Advice between Kef R3 Meta or Buchardt S400 MKII ?

[CK.]

Not to mention that if you put 99% of population in front of either KEF R3, Buchart, Revel or just about anything reasonable they think it sounds awesome. Only the small minority of people have interested in really digging into the sound and trying to find what they like or don't like (sometimes based on what they think of the manufacturer beforehand).

Recommending anything to the 1% is then again fools errand as this minority can have wildly varied preferences. For example thinking that speaker A has so harsh treble they get migraine just listening to the first few tunes of any song or the soundstage is so constraining that it sucks they soul out of everything making any kind of enjoyment impossible. Even though a skeptical person might wonder if they could notice these things in a controlled environment.

Still, if one got paid by results of their recommendations I think most would go with the measurements without intimate knowledge of the recipients preferences.

Indeed, listening/reading the reviews of each speaker i suspect that some subtle differences are exaggerated for the sake of the review and found my self going through this rabbit hole just for peace of mind. But its interesting to hear that you cannot go terribly wrong with these well regarded speakers, thanks for that

 

[CK.]

But people might not like any product you recommend them for a multitude of reasons.
The upside of recommending based on measurements is that hard data is objective, and there is even research evidence that listener preferences under controlled conditions are more uniform than people usually expect. As many others I highly recommend dr. Floyd Toole's amazing book "Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms" to anyone interested in the hobby.
Recommending based on subjective listening impressions is more likely to be wrong due various perceptual and cognitive biases that are simply part of the human experience.

The reason why @amirm didn't initially like the Kef R3 was due to a room resonance and not due to the loudspeaker - as he explained in detail in one of the subsequent reviews: see the "Speaker Listening Tests" part of this post (though the whole post is worth reading).
This just highlights how important room EQ is with any loudspeaker (and how unreliable uncontrolled listening impressions can be)!

I see, thanks again for the clarifications and references! I think i will go through that book because i do find all these interesting and have a slight ocd

 

[dominikz]

Yes i recently bought the r3 online and still in the 30 days return policy. Out of the box it sounds a bit harsh/bright but the build quality and looks are spot on with my taste so trying to eq it. I will share results in a new post for advice

Might also depend on what you used before and gotten used to. E.g. if the speakers/headphones you used before had a rolled-off high end you might be used to that and now R3 could sound bright.

Though the Kef R3 does seem to have a slight low-Q mid scoop followed by a very slight low-Q high frequency boost directly on-axis - maybe you're sensitive to that:

Perhaps you'd like to try the EQ preset generated by @pierre's amazing online tool (link): KEF R3 EQ preset
(NOTE: I personally wouldn't apply the 69Hz and 245Hz filters - that region should be EQ-ed based on in-room measurements)

On-axis after EQ:

 

[Ellebob]

We have an enthusiast customer that just took home the R3 and his comments were similar to Amir's just not exciting and thought they sound like a blanket was over them. He also commented I don't usually agree with Amir but this time he is spot on. This is going to be second system for him and has Focal Utopia in his HT. He wants to stay around $2k and tried the Focal Aria 906 at a different time and thinks he might like them better than the R3 but they didn't blow him away either. He did like the Dynaudio Special 40 and so far I think that is his top choice. He also listened to but Kali SM-5-C at my house and really liked them but more than he wants to spend if you include the Lea amp. He wondered if they would work with his Lyngdorf 1120 and if the room perfect could do the EQ. I told him the warranty is not valid if something happens and you are not using their recommended amp which has protection built in. He is still looking but going by measurements he is a little all over the map. So, I always recommend everyone to listen before they buy and not just go by reviews or measurements.

 

[bodhi]

We have an enthusiast customer that just took home the R3 and his comments were similar to Amir's just not exciting and thought they sound like a blanket was over them. He also commented I don't usually agree with Amir but this time he is spot on.

I guess you have to continue disagreeing as Amir seemed to like the R3 very much after a small EQ adjustment. As has been corrected half gazillion times in this and other threads.

 

[Purité Audio]

We have an enthusiast customer that just took home the R3 and his comments were similar to Amir's just not exciting and thought they sound like a blanket was over them. He also commented I don't usually agree with Amir but this time he is spot on. This is going to be second system for him and has Focal Utopia in his HT. He wants to stay around $2k and tried the Focal Aria 906 at a different time and thinks he might like them better than the R3 but they didn't blow him away either. He did like the Dynaudio Special 40 and so far I think that is his top choice. He also listened to but Kali SM-5-C at my house and really liked them but more than he wants to spend if you include the Lea amp. He wondered if they would work with his Lyngdorf 1120 and if the room perfect could do the EQ. I told him the warranty is not valid if something happens and you are not using their recommended amp which has protection built in. He is still looking but going by measurements he is a little all over the map. So, I always recommend everyone to listen before they buy and not just go by reviews or measurements.

