• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Some KEF R3 Meta measurements with REW - what can be improved

nygafre

Active Member
Joined
Aug 25, 2022
Messages
150
Likes
144
This morning I digged out my UMIK-1 to perform some measurements of my KEF R3 metas as I suspected their positioning were suboptimal.

I use my speakers in my living room both for movies and music. The speakers are placed on each side of the TV, on an 1.8 meters wide stand pretty close to the wall (conforming to estethics rather than rules of proper placement). This is my first proper attempt to measure with REW, so bear with me if I have missed something basic.

Original position:
The red curve shows the measurement from the speakers original placement. Here, the speakers are 1.6 meters apart, measured from top centre of each speaker. My listening-/measuring position is 2.4 meters away from the front of the speakers. Distance to back wall from back of speakers is (only) 150mm. I know KEF recommends a minimum of 225mm from the back wall.

TV-benk rød_stativ, grønn.jpg


Improved position:
The green curve shows the improved response. To improve the positioning, I put the speakers on nightstands on the side of the TV-stand, which increased the width to 2.25 meters centre to centre. I then continued by pulling the speakers out in steps of about 5-7 cm, and performed a measurement for each step. As I could see the response improving by each step, I continued until I reached what I found was close to optimal placement. The back of the speakers were now 340mm from the back wall.

Notes:
From the first (red) measurement, I was surprised to see the big valley in bass around 30-60Hz. The new position seemed to improve that, but then introduced smaller dips around e.g. 75Hz and 105-ish Hz. I have used 1/6th octave smoothing (don't know what is most commonly used). Without smoothing the dips were quite deep.

Both the red and green measurements were performed with the halfbungs in the port. I also did a measurement removing the bungs, which is shown by the purple curve in the following photo.

lilla_uten bungs.jpg


I found removing the bungs resulted in more bass, but after listening I think I prefer the response with the halfbungs in (green line). Things seemed a bit calmer that way.

My question would be how this looks in the more experienced eyes of the forum members here? I am already happy with the sound, but is there any obvious improvements I can do/explore? I currently don't own a sub, and as I have a neighbour on the other side of the wall I would be hesitant to introduce more bass. It would be interesting if I can improve things further, so any suggestions are welcomed.

Regards nygafre
 
Last edited:

HarmonicTHD

Major Contributor
Joined
Mar 18, 2022
Messages
3,326
Likes
4,846
This morning I digged out my UMIK-1 to perform some measurements of my KEF R3 metas as I suspected their positioning were suboptimal.

I use my speakers in my living room both for movies and music. The speakers are placed on each side of the TV, on an 1.8 meters wide stand pretty close to the wall (conforming to estethics rather than rules of proper placement). This is my first proper attempt to measure with REW, so bear with me if I have missed something basic.

Original position:
The red curve shows the measurement from the speakers original placement. Here, the speakers are 1.6 meters apart, measured from top centre of each speaker. My listening-/measuring position is 2.4 meters away from the front of the speakers. Distance to back wall from back of speakers is (only) 150mm. I know KEF recommends a minimum of 225mm from the back wall.

View attachment 296075

Improved position:
The green curve shows the improved response. To improve the positioning, I put the speakers on nightstands on the side of the TV-stand, which increased the width to 2.25 meters centre to centre. I then continued by pulling the speakers out in steps of about 5-7 cm, and performed a measurement for each step. As I could see the response improving by each step, I continued until I reached what I found was close to optimal placement. The back of the speakers were now 340mm from the back wall.

Notes:
From the first (red) measurement, I was surprised to see the big valley in bass around 30-60Hz. The new position seemed to improve that, but then introduced smaller dips around e.g. 75Hz and 105-ish Hz. I have used 1/6th octave smoothing (don't know what is most commonly used). Without smoothing the dips were quite deep.

Both the red and green measurements were performed with the halfbungs in the port. I also did a measurement removing the bungs, which is shown by the purple curve in the following photo.

View attachment 296073

I found removing the bungs resulted in more bass, but after listening I think I prefer the response with the halfbungs in (green line). Things seemed a bit calmer that way.

My question would be how this looks in the more experienced eyes of the forum members here? I am already happy with the sound, but is there any obvious improvements I can do/explore? I currently don't own a sub, and as I have a neighbour on the other side of the wall I would be hesitant to introduce more bass. It would be interesting if I can improve things further, so any suggestions are welcomed.

