Mackie MR524 | Studio Monitor | Measurements & Subjective Impressions

[Weeb Labs]

EDIT 13/12/21: Updated with full spin and other fun things.

I recently purchased a pair of Mackie MR524s and thought it might be helpful to capture a set of essential measurements. To my knowledge, none are currently available.

The MR524 is sold individually for approximately €100, which places it within the same price bracket as the JBL 305P, Adam T5V and KRK Rokit 5. I find them to be rather aesthetically pleasing and many of the product images online fail to do them justice.

I love the inverted dust cap!

The rear is populated with a handy selection of RCA, TRS and XLR inputs. There are also three onboard PEQ settings, a three-position HF adjustment switch and input gain control with the unity point marked. The HF adjustment switch is best left in the -2dB position, as we will soon discover.

Firstly, we have the on-axis and off-axis measurements. These are a composite of nearfield and gated methods, which should be reasonably comparable to measurements acquired using a Klippel NFS or anechoic chamber.

On-axis response is predominantly flat, with a little bit of additional bass and a bump around 10KHz. Directivity is excellent, which will enable effective correction of the on-axis response via EQ or DRC.


In-room response displays a rise at higher frequencies, which will sound rather "bright".

Moving the HF adjustment switch to the -2dB position produces a significantly flatter response, so this is the configuration that I would recommend.

Next, we have nearfield measurements of all driver components.

We have a rather loud port resonance at 1KHz, so if using these monitors in conjunction with a subwoofer, you may wish to plug the port. The midbass response also dips just above the crossover, which is likely responsible for the briefly narrowed directivity at that point.


Next, polar and line charts for directivity. Horizontal directivity is quite wide and uniform, with approximately 45 degrees of freedom up to 15KHz.

You will definitely want to remain within about 25 degrees on the vertical axis. This likely results from the driver separation being somewhat larger than is typical, despite being less than a wavelength apart at the crossover. If you must listen off-axis, prefer above to below.

Finally, we have harmonic distortion. This measurement is subject to room interaction and so accuracy is always questionable.

Subjective impressions were very positive. The 10KHz bump is readily apparent but easily corrected with a gentle filter and at that point, they sound just as lovely as my T5Vs and LSR 305s. Quiet Light from Hanaregumi is one of my favorite test tracks and it sounds delightful on these; especially in conjunction with Audiolense correction!

Thank you for reading and I hope this thread has been of help to you.

Spoiler: Pierre's Anechoic Correction PEQ

NOTE: If making use of either these filters or the attached WAV IR, please ensure that the "Acoustic Space" and "High Freq Filter" switches on the rear of the monitors are in the "0dB" position".

Filter (#)	Frequency (Hz)	Gain (dB)	Width (Q)	Width (BW)
1	102.12	-2.93	1.24	1.13
2	939.68	-2.22	4.37	0.32
3	1270.67	1.18	2.92	0.49
4	1922.01	-2.39	2.70	0.53
5	5153.60	-2.29	1.31	1.07
6	9866.26	-4.67	2.94	0.48
7	16694.51	-3.88	3.43	0.41

 

[sweetchaos]

Thanks for your efforts.

Can you please provide us with the raw spinorama data? We can use it to compute the preference score.

 

[Weeb Labs]

Thanks for your efforts.

Can you please provide us with the raw spinorama data? We can use it to compute the preference score.

I have attached the spin data. Hopefully that is all of it.

 

[sweetchaos]

Thanks. @pierre @MZKM @Maiky76

 

[napilopez]

That's solid! Thanks for sharing and yay for more measurements!

 

[thewas]

First of all, thank you for the very welcome measurements , one question though.

I did not capture vertical measurements but we can safely assume vertical directivity to be rather poor.

Aren't the verticals needed for the spinorama plots, for example power and early reflection responses?

 

[Weeb Labs]

First of all, thank you for the very welcome measurements , one question though.

Aren't the verticals needed for the spinorama plots, for example power and early reflection responses?

I am actually not entirely certain but you have succeeded in instilling just enough doubt for me to capture the verticals tomorrow.

 

[thewas]

I am actually not entirely certain but you have succeeded in instilling just enough doubt for me to capture the verticals tomorrow.

