Let's see what you have under (and around) the hood - bass management and DSP challenge

[Oddball]

Well a bit of a challenge for all. No real prize, just fame a glory. My best bet is that it will not go to any ART system as some members have gone beyond that, but with such significant resources that are prohibitive for the most.

We had a lots of positive Dirac ART threads and posts lately and they have been questioned on many grounds. I think there is only one ground that has merit apart from members doing more complicated tests and comparisons that professional reviewers do. Show us your systems and FQ response graph as well as decay graph. If we can get this information and not get into impulse, group delay etc, it will take us a long way without taking any side roads. In my personal view, Dirac ART made my system and graphs look as good as they could. Always room for improvement, but let's see what others have.

I am sure that this will turn into a difficult discussion, but I see so many bass management related threads where people seem to struggle and perhaps not aware of the new ways to manage their issue. So to simplify, I just decided to make it - well simple - what is the best you have and don't question the others as to how get there or require further experiments if they can make it better or different. I admittedly don't know much about active speakers, which I admire based on what I heard, but not enough to understand how would e.g. 9.4.4. system work under these circumstances. So looking forward to any knowledge through graphs and photos to lean about the new ways.

Without much further BS, here is the beef. Admittedly areas to address, but lol, If you have it this good, you would probably be as lazy as I am...

EDIT as gear posts from other PC were more available.

Really looking forward to constructive discussion, but already know that it will be disrupted by the inquisitors, that if we were on IG would call haters.

 

[kawauso]

Beyond decay, the multichannel consistency below 200 Hz within the ART domain looks truly impressive.
It really seems to have a positive impact on the overall coherence of the sound.

English isn’t exactly my strength, but I got so fascinated that I ended up writing this article about ART.
It’s not monetized or sponsored by Dirac in any way — just pure curiosity and enthusiasm.

Dirac ART: Beyond Room Correction

『Dirac ART: Beyond Room Correction』

Introduction Dirac ART (Active Room Treatment) is not about “correctin…

ameblo.jp

Dirac ART Field Report: JBL Room in 2.0 Mode

『Dirac ART Field Report: JBL Room in 2.0 Mode』

Table of Contents Overview Effect in 2.0 Configuration Spectrum Changes Subwoof…

ameblo.jp

When Dirac ART Freed My KEF Room

『When Dirac ART Freed My KEF Room』

Oct 25, 2025 · User ReviewTable of Contents Measurements & First Impressions Li…

ameblo.jp

How to Set Up Dirac ART on Denon & Marantz

『How to Set Up Dirac ART on Denon & Marantz』

Table of Contents Preparation Tuning Subwoofer Grouping LFE Group Setup Main Sp…

ameblo.jp

 

[ppataki]

Well a bit of a challenge for all. No real prize, just fame a glory. My best bet is that it will not go to any ART system as some members have gone beyond that, but with such significant resources that are prohibitive for the most.

We had a lots of positive Dirac ART threads and posts lately and they have been questioned on many grounds. I think there is only one ground that has merit apart from members doing more complicated tests and comparisons that professional reviewers do. Show us your systems and FQ response graph as well as decay graph. If we can get this information and not get into impulse, group delay etc, it will take us a long way without taking any side roads. In my personal view, Dirac ART made my system and graphs look as good as they could. Always room for improvement, but let's see what others have.

I am sure that this will turn into a difficult discussion, but I see so many bass management related threads where people seem to struggle and perhaps not aware of the new ways to manage their issue. So to simplify, I just decided to make it - well simple - what is the best you have and don't question the others as to how get there or require further experiments if they can make it better or different. I admittedly don't know much about active speakers, which I admire based on what I heard, but not enough to understand how would e.g. 9.4.4. system work under these circumstances. So looking forward to any knowledge through graphs and photos to lean about the new ways.

Without much further BS, here is the beef. Admittedly areas to address, but lol, If you have it this good, you would probably be as lazy as I am...

View attachment 490740View attachment 490741View attachment 490742View attachment 490743View attachment 490744

EDIT as gear posts from other PC were more available.

Really looking forward to constructive discussion, but already know that it will be disrupted by the inquisitors, that if we were on IG would call haters.

I don't have Dirac ART, just the normal Dirac Live but I configured my two DIY sealed subs manually so Dirac 'sees' them integrated with the mains

Here is the FR (Psy smoothing) and the Wavelet (happy to post the Spectrogram too but not sure what exact settings you used to generate that)

 

[Oddball]

I don't have Dirac ART, just the normal Dirac Live but I configured my two DIY sealed subs manually so Dirac 'sees' them integrated with the mains

Here is the FR (Psy smoothing) and the Wavelet (happy to post the Spectrogram too but not sure what exact settings you used to generate that)

View attachment 490798

View attachment 490800

Well just used the REW default for spectrogram. I have a new instal for ART as changed laptops. Showed the FR response and decay across the board in the ranges shows - If anything else needed happy to add if I can.