Regards nygafre
In general try to setup your speaker in an equilateral triangle. You are already moving in the right direction.

Putting them on speaker stands is usually a good idea.

Always keep the foam plugs out unless you use a sub. Lower bass.

Seeing dips in bass is perfectly normal. It is called room modes. Experiment a bit with positioning to find the best spot but keep in mind there will always be some dips and peaks left.

To remove the peaks in the low frequency range (usually below about 300Hz) you can get a parametric EQ eg included in an AVR or in software eg. APOEQ or Camilla DSP or a separate hardware DSP eg MiniDSP.

To even further smooth low frequency response you would need a sub or even two, which placed in such a way that the dips created by the mains are “filled” simply speaking. You need an AVR or MiniDSP to integrate sub and mains.

Good luck and enjoy your nice speakers.

Edit. Typos.
 
Last edited:

muad

Senior Member
Joined
Jun 8, 2019
Messages
425
Likes
488
This morning I digged out my UMIK-1 to perform some measurements of my KEF R3 metas as I suspected their positioning were suboptimal.

I use my speakers in my living room both for movies and music. The speakers are placed on each side of the TV, on an 1.8 meters wide stand pretty close to the wall (conforming to estethics rather than rules of proper placement). This is my first proper attempt to measure with REW, so bear with me if I have missed something basic.

Original position:
The red curve shows the measurement from the speakers original placement. Here, the speakers are 1.6 meters apart, measured from top centre of each speaker. My listening-/measuring position is 2.4 meters away from the front of the speakers. Distance to back wall from back of speakers is (only) 150mm. I know KEF recommends a minimum of 225mm from the back wall.

View attachment 296075

Improved position:
The green curve shows the improved response. To improve the positioning, I put the speakers on nightstands on the side of the TV-stand, which increased the width to 2.25 meters centre to centre. I then continued by pulling the speakers out in steps of about 5-7 cm, and performed a measurement for each step. As I could see the response improving by each step, I continued until I reached what I found was close to optimal placement. The back of the speakers were now 340mm from the back wall.

Notes:
From the first (red) measurement, I was surprised to see the big valley in bass around 30-60Hz. The new position seemed to improve that, but then introduced smaller dips around e.g. 75Hz and 105-ish Hz. I have used 1/6th octave smoothing (don't know what is most commonly used). Without smoothing the dips were quite deep.

Both the red and green measurements were performed with the halfbungs in the port. I also did a measurement removing the bungs, which is shown by the purple curve in the following photo.

View attachment 296073

I found removing the bungs resulted in more bass, but after listening I think I prefer the response with the halfbungs in (green line). Things seemed a bit calmer that way.

My question would be how this looks in the more experienced eyes of the forum members here? I am already happy with the sound, but is there any obvious improvements I can do/explore? I currently don't own a sub, and as I have a neighbour on the other side of the wall I would be hesitant to introduce more bass. It would be interesting if I can improve things further, so any suggestions are welcomed.

Regards nygafr

Can you repost with a 40-90db scale (50db)
 
OP
nygafre

nygafre

Active Member
Joined
Aug 25, 2022
Messages
150
Likes
144
In general try to setup your speaker in an equilateral triangle. You are already moving in the right direction.

Putting them on speaker stands is usually a good idea.

Always keep the foam plugs out unless you use a sub. Lower bass.

Seeing dips in bass is perfectly normal. It is called room modes. Experiment a bit with positioning to find the best spot but keep in mind there will always be some dips and peaks left.

To remove the peaks in the low frequency range (usually below about 300Hz) you can get a parametric EQ eg included in an AVR or in software eg. APOEQ or Camilla DSP or a separate hardware DSP eg MiniDSP.

To even further smooth low frequency response you would need a sub or even two, which placed in such a way that the dips created by the mains are “filled” simply speaking. You need an AVR or MiniDSP to integrate sub and mains.

Good luck and enjoy your nice speakers.

Edit. Typos.
Thank's for the suggestions!

Ah, I thought the R3's/R3 Metas worked better with noe toe in, or even with a slight toe out? I will try out APO EQ :)
 

HarmonicTHD

Major Contributor
Joined
Mar 18, 2022
Messages
3,326
Likes
4,846
Thank's for the suggestions!