Excellent , these references might also become handy:

How to make quasi-anechoic speaker measurements/spinoramas with REW and VituixCAD

Quasi-anechoic measurements are a way to capture the 'real' frequency response of a speaker without the influence of a room. They are called 'quasi-anechoic' because they have the potential to get you really close to the results one might achieve with a non-reflective environment like an...

www.audiosciencereview.com

How to make quasi-anechoic speaker measurements/spinoramas with REW and VituixCAD

Quasi-anechoic measurements are a way to capture the 'real' frequency response of a speaker without the influence of a room. They are called 'quasi-anechoic' because they have the potential to get you really close to the results one might achieve with a non-reflective environment like an...

www.audiosciencereview.com

 

[Maiky76]

I recently purchased a pair of Mackie MR524s and thought it might be helpful to capture a set of essential measurements. To my knowledge, none are currently available.

The MR524 is sold individually for approximately €100, which places it within the same price bracket as the JBL 305P, Adam T5V and KRK Rokit 5. I find them to be rather aesthetically pleasing and many of the product images online fail to do them justice.

I love the inverted dust cap!

View attachment 171403
The rear is populated with a handy selection of RCA, TRS and XLR inputs. There are also three onboard PEQ settings, a three-position HF adjustment switch and input gain control with the unity point marked.

Firstly, we have the on-axis and off-axis measurements. These are a composite of nearfield and gated methods, which should be reasonably comparable to measurements acquired using a Klippel NFS or anechoic chamber.

View attachment 171406

On-axis response is predominantly flat, with a little bit of additional bass and a bump around 10KHz. Directivity is excellent, which will enable effective correction of the on-axis response via EQ or DRC.

Next, we have nearfield measurements of all driver components.

View attachment 171412

I suspect the peak at 1KHz to be leakage from the midbass rather than a resonance but it is difficult to know for certain. It does not appear to affect the on-axis or off-axis response, so the former seems more likely. Even so, you may wish to plug the port if using these monitors in conjunction with a subwoofer.

Finally, polar and line charts for horizontal directivity.

View attachment 171413

View attachment 171414

I did not capture vertical measurements but we can safely assume vertical directivity to be rather poor. Distortion and/or vertical measurements may be added at a later date.

Subjective impressions were very positive. The 10KHz bump is readily apparent but easily corrected with a gentle filter and at that point, they sound just as lovely as my T5Vs and LSR 305s. Quiet Light from Hanaregumi is one of my favorite test tracks and it sounds delightful on these; especially in conjunction with Audiolense correction!

Thank you for reading and I hope this thread has been of help to you.

Hi,

Great effort!
I think the data is however unusual, the LW is very (too?) close to the ON.
The different curves have unusual relationships to each other, SP vs ON for example. i.e. the DI.
I am not sure the data is correct below 400Hz.
The cross over look particularly perfect... I'd like to see the Vertical data to check.
The score may not be comparable to the NFS derived data.
Again great effort nonetheless!

Here is my take on the EQ.


These EQ are anechoic EQ to get the speaker right before room integration. If you able to implement these EQs you must add EQ at LF for room integration, that is usually not optional… see hints there: https://www.audiosciencereview.com/...helf-speaker-review.11144/page-26#post-800725

The raw data with:

Score no EQ: 4.3
With Sub: 6.3

Spinorama with no EQ:

EQ design:

I have generated one EQ. The APO config file is attached.
The LW and Score EQ are too close to differentiate.
Not sure if the EQ would translate on may units.

Score LW/EQ Score: 6.3
with sub: 8.3

Mackie MR524 by Weeb Labs APO LW EQ 96000Hz
December102021-132609

Preamp: -1.5 dB

Filter 1: ON PK Fc 100.72,    -2.63,    1.29
Filter 2: ON PK Fc 932.68,    -2.46,    6.62
Filter 3: ON PK Fc 1263.70,    1.59,    3.15
Filter 4: ON PK Fc 2556.17,    0.62,    2.13
Filter 5: ON PK Fc 4171.60,    -1.48,    6.27
Filter 6: ON PK Fc 5724.70,    -1.02,    4.18
Filter 7: ON PK Fc 9769.89,    -4.39,    3.73
Filter 8: ON PK Fc 16795.89,    -3.70,    5.16

Spinorama EQ Score

Zoom PIR-LW-ON

Regression - Tonal

Radar no EQ vs EQ score
Nice improvements

 

[Weeb Labs]

Hi,

Great effort!
I think the data is however unusual, the LW is very (too?) close to the ON.
The different curves have unusual relationships to each other, SP vs ON for example. i.e. the DI.
I am not sure the data is correct below 400Hz.
The cross over look particularly perfect... I'd like to see the Vertical data to check.
The score may not be comparable to the NFS derived data.
Again great effort nonetheless!

Here is my take on the EQ.