Customers almost always believe that what they have been listening to for the last thirty years is ‘correct’ even when it isn’t ,

Focal Maestro Utopia III loudspeaker Measurements | Stereophile.com

Sidebar 3: Measurements Measuring the behavior of a speaker as large and heavy as the Focal Maestro Utopia III is problematic. Ultimately, I had to perform the acoustic measurements with the speaker in my listening room.

www.stereophile.com

If he used to HF lift then any flat FR is initially going to sound a little dull in comparison.
Keith

 

[CK.]

Might also depend on what you used before and gotten used to. E.g. if the speakers/headphones you used before had a rolled-off high end you might be used to that and now R3 could sound bright.

Though the Kef R3 does seem to have a slight low-Q mid scoop followed by a very slight low-Q high frequency boost directly on-axis - maybe you're sensitive to that:

Perhaps you'd like to try the EQ preset generated by @pierre's amazing online tool (link): KEF R3 EQ preset
(NOTE: I personally wouldn't apply the 69Hz and 245Hz filters - that region should be EQ-ed based on in-room measurements)

On-axis after EQ:

Before I used to have KEF Q150 as LR which I am now using as surrounds. Thanks for the tool reference had no idea someone has done such a thing! will try the eq but I can see the responce is flat, the EQ I used had an in-room FR with a negative slope above say 3K. isn't that what the harman curve is all about? I am attaching the REW FR post EQ. is the shelve too aggresive? I am using 1/6 smoothing and my room is untreated. thanks

 

[CK.]

We have an enthusiast customer that just took home the R3 and his comments were similar to Amir's just not exciting and thought they sound like a blanket was over them. He also commented I don't usually agree with Amir but this time he is spot on. This is going to be second system for him and has Focal Utopia in his HT. He wants to stay around $2k and tried the Focal Aria 906 at a different time and thinks he might like them better than the R3 but they didn't blow him away either. He did like the Dynaudio Special 40 and so far I think that is his top choice. He also listened to but Kali SM-5-C at my house and really liked them but more than he wants to spend if you include the Lea amp. He wondered if they would work with his Lyngdorf 1120 and if the room perfect could do the EQ. I told him the warranty is not valid if something happens and you are not using their recommended amp which has protection built in. He is still looking but going by measurements he is a little all over the map. So, I always recommend everyone to listen before they buy and not just go by reviews or measurements.

Hm I cant really say for sure if they sound like having a blanket over them because my previous speaker were also from KEF so it might be the KEF sig? But tbh to me they sound pretty clean and transparent with enough midrange presence (tad forward).

 

[CK.]

If you EQ based on anechoic data and don’t do high Q correction (coz inevitably unit to unit variation of various minor peaks and dips will persent anyway) then it would correct minor flaws given the good directivity of the R3, but if you do 2-16khz based on in room measurements I would suggest you not to, since our ears can differentiate direct vs reflected sound quite well in those frequencies. Those corrections would likely result in Weirdness could appear something like shifted image center or noticeable boost in one speaker at certain frequencies

hm...could you please elaborate a bit more? if I do find that the 2-16Khz region is a bit boosted then I should not EQ that? I have attached my in-room FR measured in REW of my left channel. Red is post EQ, green is without any EQ. Room is untreated. thanks!

 

[YSC]

Before I used to have KEF Q150 as LR which I am now using as surrounds. Thanks for the tool reference had no idea someone has done such a thing! will try the eq but I can see the responce is flat, the EQ I used had an in-room FR with a negative slope above say 3K. isn't that what the harman curve is all about? I am attaching the REW FR post EQ. is the shelve too aggresive? I am using 1/6 smoothing and my room is untreated. thanks

if you corrected the 2-16khz region, those are mostly reflected and late sound which our brain can separate where the mic can't, so when you corrected to the apparent in room sweep to flat you will have messed up the direct sound which should be flat and the brain can isolate out, you can even visualize some form of problem say by moving the mic a bit and I bet the response will become worse than the original.

Harman curve is what a anechoic flat speaker with good directivity should behave in a well treated room, which don't have all the messed up peaks and dips you would be correcting due to furniture etc.

I would say try disable the current filter, do a MMM (multi measurement in the MLP region and not a spot, then average out) and develop a cut only filter with a bass shelve, only trim away the peak like the one you have in around 50hz. and see how the imaging would result

 

[CK.]

The other problem is measurements dont add '' human errors or weird room issues ''

Measurements just tell you the truth, the rest is your work: like understand the measurements and fix the in-room errors. That being said, the research also include somethings like people will prefer 25% a speaker more brighter and 25% more warm(I cant remember the exact numbers), so all you have to do is know how to read these graphs and measurements and then suit the speaker to your personal liking.