Ah, I thought the R3's/R3 Metas worked better with noe toe in, or even with a slight toe out? I will try out APO EQ :)
Position first for minimal room mode and then toe-in. My Reference 3 are also toed in. What it does, it increases the direct sound and reduces the side wall reflections. However the sound stage might be perceived more narrow. There is no hard formula for toe in - just adjust to your liking of the soundstage.
 

muad

Senior Member
Joined
Jun 8, 2019
Messages
425
Likes
488
Thanks! The green is pretty close to an optimal house curve. 10db drop over the overall response. The port bung one looks the best to me. If you are optimizing for one position, then PEQ is a great way of getting the response perfect. Personally I would be happy with the green response, looks nice!

If you add a sub, you can run the r3s full range with bungs and only add just enough subwoofer to bring the response up. This would help even out the response under 100db.

Additionally I like to recommend a Moving mic method speaker measurement. It's a great way of getting an average response around your listening position. It's closer to what you will perceive in room
 
OP
nygafre

nygafre

Active Member
Joined
Aug 25, 2022
Messages
150
Likes
144
Thanks! The green is pretty close to an optimal house curve. 10db drop over the overall response. The port bung one looks the best to me. If you are optimizing for one position, then PEQ is a great way of getting the response perfect. Personally I would be happy with the green response, looks nice!

If you add a sub, you can run the r3s full range with bungs and only add just enough subwoofer to bring the response up. This would help even out the response under 100db.

Additionally I like to recommend a Moving mic method speaker measurement. It's a great way of getting an average response around your listening position. It's closer to what you will perceive in room
Yeah, I was quite happy with the green one too. Just wanted to know if I should/could look for further improvements.

How do I do the moving mic measurements? Perform several measurement, covering the entire sofa, and then average them?
 
Last edited:

muad

Senior Member
Joined
Jun 8, 2019
Messages
425
Likes
488
Yeah, I was quite happy with the green one too. Justed wanted to know if I should/could look for further improvements.

How do I do the moving mic measurements? Perform several measurement, covering the entire sofa, and then average them?



This is a pretty decent source. Just make sure you choose the right settings as per the link.
 

alont

Active Member
Forum Donor
Joined
Sep 1, 2018
Messages
100
Likes
182
Location
Seattle, WA
Understanding Speaker Boundary Interference Response (SBIR) is key for optimal speaker placement. When speakers are situated near room boundaries, reflected sound waves can interfere with the direct sound, which may either amplify or diminish certain frequencies.

For your speaker, which starts acting omnidirectional around 220Hz and extends down to about 50Hz, this is the frequency range to consider. Within this range, the sound waves are long enough to reflect off nearby boundaries and interfere with the direct sound wave, significantly affecting the perception of these frequencies at the listening position.

To minimize SBIR, aim to avoid placing your speakers a distance from a room boundary that equals a quarter of the wavelength of frequencies within this critical range. With the speed of sound at approximately 343 meters per second, the wavelength of a frequency can be calculated using the formula “wavelength = speed of sound / frequency”. For 220Hz, the wavelength is about 1.56 meters and for 50Hz, it’s around 6.86 meters. Therefore, try to avoid positioning your speakers a quarter of any value within this range (roughly 0.39 to 1.72 meters) away from any boundary. Note that the distance should be measured from the boundary to the emitting driver.
 
Last edited:
OP
nygafre

nygafre

Active Member
Joined
Aug 25, 2022
Messages
150
Likes
144
... , don't you feel this is close to ideal for home use already?!
Yes, I’m very happy with the sound already. It’s my first time trying to measure properly, thought it was kinda fun and wanted to share to see if there was any obvious things to correct further. It seems there are no major issues in my measured response/setup?
 

geox

Member
Joined
May 30, 2022
Messages
42
Likes
59
Understanding Speaker Boundary Interference Response (SBIR) is key for optimal speaker placement. When speakers are situated near room boundaries, reflected sound waves can interfere with the direct sound, which may either amplify or diminish certain frequencies.

For your speaker, which starts acting omnidirectional around 220Hz and extends down to about 50Hz, this is the frequency range to consider. Within this range, the sound waves are long enough to reflect off nearby boundaries and interfere with the direct sound wave, significantly affecting the perception of these frequencies at the listening position.

