Some rough indoor measurements of KEF Q350

[Bjorn]

I have conducted a few indoor measurements of the KEF Q350 that I'll share. I'm aware the speaker has been measured by Amir and Stereophile, but wanted to measure it myself after I set a pair up at work where we very dissapointed by the sound quality.

First an on-axis measurement at 20 cm distance and placed on a 1 m tall stand.

1/48 Oct. smoothing:

1/6 Oct. smoothing:

At 40 cm distance, we can see the room starts to contribute more.
1/48 Oct. smoothing:

The directivity measurements below was conducted at a 40 cm distance on-axis. There's no gating used here, and the room defintely starts to interfere with the measurements. So keep that in mind.

Horizontal polar out to 90°: Normalized.

Horizontal polar out to 150°. Normalized.

Vertical polar to 90°. Normalized.

Vertical polar to 150°. Normalized.

In the next post I'll share a few more measurements and some comparisons with other speakers.

 

[Bjorn]

Here's on-axis measurement at 28 cm of the KEF Q350 (green graph) and a prototype 2-way satelite speaker (blue graph) I had build many years ago. The 2-way design used an 8" woofer, a wide band driver and active crossover.

Here I've used 1/24 Oct. smoothing

Same as above but with lower resolution on the y-axis:

While the directivity mesaurements of other speakers haven't been conducted with the exact same position in the room within centimeters, nor with same type of furnitures/treatment we still get a general idea of the difference. I will also like to point out that it's much easier to measure a speaker that has a narrow directivity more correctly indoor, since it interacts less with the room. This is also general leads to a more even response in the listening position, something I think is somewhat overlooked by many.

Let's bring up the horizontal polar of the KEF Q350 up to 90°.

And let's look at a large horn that I've been involved in designing. Here's the measurement is at 60 cm distance, 20 cm further away than the KEFs were measured at.

And here's a prototype with another coax. Also at 60 cm distance.

 

[Bjorn]

I find it interesting to see how a speaker compares in frequency response compared to other designs in the listening position.

First a measurement of a large horn speaker. This horn speaker requires a subwoofer for the lows, but I'm doing all of these without a subwoofer for an easy comparison. Graphs are with 1/24 Oct. smoothing.

2- way horn speaker:

KEF Q350 with the exact same position:

A problem with a speaker design like the KEF Q350 is that it will suffer from SBIR from side walls and front wall. Something that doesn't happened to this horn in the same way. In this case, the distance from the front wall is about 195 cm. Meaning we avoid the SBIR from the front wall. But there's still an issue with SBIR of the side wall. I therefore moved the KEF Q350 as close to the side wall I could and measured again:

As we can see, this improved the frequency response considerably. Still not on pair with the horn, but overall good.

We do get some more specular reflections though from the KEF Q350 due to a wider dispersion and especially when it's close to the side wall. We can see that in the ETCs below.

ETC KEF Q350 with placement close to side wall:

Horn ETC:

 

[ernestcarl]

I have conducted a few indoor measurements of the KEF Q350 that I'll share. I'm aware the speaker has been measured by Amir and Stereophile, but wanted to measure it myself after I set a pair up at work where we very dissapointed by the sound quality.

First an on-axis measurement at 20 cm distance and placed on a 1 m tall stand.

1/48 Oct. smoothing:

View attachment 326611

1/6 Oct. smoothing:
View attachment 326613

At 40 cm distance, we can see the room starts to contribute more.
1/48 Oct. smoothing:
View attachment 326616

The directivity measurements below was conducted at a 40 cm distance on-axis. There's no gating used here, and the room defintely starts to interfere with the measurements. So keep that in mind.

Horizontal polar out to 90°: Normalized.
View attachment 326617

Horizontal polar out to 150°. Normalized.
View attachment 326618


Vertical polar to 90°. Normalized.
View attachment 326619


Vertical polar to 150°. Normalized.
View attachment 326620



In the next post I'll share a few more measurements and some comparisons with other speakers.

The dip in the midrange appears to me wider/broader than other measurements we’ve seen thus far. Do you only have a single Q350? I’m mainly wondering if it’s only worse in this particular unit.

 

[Bjorn]

The dip in the midrange appears to me wider/broader than other measurements we’ve seen thus far. Do you only have a single Q350? I’m mainly wondering if it’s only worse in this particular unit.

I have access to three more units. I would have to bring with me one or two more when I get time.

 

[thewas]

I'm aware the speaker has been measured by Amir and Stereophile, but wanted to measure it myself after I set a pair up at work where we very dissapointed by the sound quality.

I don't understand what the purpose of those measurements and thread is when a much better/correct anechoic NFS measurements exist?

Here's on-axis measurement at 28 cm of the KEF Q350 (green graph) and a prototype 2-way satelite speaker (blue graph) I had build many years ago. The 2-way design used an 8" woofer, a wide band driver and active crossover.

Also what is the purpose of this on-axis apple and peas comparison of an entry passive loudspeaker with an active one of unknown drivers and price? Anything can be tuned to any target on a specific angle, what counts though is the total behaviour like directivity and distortions.

And let's look at a large horn that I've been involved in designing. Here's the measurement is at 60 cm distance, 20 cm further away than the KEFs were measured at.

At which frequency is it crossed and can you show its vertical directivity?

 

[Bjorn]

I don't understand what the purpose of those measurements and thread is when a much better/correct anechoic NFS measurements exist?