From your graph it seems that you have managed way better than ART with a bump at 300hz and 100 ms. (with your scale being inverted to mine horizontal vs vertical). I have been following your posts and that does not surprise me as you have done so much work - but yet such response? Would be really helpful to see your room and gear.

Another potentially contentious points we will encounter will be the dB at which we sweep. Subs will obviously do more delay if turned up, not sure how much of that can be EQs. That's just the way they work. But let's not turn this into thread like that as then we will not get to many conclusions.

This is a great curve and delay is incredible. So regardless how loud can they go, it is an amazing accomplishment.

 

[ppataki]

This is a great curve and delay is incredible.

Thank you, much appreciated!

Would be really helpful to see your room and gear.

You can see them in the other posts:

Here is the room (+ the project thread for my current mains)

3D printed classic 3-way budget speaker project using SB and Dayton drivers

The woofer's cabinet is also done! And some knock test, as always: https://photos.app.goo.gl/vsZwHqL5fQb29oRn8 Unfortunately there is some air leakage issue that I will need to fix tomorrow + I still need to print some holders I might be able to make all the 3-ways function by Sunday...

www.audiosciencereview.com

Here is the project thread for my front sub:

Lavoce SSF153.00 15" based dual opposed subwoofer project

I am excited to share my next project here! :) After converting my MCLAs to a sub I have decided to look for a more traditional approach to subs - hence the below project came to my mind I have designed and built several dual opposed subwoofers (DOS) already - I can say I really love this design...

www.audiosciencereview.com

And here is the project thread for my rear sub:

Using four bass guitar drivers to build a dual opposed subwoofer project

Even the title is unorthodox :) I have recently built a dual opposed subwoofer (DOS) for my home office which was a hugely successful project for me. Once it was done I have taken it downstairs in my living room and started playing around with what I can do with two subs instead of just one...

www.audiosciencereview.com

+ here is a bonus project from my home office, with similarly low delay results:

Another 8" dual opposed subwoofer project

After finishing my new DIY system in my home office/bedroom I have decided to also add (build) a subwoofer to it I have really huge dips in the frequency response due to room modes; see the before/after DSP correction curve below: When running REW simulations for adding a sub to the system...

www.audiosciencereview.com

Based on my experience in the last 5-6 years playing around a lot with my room(s) and changing gears very frequently, in order to get to this really low delay, you need:
- one or more sealed sub(s) - preferably DIY since then you can be flexible with the driver and amp selection too
- strategic placement in the room based on simulation (REW Room Sim) and even trial and error measurements (in case your room cannot be simulated, like mine)
- manual integration with the mains (I personally use the Acoustic Timing Reference method with REW): this includes proper crossover settings (LR 24dB/oct for example), volume, delay and polarity settings
- and finally Dirac Live (or whatever that is equivalent)

When (or actually IF, ever) Dirac ART ever becomes available on a PC, I will try it immediately to see what results it will produce with my gear
But I am already extremely happy with these results

 

[Keith_W]

Don't forget you also need to look at the step response - both the head and the tail.

Here's my overall FR - L, R, and vector sum.

The head of the step response shows how well the drivers and subwoofers have been integrated. What you want to see: (1) no pre-ringing before the main impulse, (2) nice, sharp step response, (3) L-R symmetry.

The tail of the step response examines the room behaviour of the impulse. Here I am showing the first 450ms of the step response. What you want to see: (1) rapid decay, (2) no subsequent peaks higher than the main impulse, (3) left-right symmetry.

The spectrogram (Fourier mode) should be zoomed to an appropriate time scale to show the longest decay AND normalized to peak at every frequency.

Everything you see here has been hand fettled with DSP.

How to display and interpret these measurements described in Book 2 of the REW eBook.

 

[kawauso]

Don't forget you also need to look at the step response - both the head and the tail.

View attachment 490836

Here's my overall FR - L, R, and vector sum.

View attachment 490838

The head of the step response shows how well the drivers and subwoofers have been integrated. What you want to see: (1) no pre-ringing before the main impulse, (2) nice, sharp step response, (3) L-R symmetry.

View attachment 490837

The tail of the step response examines the room behaviour of the impulse. Here I am showing the first 450ms of the step response. What you want to see: (1) rapid decay, (2) no subsequent peaks higher than the main impulse, (3) left-right symmetry.

View attachment 490839

The spectrogram (Fourier mode) should be zoomed to an appropriate time scale to show the longest decay AND normalized to peak at every frequency.

Everything you see here has been hand fettled with DSP.

How to display and interpret these measurements described in Book 2 of the REW eBook.

