputing, just stating my own observations. As to whether the rise you referred to was due to resonance, could you ask Dr. Floyd Toole to comment please?

http://www.soundstagenetwork.com/index.php?option=com_content&view=category&id=77&Itemid=153

By the way, did you read the white paper that JA mentioned? I don't understand everything said in the paper but I am impressed they apparently put a lot of effort in tacking "resonance", FEA and BEA were used, I doubt too many companies, especially the smaller ones have the resource to do it. FEA is a very effective computerized method for detailed structural analysis. The 23 page file is too large to upload, just follow the link.

https://www.kefdirect.com/media/wysiwyg/documents/ls50/ls50_white_paper.pdf

 

[PENG]

A serious resonance should show up on the waterfall plot. They certainly do on mine. I think the jury is still out on this one.

May be that peak is in fact due to resonance, but not as serious as some think it is..

 

[Beave]

Dr. Toole commented on another forum a few months back that the off-axis peak seen in the NRC measurements does indicate a resonance.

I can't figure out what kind/type/origin, however. It's puzzling.

 

[TheWarrior]

A serious resonance should show up on the waterfall plot. They certainly do on mine. I think the jury is still out on this one.

Dennis, surely you understand that with waterfalls you have to choose between resolution in frequency or time. If a resonance is an octave wide, a low (frequency) resolution measurement won't clearly identify that. But all the on and off axis measurements do!

May be that peak is in fact due to resonance, but not as serious as some think it is..

An abrupt +5dB level change in anechoic measurements is a 'colored' sounding speaker. This thread started asking for thoughts on this speaker, and my thought was simply that the coloration I heard was too significant to justify the price. You enjoy the sound of your speakers, so the OP's take away should be to seriously audition these speakers before purchasing!

Not to take a turn to 'negative town' but I breezed through the KEF 'white paper'. I don't think it was your fault for not understanding everything in there, which may explain why no author is listed. FEA analysis is great, but showing pictures of half enclosures with rainbow colors and arbitrarily cited frequencies is no way to demonstrate technical prowess. Nor is an inability to stay on topic. I don't think that paper in any way credits a manufacturer thats been around as long as KEF.

 

[TheWarrior]

Dr. Toole commented on another forum a few months back that the off-axis peak seen in the NRC measurements does indicate a resonance.

I can't figure out what kind/type/origin, however. It's puzzling.

Those are the types of 'predictions' mentioned earlier that would be provided with comprehensive measurement data, such as CTA 2034. More data is needed to understand the 'origin' but not to confirm the existence of the resonance.

 

[Dennis Murphy]

As I stated in this (or the other parallel thread?), my guess now is that JA's waterfall plot (and others I've seen) for the LS50 are clean because they are based on the on-axis signal from the speaker. The peak in discussion (or resonance, if you want to call it that) develops off-axis, but isn't present on-axis.

But there's no peaking shown on any of John's off-axis plots. If there's no peak, there's no need to even look at an off-axis waterfall. The Stereophile and NRC measurements are simply irreconcilable.

 

[Beave]

Those are the types of 'predictions' mentioned earlier that would be provided with comprehensive measurement data, such as CTA 2034. More data is needed to understand the 'origin' but not to confirm the existence of the resonance.

Given that the output of this speaker is nearly the same in the vertical axis as it is in the horizontal axis, you can predict pretty well what the other data would look like.

 

[TheWarrior]

But there's no peaking shown on any of John's off-axis plots. If there's no peak, there's no need to even look at an off-axis waterfall. The Stereophile and NRC measurements are simply irreconcilable.

The Stereophile file measurements are 1/6 octave, that is too low of a resolution and will mask resonances, especially ones that are an octave wide.

But even the Listening Window measurement shows a 5 dB rise to 2khz, and it's not even the same scale!

 

[Beave]

But there's no peaking shown on any of John's off-axis plots. If there's no peak, there's no need to even look at an off-axis waterfall. The Stereophile and NRC measurements are simply irreconcilable.

JA's off-axis plot is a difference plot, which obscures the data somewhat. But I do agree that I would expect to see *some* evidence of a localized off-axis peak there, even the way he plots it... but there's nothing to see.

 

[Beave]

The Stereophile file measurements are 1/6 octave, that is too low of a resolution and will mask resonances, especially ones that are an octave wide.

His waterfall plots for lots of other speakers show resonances, both narrow and wide ones.

 

[Dennis Murphy]

The Stereophile file measurements are 1/6 octave, that is too low of a resolution and will mask resonances, especially ones that are an octave wide.

But even the Listening Window measurement shows a 5 dB rise to 2khz, and it's not even the same scale!

View attachment 23496

\\
It's more like a 2.5 dB dip in the response above 1 kHz. Many woofers show similar behavior, and it's very difficult to deal with using a passive crossover. But it's not a resonance. I don't think you can find any evidence of a resonance in the Stereophile plots.

 

[Beave]

90% of the time I ignore subjective reviews but do read them for fun. I am quoting the two below (extracts only) just for demo. I do have reasons to trust them a little (just a little) more, but I could be wrong. Doug Schneider apparently oversees the loudspeaker measurement program at the NRC and John Atkinson A has reportedly performed extensive technical analyses of well over 700 different loudspeakers

Doug oversees the Soundstage! network of online audio review sites, and they contract/pay the NRC to do measurements for them for some speakers they review. I don't think Doug works directly for the NRC or runs anything for them. Rather, he is a client of their lab.

 

[Johnny2Bad]

I've not followed the more technical discussion here closely (time, you know) but I just want to mention that the UK speaker industry has a long history of allowing, you might even say using, cabinet resonances to voice a loudspeaker. This is in contrast to the North American practice of hunting down and (ideally) eliminating all enclosure contributions (the Japanese are sometimes even more fastidious in that regard, if their better speaker products {as in 'intended for serious HiFi Audio buyers'} are considered).

It's one of the reasons why we still have "British" and "American" ~sound descriptions. Painted with a rather broad brush, it's my experience that people tend to rather consistently like one and dislike the other approach.

 

[TheWarrior]

Well, I think we've done a good job of making it clear we have differing opinions on the speaker under scrutiny.

Take it home, listen and compare, folks!

 

[PENG]

Doug oversees the Soundstage! network of online audio review sites, and they contract/pay the NRC to do measurements for them for some speakers they review. I don't think Doug works directly for the NRC or runs anything for them. Rather, he is a client of their lab.

Yes, I knew that but sort of misstated it by omitting to mention SoundStage Network. Thank you for clarifying it for me. It is not exclusive, anyone can, at least try, to use NRC..

 

[PENG]

Well, I think we've done a good job of making it clear we have differing opinions on the speaker under scrutiny.

Take it home, listen and compare, folks!

Thank you!

 

[ematthews]

So I went back and set up my system with my Ascend RAAL towers yesterday and BAMMM!! What a sound. It's been a while since I had them hooked up and was like wow.. Maybe I don't need anything else. Now to make them perfect, just make them bigger with 8 in woofers so I don't need to use my sub.

So I think I am a tower lover after all. Will probably be putting both the KEF and Polk's up for sale and looking at bigger towers as my Ascend alternative.