For example i think my KEF R7s are very very nice speakers speakers with 3 filter:

Because i know i like smooth speakers i down -1.5dB (2 filters) around 3000hz-9000hz and because i also know i like things like cymbals i EQ the sheen/airy things at like 10.5khz. So the 3rd filter is +4dB 10,500hz-20k.



You have to understand your liking and your room too.

Yeah what you are saying makes sense, I do also like a smooth sound and not very forward midrange. I've already attached my FR, but maybe I overdid it with the negative slope above 3K.

 

[CK.]

if you corrected the 2-16khz region, those are mostly reflected and late sound which our brain can separate where the mic can't, so when you corrected to the apparent in room sweep to flat you will have messed up the direct sound which should be flat and the brain can isolate out, you can even visualize some form of problem say by moving the mic a bit and I bet the response will become worse than the original.

Harman curve is what a anechoic flat speaker with good directivity should behave in a well treated room, which don't have all the messed up peaks and dips you would be correcting due to furniture etc.

I would say try disable the current filter, do a MMM (multi measurement in the MLP region and not a spot, then average out) and develop a cut only filter with a bass shelve, only trim away the peak like the one you have in around 50hz. and see how the imaging would result

Oh OK...I did not realize that above 3K the mic was somehow "exaggerating" the response due to reflections, thanks for that! I will remove the filter and do MMM but I wanted to use EQ in the first place because it sound a bit harsh/bright to me. Would you say I could try reducing the 2.5K-4K region by -2dB?
So let me see if I got this right, I cannot really measure how my brain listens to the sound say above 3K but if the mic measures a flat FR then according to research in a typical room the response "SHOULD" be like a harman curve?

 

[dogmamann]

R3 Meta anyday

 

[YSC]

Oh OK...I did not realize that above 3K the mic was somehow "exaggerating" the response due to reflections, thanks for that! I will remove the filter and do MMM but I wanted to use EQ in the first place because it sound a bit harsh/bright to me. Would you say I could try reducing the 2.5K-4K region by -2dB?
So let me see if I got this right, I cannot really measure how my brain listens to the sound say above 3K but if the mic measures a flat FR then according to research in a typical room the response "SHOULD" be like a harman curve?

the theory says you better correct only below 200hz or so, but sounding bright could be due to 1) you prefer more tilted down treble 2) you have too few bass, when you trying to EQ in room flat rather than a low shelve below ~150hz of 6db.

I would say find a target curve and use REQ to develop the EQ to cut the bass peaks over that curve and allow the in room bass boost there and see how you like that

 

[Bridges]

You appear to be talking about the S400 MKI. Virtually all of that has been remedied in S400 MKII. Lower crossover point. Braced cabinet. Smoother frequency response and dispersion.

Measurements are available on Buchardt’s website, which have historically been from Warkwyn.

Why did they not get it righ the first time?

 

[BrokenEnglishGuy]

Yeah what you are saying makes sense, I do also like a smooth sound and not very forward midrange. I've already attached my FR, but maybe I overdid it with the negative slope above 3K.

There is a lot lf people who can dont like the r series stock, if you look at the measurements, at 1khz its increase the energy up 8-9khz, but because the excellent directivity you can eq that down. There is also a boost that you may apply 10khz+

The quantity is up your liking

 

[BrokenEnglishGuy]

hm...could you please elaborate a bit more? if I do find that the 2-16Khz region is a bit boosted then I should not EQ that? I have attached my in-room FR measured in REW of my left channel. Red is post EQ, green is without any EQ. Room is untreated. thanks!

Yes that how the mf hf from the r3 ( i see r3) tend to sound, at1khz ->8khz its FR tend to won energy instead of being flat like the R3M or Reference series, fr is very important and is very noticeable by the ears.

Pd: bass response is a mess, id recommend to measure the right and left channel separately, because the bass response in your room is a mess like mine. I never got good result with r&l measurements in bass
Pd2: Im seeying its 1dB scale, the post eq how it's sound in the listening test?
Pd3: I always see a mess in the in room bass response from KEFs lol. That include mine.

What i would do is this, correct these to your taste, should be very easy
The 3khz is too boosted
Personally i have for my R7s
-0.5dB 3khz
-1.5dB 5khz
+4dB 10.500hz
But thats are just my personal preference. You should EQ your speaker at your taste and room.

 

[dominikz]

Before I used to have KEF Q150 as LR which I am now using as surrounds. Thanks for the tool reference had no idea someone has done such a thing!