Keith, could you take a look at my MDATs?
• KEF Room – BC and ART(with an ART low-frequency cut mistake)

BCART.mdat

drive.google.com

• JBL Room – RC and ART in 2.0ch

2ch.mdat

drive.google.com

• JBL Room at MLP and off-MLP door side

main vs by the door.mdat

drive.google.com

 

[ban25]

Cool thread topic!

There's too much rain right now for me to take a fresh measurement, but this is my most recent one and there have been no notable changes to the system in that time, beyond being on a now slightly older version of Dirac + Bass Control from when this was taken (May 2024):

Graphs:

System:

Specs:
Pre-Amp:
- Integra DRX-8.4
- Dirac Live + DLBC
- McIntosh C53 (Passthrough)
Amp:
- McIntosh MC302 (FL/FR)
- Fosi ZA3 (C)
Speakers (5.2.4):
- KEF Blade Two Meta (FL/FR)
- KEF R2c (C)
- KEF R3 (SL/SR)
- KEF Ci200RR-THX (Atmos)
- KEF KC92 (Sub1)
- KEF KF92 (Sub2)

 

[kawauso]

Cool thread topic!

There's too much rain right now for me to take a fresh measurement, but this is my most recent one and there have been no notable changes to the system in that time, beyond being on a now slightly older version of Dirac + Bass Control from when this was taken (May 2024):

Graphs:

View attachment 490849
View attachment 490850

System:
View attachment 490851
View attachment 490852
View attachment 490853
View attachment 490854
View attachment 490855
View attachment 490856

Specs:
Pre-Amp:
- Integra DRX-8.4
- Dirac Live + DLBC
- McIntosh C53 (Passthrough)
Amp:
- McIntosh MC302 (FL/FR)
- Fosi ZA3 (C)
Speakers (5.2.4):
- KEF Blade Two Meta (FL/FR)
- KEF R2c (C)
- KEF R3 (SL/SR)
- KEF Ci200RR-THX (Atmos)
- KEF KC92 (Sub1)
- KEF KF92 (Sub2)

Oh wow, Blade and McIntosh — that looks absolutely stunning! ❤
I really like that touch of mid-high reverberation; it’s exactly my kind of sound.

 

[ban25]

Oh wow, Blade and McIntosh — that looks absolutely stunning! ❤
I really like that touch of mid-high reverberation; it’s exactly my kind of sound.

Thanks! When I want dedicated 2-channel without DSP, I can go straight through the McIntosh Pre-Amp for e.g. vinyl or Roon. The 12V trigger puts it into passthrough mode when my AVR is turned on and I have dedicated Dirac slots for single listener and group seating. It's impressive how much bass output I can get just out of the Blades by themselves, especially with a bit of EQ (either on the McIntosh or in Roon).

 

[Keith_W]

Keith, could you take a look at my MDATs?

Why don't you download that eBook and look at them yourself? That's what the eBook is for. I would like people to have a go at reading their own measurements instead of relying on the ASR community to read it for you. There is far more detail in that eBook than is possible for me to keep typing in forum posts over and over. I'm happy to look over your measurement after you've had a go interpreting them to point out things that you may have missed or incorrectly interpreted. It's very simple - choose the same views that the book suggests you to choose, and look at features of the measurement. There is no need to learn how to read every measurement at once - just focus on a few and try to understand what they mean.

You have taken individual L and R measurements. If your intention is to compare BC to ART, the first thing to realize is that these bass management systems are not designed to work on their own, it is the sum of all speakers and subwoofers that produces the flat response. So the first step is to vector sum the BC L+R and ART L+R to look at the combined bass response, then compare them to see if there is an improvement. These were taken from the KEF series. Yellow = ART, Green = BC.

Frequency response.

Step response head and tail.

Spectro. ART on left, BC on right. Note the view I have chosen - Fourier, restricted from 20Hz - 400Hz, time up to 600ms (since that is the maximum decay on your measurement ... if it was up to 900ms I would have shown that instead). And also - normalized to peak at every frequency.

And this is the ETC of your left speaker only.

As you can see, once you have chosen appropriate views that demonstrates what you want to show, interpretation is simple. You have some problems with your system. What do you think they are?

 

[bmc0]

Here's my setup with two subs and manual correction (5 biquads):

The second plot is the same data but normalized.
Magnitude response (spatial average; 1/12th oct smoothing):

 

[kawauso]

Why don't you download that eBook and look at them yourself? That's what the eBook is for. I would like people to have a go at reading their own measurements instead of relying on the ASR community to read it for you. There is far more detail in that eBook than is possible for me to keep typing in forum posts over and over. I'm happy to look over your measurement after you've had a go interpreting them to point out things that you may have missed or incorrectly interpreted. It's very simple - choose the same views that the book suggests you to choose, and look at features of the measurement. There is no need to learn how to read every measurement at once - just focus on a few and try to understand what they mean.