These EQ are anechoic EQ to get the speaker right before room integration. If you able to implement these EQs you must add EQ at LF for room integration, that is usually not optional… see hints there: https://www.audiosciencereview.com/...helf-speaker-review.11144/page-26#post-800725

The raw data with:

Score no EQ: 4.3
With Sub: 6.3

Spinorama with no EQ:
View attachment 171436

EQ design:

I have generated one EQ. The APO config file is attached.
The LW and Score EQ are too close to differentiate.
Not sure if the EQ would translate on may units.

Score LW/EQ Score: 6.3
with sub: 8.3

Mackie MR524 by Weeb Labs APO LW EQ 96000Hz
December102021-132609

Preamp: -1.5 dB

Filter 1: ON PK Fc 100.72,    -2.63,    1.29
Filter 2: ON PK Fc 932.68,    -2.46,    6.62
Filter 3: ON PK Fc 1263.70,    1.59,    3.15
Filter 4: ON PK Fc 2556.17,    0.62,    2.13
Filter 5: ON PK Fc 4171.60,    -1.48,    6.27
Filter 6: ON PK Fc 5724.70,    -1.02,    4.18
Filter 7: ON PK Fc 9769.89,    -4.39,    3.73
Filter 8: ON PK Fc 16795.89,    -3.70,    5.16

View attachment 171431

Spinorama EQ Score
View attachment 171435

Zoom PIR-LW-ON
View attachment 171432

Regression - Tonal
View attachment 171433

Radar no EQ vs EQ score
Nice improvements
View attachment 171434

It may be the absence of vertical measurements that has resulted in these oddities. I am quite certain that the response below 400Hz is accurate, as it is a nearfield measurement with baffle step transform and follows my ungated on-axis tests quite well. We'll see what happens tomorrow.

 

[abdo123]

I am actually not entirely certain but you have succeeded in instilling just enough doubt for me to capture the verticals tomorrow.

You need just till -\+ 40 degrees so like 8 more measurements.

 

[ctrl]

I am not sure the data is correct below 400Hz.

I am quite certain that the response below 400Hz is accurate, as it is a nearfield measurement with baffle step transform and follows my ungated on-axis tests quite well.

Have to agree with @Maiky76. The simulation/calculation of the frequency responses below 400Hz is wrong.

A BR speaker of this size hardly shows any directivity below 100Hz - apart from active cardioid loudspeakers.



Quick guide for VCAD

I assume that the near field measurements were correctly summed up.
Then you have to generate all frequency responses (hor and ver) with baffle step correction in the diffraction tool of VCAD.

Open your combined nearfield measurement. Pay attention to the correct position of woofer and mic. Enter the correct measurement distance. Check all the needed buttons.


Then start the VCAD merger tool and open the just created baffle step corrected frequency responses (I do hor and ver FR separately) in the low frequency part window. Check far field button. Check correct step wide.

In the high frequency part window open the gated measurements.

If you did everything right (and set scale and delay if necessary), you should get a transition frequency range (depending on your gate, mostly between 200Hz and 400Hz) where near field and far field measurements are congruent (frequency response and phase frequency response).
Save merged FR.

These merged FR then show approximately realistic frequency responses for the low frequency range.
Here is an example with a 0.45m wide baffle:

 

[Morpheus]

Hi, why don't you think that disturbance in frequency and directivity at just below 1Khz is not port related? From the nearfield driver ( not present) and port measurements ( pretty big) it looks like the most probable cause to me..

 

[Weeb Labs]

Have to agree with @Maiky76. The simulation/calculation of the frequency responses below 400Hz is wrong.

A BR speaker of this size hardly shows any directivity below 100Hz - apart from active cardioid loudspeakers.
View attachment 171464


Quick guide for VCAD

I assume that the near field measurements were correctly summed up.
Then you have to generate all frequency responses (hor and ver) with baffle step correction in the diffraction tool of VCAD.
View attachment 171465
Open your combined nearfield measurement. Pay attention to the correct position of woofer and mic. Enter the correct measurement distance. Check all the needed buttons.

Then start the VCAD merger tool and open the just created baffle step corrected frequency responses (I do hor and ver FR separately) in the low frequency part window. Check far field button. Check correct step wide.

In the high frequency part window open the gated measurements.
View attachment 171470
If you did everything right (and set scale and delay if necessary), you should get a transition frequency range (depending on your gate, mostly between 200Hz and 400Hz) where near field and far field measurements are congruent (frequency response and phase frequency response).
Save merged FR.

These merged FR then show approximately realistic frequency responses for the low frequency range.
Here is an example with a 0.45m wide baffle:
View attachment 171471

I performed the baffle transform using Bagby's simulator as described here, which should produce results very similar to those of VCAD. After capturing the vertical measurements this evening, I will repeat the transform using VCAD and check for other potential issues along the way.