To minimize SBIR, aim to avoid placing your speakers a distance from a room boundary that equals a quarter of the wavelength of frequencies within this critical range. With the speed of sound at approximately 343 meters per second, the wavelength of a frequency can be calculated using the formula “wavelength = speed of sound / frequency”. For 220Hz, the wavelength is about 1.56 meters and for 50Hz, it’s around 6.86 meters. Therefore, try to avoid positioning your speakers a quarter of any value within this range (roughly 0.39 to 1.72 meters) away from any boundary. Note that the distance should be measured from the boundary to the emitting driver.
thanks for this.. super insightful. this is exactly the kinda stuff i hang around in asr for.

would sound absorption behind the speakers help with this?
 
Last edited:

muad

Senior Member
Joined
Jun 8, 2019
Messages
425
Likes
488
thanks for this.. super insightful. this is exactly the kinda stuff i hang around in asr for.

would sound absorption behind the speakers help with this?
No the wave lengths are too long. The rear distance, subs crossed very high or cardioid speakers are the only way to mitigate it.
 

alont

Active Member
Forum Donor
Joined
Sep 1, 2018
Messages
100
Likes
182
Location
Seattle, WA
thanks for this.. super insightful. this is exactly the kinda stuff i hang around in asr for.

would sound absorption behind the speakers help with this?
They could help if the emitting driver is fairly close to a boundary, otherwise they would have to be quite thick. Good 2" thick panels would only be able to absorb about 60% of the energy at 200 Hz, and basically no energy under 100 Hz.
 
OP
nygafre

nygafre

Active Member
Joined
Aug 25, 2022
Messages
150
Likes
144
Understanding Speaker Boundary Interference Response (SBIR) is key for optimal speaker placement. When speakers are situated near room boundaries, reflected sound waves can interfere with the direct sound, which may either amplify or diminish certain frequencies.

For your speaker, which starts acting omnidirectional around 220Hz and extends down to about 50Hz, this is the frequency range to consider. Within this range, the sound waves are long enough to reflect off nearby boundaries and interfere with the direct sound wave, significantly affecting the perception of these frequencies at the listening position.

To minimize SBIR, aim to avoid placing your speakers a distance from a room boundary that equals a quarter of the wavelength of frequencies within this critical range. With the speed of sound at approximately 343 meters per second, the wavelength of a frequency can be calculated using the formula “wavelength = speed of sound / frequency”. For 220Hz, the wavelength is about 1.56 meters and for 50Hz, it’s around 6.86 meters. Therefore, try to avoid positioning your speakers a quarter of any value within this range (roughly 0.39 to 1.72 meters) away from any boundary. Note that the distance should be measured from the boundary to the emitting driver.
Thank’s, great advice.

So when positioning the speakers from the back and side walls, I need to keep in mind the distance both from the rear port to the back wall, and also the front facing driver to the back wall?

As things are setup now, the port is 34cm from the back wall, and the front of the speakers 66cm. There is no way for me to keep them 1.72+meters away from any boundary.. but it would be ideal if I could?
 

HarmonicTHD

Major Contributor
Joined
Mar 18, 2022
Messages
3,326
Likes
4,846
Thank’s, great advice.

So when positioning the speakers from the back and side walls, I need to keep in mind the distance both from the rear port to the back wall, and also the front facing driver to the back wall?

As things are setup now, the port is 34cm from the back wall, and the front of the speakers 66cm. There is no way for me to keep them 1.72+meters away from any boundary.. but it would be ideal if I could?
Don’t worry. You are good. Look at your green curve. If you were to move them “far” away from the walls you loose the room gain effect for the lower frequency and your R3 would lack bass, especially as they don’t have big / many bass drivers.

As said above, to smooth the lower frequencies get at least RoomEQ and if you can a sub or even two.
 
OP
nygafre

nygafre

Active Member
Joined
Aug 25, 2022
Messages
150
Likes
144
IMG_0999.jpeg


Regarding SBIR.. then the huge valley/dip in bass on the red curve is due to the original positioning beeing around one of these frequency nodes/‘quarters’? On this measurement, the front of the speakers were aprox. 47cm from the back wall.

The dip around 150Hz seems to correlate quite well with the SBIR calculations and the positioning, I think.
 
Last edited:
Top Bottom