I like to investigate myself and promised I would share the measuerement. There could also be a fault with the units I measured. If you have no interest in this, then you can simply avoid the thread.

Also what is the purpose of this on-axis apple and peas comparison of an entry passive loudspeaker with an active one of unknown drivers and price? Anything can be tuned to any target on a specific angle, what counts though is the total behaviour like directivity and distortions.

It's good to have a reference. On-axis response matters.

At which frequency is it crossed and can you show its vertical directivity?

Probably not something you're interested in anyway.

 

[thewas]

I like to investigate myself and promised I would share the measuerement. There could also be a fault with the units I measured. If you have no interest in this, then you can simply avoid the thread.

Why not in the big Q350 thread? I have interest in full, meaningful and correct measurements and not just cherrypicked limited ones.

It's good to have a reference. On-axis response matters.

As said one single axis can be easily tuned to any target (even easier in an active or DSP design) but that by far isn't enough for to make a loudspeaker being good.

Probably not something you're interested in anyway.

So you are avoiding a full comparison with the also necessary vertical directivity?

 

[Bjorn]

Post #3 has been updated with some comparison measurements in the listening position for those who are interested..

Some rough indoor measurements of KEF Q350

I have conducted a few indoor measurements of the KEF Q350 that I'll share. I'm aware the speaker has been measured by Amir and Stereophile, but wanted to measure it myself after I set a pair up at work where we very dissapointed by the sound quality. First an on-axis measurement at 20 cm...

www.audiosciencereview.com

Also plan to do more such comparisons.

 

[abdo123]

it would be nice to see some measurements from the modded crossover, i can't find the link anymore. maybe @thewas can help?

it's a bottom of the barrel speaker after all. the uni-q driver itself is allegdly very good, just the cabinet and crossover can use some love.

 

[Bjorn]

it would be nice to see some measurements from the modded crossover, i can't find the link anymore. maybe @thewas can help?

it's a bottom of the barrel speaker after all. the uni-q driver itself is allegdly very good, just the cabinet and crossover can use some love.

I'll probably open the speaker, measure the drivers separately without a crossover and then add active crossover. Passive belongs to the stone age

 

[abdo123]

I'll probably open the speaker, measure the drivers separately without a crossover and then add active crossover. Passive belongs to the stone age

i'm a normie multichannel AVR sucker now so not really an option anymore unfortunately.

 

[thewas]

it would be nice to see some measurements from the modded crossover, i can't find the link anymore. maybe @thewas can help?

it's a bottom of the barrel speaker after all. the uni-q driver itself is allegdly very good, just the cabinet and crossover can use some love.

Pimp a KEF Q350 – oder: Gutes noch besser machen… – HiFi Alex

 

[sigbergaudio]

Do you have a measurement slightly off-axis, say 10, 15 and/or 20 degrees?

 

[sigbergaudio]

I guess it's also fair to keep in mind that this is a 700USD/pair speaker.

 

[Bjorn]

Do you have a measurement slightly off-axis, say 10, 15 and/or 20 degrees?

That's incorpated in the polar plots. I measured with 10° steps.

Here you can see 10°, 20° and 30° separately. 1/24 Oct. smoothing.

 

[Rednaxela]

If you have no interest in this, then you can simply avoid the thread.

Something about this remark feels very wrong.

 

[RobL]

Pimp a KEF Q350 – oder: Gutes noch besser machen… – HiFi Alex

That’s a great link! One of the commenters noted that the port appears to resonate at the same frequency as the 1k dip…perhaps that dip is less of a coax design issue and more likely a port cancellation issue?

 

[thewas]

That’s a great link! One of the commenters noted that the port appears to resonate at the same frequency as the 1k dip…perhaps that dip is less of a coax design issue and more likely a port cancellation issue?

You are welcome, Amir's measurement with port plugged shows it is not port cancellation related though:

 

[Thomas_A]

I find it interesting to see how a speaker compares in frequency response compared to other designs in the listening position.

First a measurement of a large horn speaker. This horn speaker requires a subwoofer for the lows, but I'm doing all of these without a subwoofer for an easy comparison. Graphs are with 1/24 Oct. smoothing.

2- way horn speaker:
View attachment 326680


KEF Q350 with the exact same position:
View attachment 326681

A problem with a speaker design like the KEF Q350 is that it will suffer from SBIR from side walls and front wall. Something that doesn't happened to this horn in the same way. In this case, the distance from the front wall is about 195 cm. Meaning we avoid the SBIR from the front wall. But there's still an issue with SBIR of the side wall. I therefore moved the KEF Q350 as close to the side wall I could and measured again:

View attachment 326689


As we can see, this improved the frequency response considerably. Still not on pair with the horn, but overall good.

We do get some more specular reflections though from the KEF Q350 due to a wider dispersion and especially when it's close to the side wall. We can see that in the ETCs below.

ETC KEF Q350 with placement close to side wall:
View attachment 326687

Horn ETC:
View attachment 326688

How large is the room?

Here is my wide dispersion two-way speakers in my small room approx 4x4.2 m using the same scale 10-95 dB, cirka 3.3 m speaker to listening pos. Sub is used < 100 Hz without correction of room mode. Speaker close to front wall (25 cm woofer center to wall), 40° toe in. Difference 100-200 Hz mostly due to asymmetric furnituring around speakers.