 

[ctrl]

I performed the baffle transform using Bagby's simulator as described here, which should produce results very similar to those of VCAD. After capturing the vertical measurements this evening, I will repeat the transform using VCAD and check for other potential issues long the way.

It is not sufficient to perform the baffle step correction only for on-axis near-field measurement.

All angular frequency responses of the near-field measurement must be corrected accordingly. VCAD generates these corrected frequency responses automatically.
Perhaps this is not clearly expressed in the tutorial by @napilopez?

 

[Weeb Labs]

It is not sufficient to perform the baffle step correction only for on-axis near-field measurement.

All angular frequency responses of the near-field measurement must be corrected accordingly. VCAD generates these corrected frequency responses automatically.
Perhaps this is not clearly expressed in the tutorial by @napilopez?

It doesn't appear to have been mentioned, as far as I can see. Given that the off-axis measurements are windowed to 3.5ms (only valid to about 250Hz) and blended to the nearfield bass measurement at about 550Hz, the effect of applying baffle step to them seems unclear to me. The transform doesn't do very much at that frequency.

 

[ctrl]

blended to the nearfield bass measurement at about 550Hz, the effect of applying baffle step to them seems unclear to me. The transform doesn't do very much at that frequency.

The transition frequency between near-field and far-field measurement is almost a bit high at 550Hz. I would assume that the curves show best congruent behavior between 200-400Hz.

The baffle step correction reduces the low bass range by about 6dB compared to the higher frequencies where the baffle still provokes half-space radiation. In addition, there are interference effects due to the edge diffraction...
Thus, if the transition frequency is 550Hz, the baffle step correction has a particularly strong effect on the near-field measurement.

Under different angles the "effective" baffle area changes, therefore the baffle step correction is calculated for each angle (step).

 

[Weeb Labs]

The transition frequency between near-field and far-field measurement is almost a bit high at 550Hz. I would assume that the curves show best congruent behavior between 200-400Hz.

The baffle step correction reduces the low bass range by about 6dB compared to the higher frequencies where the baffle still provokes half-space radiation. In addition, there are interference effects due to the edge diffraction...
Thus, if the transition frequency is 550Hz, the baffle step correction has a particularly strong effect on the near-field measurement.

Under different angles the "effective" baffle area changes, therefore the baffle step correction is calculated for each angle.

I have just identified the problem. The version of VCAD used by @napilopez in the creation of his guide differs from the current version in one important respect. The current build includes a "Force to gradient" function within the "Merge" window and it is disabled by default. It had previously been the default behavior (no setting).

Updating the OP, now.

I will capture the vertical measurements this evening.

 

[napilopez]

Apologies for any confusion that might've been caused by my guide or lack of clarity. Will comment on a few things:

-I missed the part about not capturing vertical measurements on my original lookthrough. Unfortunately that does invalidate the spinorama quite strongly, for everything except the On-axis and a bit of the listening window. If the issue is saving time you'd get more accurate results by using fewer horizontal angles with a few key vertical ones. the general trend might be similar, there will certainly be a crossover dip somewhere in there. that's not currently present.

--@ctrl's above guide for LF directivity appears to be more thorough and accurate than my own; I have not previously calculated LF data this way.

In my guide, I simply calculated baffle step for the on-axis response; in this case, using the merger tool meant Vituixcad would estimate the LF directivity for off-axis angles based on LF trends in the gated data.

However, depending on how wide your gate is and how noisy your data is, this could still lead to less than ideal results. This is why in my reviews I normally mention that off-axis data bellow 200 Hz or so may not be accurate. It was usually 'good enough.' but I do mention near the end of my guide that "Sometimes the off-axis bass just doesn't look right for whatever reason. In these situations, I prefer to cut off the data below 200Hz."
However that is a bit of a lame thing to say, because that assumes the reader would know when the bass looks off.

Per VituixCAD's old manual:

"Merger tool merges near field measurements + diffraction simulation or 2π + diffraction simulation or 4π simulation to time-windowed far field measurements. Merged off-axis responses contain directivity information below transition frequency based on time-windowed axial response divided by time-windowed off-axis response. Reliable directivity information at low frequencies requires long time window."

However it seems there've recently been some important updates which are great to see. The new manual and new versions of VCAD show way more powerful tools that should significantly mitigate these problems even if not 100% accurate. I've not used them yet -- it's been months since I last measured a speaker but it looks really interesting:

"
'Force to Gradient' checkbox and 'Monopole portion' text box enable forcing ideal gradient directivity to near field LF responses if directivity of far field HF responses is too unreliable at LF due to window function or measurement conditions, and directivity at LF can be predicted without measurements.