You have taken individual L and R measurements. If your intention is to compare BC to ART, the first thing to realize is that these bass management systems are not designed to work on their own, it is the sum of all speakers and subwoofers that produces the flat response. So the first step is to vector sum the BC L+R and ART L+R to look at the combined bass response, then compare them to see if there is an improvement. These were taken from the KEF series. Yellow = ART, Green = BC.

View attachment 490866

Frequency response.

View attachment 490868View attachment 490867

Step response head and tail.

View attachment 490869View attachment 490870

Spectro. ART on left, BC on right. Note the view I have chosen - Fourier, restricted from 20Hz - 400Hz, time up to 600ms (since that is the maximum decay on your measurement ... if it was up to 900ms I would have shown that instead). And also - normalized to peak at every frequency.

View attachment 490872

And this is the ETC of your left speaker only.

As you can see, once you have chosen appropriate views that demonstrates what you want to show, interpretation is simple. You have some problems with your system. What do you think they are?

Am I looking at the correct part?
The early-reflection peak on the left channel doesn’t really change with ART.
My understanding is that ART improves bass decay but does not alter early reflections themselves.

I don’t have L+R vector-sum data for the KEF room yet, so I’ll check that later.

However, in the JBL room, the L+R vector-sum changes shape quite a lot between RC and ART.
The left-only and right-only ETCs do not change much, but the combined L+R result does.

How should I interpret this difference?

 

[Keith_W]

Am I looking at the correct part?
The early-reflection peak on the left channel doesn’t really change with ART.
My understanding is that ART improves bass decay but does not alter early reflections themselves.

That is correct. No DSP can change reflection peaks. You can not move walls with DSP.

I don’t have L+R vector-sum data for the KEF room yet, so I’ll check that later.

You create it. Go to "All SPL", choose the measurements that you want, then right click in the window and choose "Vector Sum".

How should I interpret this difference?

View attachment 490878
View attachment 490879
View attachment 490880

There are no labels on your graphs so I don't know what you are looking at. That peak at 0.8ms you circled in your first graph is an early reflection, and the yellow line is decay.

 

[kawauso]

That is correct. No DSP can change reflection peaks. You can not move walls with DSP.



You create it. Go to "All SPL", choose the measurements that you want, then right click in the window and choose "Vector Sum".



There are no labels on your graphs so I don't know what you are looking at. That peak at 0.8ms you circled in your first graph is an early reflection, and the yellow line is decay.

Thanks, Keith. I'm not very good at English, so reading e-books is pretty hard for me, but I was able to learn about Envelope and Vector Sum today.

 

[kawauso]

I measured the JBL room. ART does a great job integrating the four subwoofers with different characteristics compared to BC. I was also able to confirm that the decay has become much shorter.

 

[kawauso]

And I also measured at a position far off the MLP. BC looks pretty bad in terms of FR there, but ART is totally fine to me.

 

[Bjorn]

I would prefer to see a waterfall below about 300 Hz and before and after the correction. RT60 is invalid for small room acoustics since there no real diffuse field.

 

[Oddball]

Cool thread topic!

There's too much rain right now for me to take a fresh measurement, but this is my most recent one and there have been no notable changes to the system in that time, beyond being on a now slightly older version of Dirac + Bass Control from when this was taken (May 2024):

Graphs:

View attachment 490849
View attachment 490850

System:
View attachment 490851
View attachment 490852
View attachment 490853
View attachment 490854
View attachment 490855
View attachment 490856

Specs:
Pre-Amp:
- Integra DRX-8.4
- Dirac Live + DLBC
- McIntosh C53 (Passthrough)
Amp:
- McIntosh MC302 (FL/FR)
- Fosi ZA3 (C)
Speakers (5.2.4):
- KEF Blade Two Meta (FL/FR)
- KEF R2c (C)
- KEF R3 (SL/SR)
- KEF Ci200RR-THX (Atmos)
- KEF KC92 (Sub1)
- KEF KF92 (Sub2)

Really great looking and performing system. What a feat for eyes and ears.

 

[Oddball]

We have some great stuff cumming in. Really glad that people want to pursue this topic and thanks to all for contributing. Would love to see some multi-channel active systems as well.

As I noted, I personally find it less important to post pre and after graphs. The idea is to show the best you have and how your got there. I also wanted to limit discussion to FQ response and decay. That really matters most, at least to me. But obviously people are free to post what they want and any discussion will be interesting.

FQ response is an obvious one that does not need much discussion. With recent reduction of decay on ART (which probably many experienced before through different tools), there seems to be a point of "dry" bass which less decay creates, vs. "wet" bass that most previous EQ solutions offered. Don't think there is universal answer to what works best according to one's preferences. But it is always possible to increase decay to one's preference if the system provides ways to reduce decay to "dry" bass territory. Other way around is much more difficult.