It is really an amazing resource, with a lot of functionality for analyzing and comparing loudspeaker measurements. EQ presets are a nice bonus!

will try the eq but I can see the responce is flat, the EQ I used had an in-room FR with a negative slope above say 3K. isn't that what the harman curve is all about?

Actually, the Harman curve is what you measure in-room when you have a front-firing loudspeaker with flat on-axis response and 'good' directivity (i.e. what we usually call a well-measuring loudspeaker). Therefore the Harman curve should not in general be used as a target for EQ (which is a common misinterpretation); it should instead be a natural result of using well-measuring loudspeakers in-room.
You may notice that the anechoic predicted in-room response (PIR) of good loudspeakers looks suspiciously like the "Harman curve" - this is not a coincidence.
Note that loudspeakers with 'good' directivity can be EQ-ed for a flat on-axis / listening-window, and that usually also results in the in-room response looking like the "Harman curve".

Oh OK...I did not realize that above 3K the mic was somehow "exaggerating" the response due to reflections, thanks for that!

The microphone is simply capturing direct sound and reflections for all requencies, whereas human hearing can apparently separate direct sound from reflections progressively better as frequency goes up. This is why we can EQ low frequencies based on an in-room measurement, but prefer to EQ higher frequencies based on anechoic measurements (i.e. to EQ based on direct sound frequency response only, without the influence of reflections).

So let me see if I got this right, I cannot really measure how my brain listens to the sound say above 3K but if the mic measures a flat FR then according to research in a typical room the response "SHOULD" be like a harman curve?

As mentioned above, a flat measurement of a well-measuring loudspeaker in an anechoic chamber should result in a Harman curve when the same loudspeaker is measured in-room.

I have attached my in-room FR measured in REW of my left channel. Red is post EQ, green is without any EQ. Room is untreated. thanks!

Some comments:

1. I assume this is a sweep measurement?
   I'd recommend using MMM to measure instead - perhaps these short instructions might be helpful. Note that I used a flat target in this case because it was an ultra-nearfield setup. For living room listening distances I'd recommend a slope instead which approximates the natural slope of the loudspeaker above ~1kHz (but only calculate filters below about 300Hz; if you want to EQ above that use anechoic data - e.g. the filters from the Spinorama web tool I linked to previously).
2. What is the distance from loudspeaker that you measured from?
   Ideally you should measure from the listening location.
3. How far is the front of the loudspeaker from the back wall, adjacent wall, floor and ceiling?
   Based on your measurement it looks it might be 50-60cm from both the back and adjacent walls, resulting in the bass suckout around 150Hz. That might even be the main issue making the loudspeaker sound bright. Typically we prefer to look for positions where we have peaks instead of dips (i.e. better "boomy" than "bass-shy" prior to EQ! ), as peaks we can easily knock down later with PEQ.

My personal algorithm for setting up speakers in room is something like:

• Try to find a position with maximum bass response and with least bass suckout. If there are no subwoofers, I'll probably start with speakers close to the back wall, with as different as possible distances from other walls, floor and ceiling (to avoid SBIR null overlap). You can use sweep measurements to quickly see how the bass looks at each position. If you have a subwoofer rules change slightly!
• (Optional) If needed, apply loudspeaker correction EQ based on anechoic data above 300Hz (good loudspeaker don't usually need this).
• Measure in-room response with MMM for each channel individually and both channels combined. Use "variable" smoothing in REW.
• Find the target slope by matching one of the channel measured in-room response above ~1kHz
• Calculate the room EQ for each channel against the target slope, but only use subtractive EQ (no boosts!), and only below ~300Hz. Do this for each channel.
• Add the newly calculated filters to the existing loudspeaker EQ - now every channel will have slightly different filters <300Hz, but the same ones >300Hz (if you applied loudspeaker EQ in step 2).
• Again measure in-room response with MMM with EQ applied for each channel and both channels together to validate your results, and compare with the previous measurement.
• Listen and enjoy (and try to force myself to leave well-enough alone and not to fiddle with it anymore)
• Now you can use simple tone controls (if you have them) to adjust treble and bass to your liking.

Hope this is helpful, and good luck!

 

[staticV3]

will try the eq but I can see the responce is flat, the EQ I used had an in-room FR with a negative slope above say 3K. isn't that what the harman curve is all about?

The EQ aims to make the on-axis anechoic response as flat as possible. Room acoustics will then naturally tilt the response downward.
All this is in line with Harman research.

 

[BrokenEnglishGuy]

btw, the In-room response indicate the HF is 2dB above, the 3dB tolerance is still there. You can change the line whatever you like, but as you cann see there is 2dB extra in the HF for the R UNIQ. I don't find the mids a bit weak since i like the (top mid frequency) a bit more tilted, sounds more smooth