Monopole portion values for known gradient radiators: Omni 100 % (DI 0 dB), Cardioid 50 % (DI 4.8 dB), Super-cardioid 37 % (DI 5.7 dB), Hyper-cardioid 25 % (DI 6.0 dB), Dipole 0 % (DI 4.8 dB). Directivity of far field HF responses is used at LF when 'Force to Gradient' is unchecked.
Highest frequency using ideal gradient pattern is adjusted with 'below Hz' text box. Directivity is blended within ideal gradient range and transition frequency. Narrowest blending range is one octave."

Based on this it actually appears the old default was not the Force to gradient option, but it looks like a very useful new set of features.You can also now specify if the radiator/port is located on the front of the baffle to alter its contribution to the summed response for even more accurate results.

Basically, I'll have to explore these tools and upgrade my guide accordingly; though the bagby results have long served me as very close to the real results, it would be good to do the full baffle step correction in vituixcad as it is simply more thorough.

--My experience is that not including the vertical measurements will also worsen the predicted spinorama in the LF so looking forward to seeing this =]

 

[Weeb Labs]

Apologies for any confusion that might've been caused by my guide or lack of clarity. Will comment on a few things:

-I missed the part about not capturing vertical measurements on my original lookthrough. Unfortunately that does invalidate the spinorama quite strongly, for everything except the On-axis and a bit of the listening window. If the issue is saving time you'd get more accurate results by using fewer horizontal angles with a few key vertical ones. the general trend might be similar, there will certainly be a crossover dip somewhere in there. that's not currently present.

--@ctrl's above guide for LF directivity appears to be more thorough and accurate than my own; I have not previously calculated LF data this way.

In my guide, I simply calculated baffle step for the on-axis response; in this case, using the merger tool meant Vituixcad would estimate the LF directivity for off-axis angles based on LF trends in the gated data.

However, depending on how wide your gate is and how noisy your data is, this could still lead to less than ideal results. This is why in my reviews I normally mention that off-axis data bellow 200 Hz or so may not be accurate. It was usually 'good enough.' but I do mention near the end of my guide that "Sometimes the off-axis bass just doesn't look right for whatever reason. In these situations, I prefer to cut off the data below 200Hz."
However that is a bit of a lame thing to say, because that assumes the reader would know when the bass looks off.

Per VituixCAD's old manual:

"Merger tool merges near field measurements + diffraction simulation or 2π + diffraction simulation or 4π simulation to time-windowed far field measurements. Merged off-axis responses contain directivity information below transition frequency based on time-windowed axial response divided by time-windowed off-axis response. Reliable directivity information at low frequencies requires long time window."

However it seems there've recently been some important updates which are great to see. The new manual and new versions of VCAD show way more powerful tools that should significantly mitigate these problems even if not 100% accurate. I've not used them yet -- it's been months since I last measured a speaker but it looks really interesting:

"
'Force to Gradient' checkbox and 'Monopole portion' text box enable forcing ideal gradient directivity to near field LF responses if directivity of far field HF responses is too unreliable at LF due to window function or measurement conditions, and directivity at LF can be predicted without measurements.

Monopole portion values for known gradient radiators: Omni 100 % (DI 0 dB), Cardioid 50 % (DI 4.8 dB), Super-cardioid 37 % (DI 5.7 dB), Hyper-cardioid 25 % (DI 6.0 dB), Dipole 0 % (DI 4.8 dB). Directivity of far field HF responses is used at LF when 'Force to Gradient' is unchecked.
Highest frequency using ideal gradient pattern is adjusted with 'below Hz' text box. Directivity is blended within ideal gradient range and transition frequency. Narrowest blending range is one octave."

Based on this it actually appears the old default was not the Force to gradient option, but it looks like a very useful new set of features.You can also now specify if the radiator/port is located on the front of the baffle to alter its contribution to the summed response for even more accurate results.

Basically, I'll have to explore these tools and upgrade my guide accordingly; though the bagby results have long served me as very close to the real results, it would be good to do the full baffle step correction in vituixcad as it is simply more thorough.

--My experience is that not including the vertical measurements will also worsen the predicted spinorama in the LF so looking forward to seeing this =]

Thanks for your input! I've been doing some further reading into CEA2034-A and you're definitely correct that the verticals are essential. I'll be adding those after work and then we'll see what it looks like.

It bears repeating that we are targeting "close enough" with these measurements and not "mathematically perfect". If we're within a couple of decibels of an NFS spin, that is more than adequate as a basis for assessment by prospective buyers; especially when unit tolerances are taken into consideration.