Neumann MA1 vs. Dirac Live Bass Control (DLBC) - Dual KH750 subs

[dziemian]

You are avoiding my question.

I asked those question to creator of Audiolense, Acourate and Deqx. I dont own Neumann Systems but I am Pręty confidential it is no diffrent to mentioned above. You may prove me wrong.

 

[DJBonoBobo]

I asked those question to creator of Audiolense, Acourate and Deqx. I dont own Neumann Systems but I am Pręty confidential it is no diffrent to mentioned above. You may prove me wrong.

No. You have to prove what you say. I provided measurements that show interactions between subs, as far as i understand them. I am no expert.
If you claim you know better you have to provide solid arguments, or it is just lazy shit talk.

 

[dziemian]

No. You have to prove what you say. I provided measurements that show interactions between subs, as far as i understand them. I am no expert.
If you claim you know better you have to provide solid arguments, or it is just lazy shit talk.

Yes. You are right. Enjoy your system

 

[192kbps]

The new multichannel extension for MA1 supports up to 4 subs, but with additional cost. I wonder how this and the whole multichannel thing compares to DLBC.
Hopefully someone will try it.

4 subs? I also heard that we are preparing to update the surround sound calibration.

 

[DJBonoBobo]

4 subs? I also heard that we are preparing to update the surround sound calibration.

Multichannel Extension for MA 1

Precision alignment for truly immersive sound that translates reliably to all playback systems

www.neumann.com

Available for around 300 EUR (additionally to normal MA1).

 

[192kbps]

Multichannel Extension for MA 1

Precision alignment for truly immersive sound that translates reliably to all playback systems

www.neumann.com

Available for around 300 EUR (additionally to normal MA1).

My goodness, 7.1.4, they finally updated it, great.

 

[palm]

Dirac and multi sub takes into account all the subs available and how they interact with each other. Acourate, audiolense,deqx,neumann measure aech sub channel individually.

Not sure if this is an accurate or useful categorization but here is how I see the different flavors of subwoofer integrations:

1. no filtering of the speakers, adjust low pass and volume of the sub to blend
2. high pass filtering of speakers, sub is like an extra way with proper crossover, and EQ is stereo full range, not 3 channels
3. more than one sub, but each receive the same signal and EQ, just with different delays
4. each sub is EQ’d to flat separately
5. each sub has separate EQ but the target isn’t to be individually flat but for the sum to be flat.

For 5. I guess this is because of an asymmetry in how peaks and nulls are considered when creating a filter, so 4 could be suboptimal.

 

[shxxc]

Excuse me for my silly question. After calibrating with MA-1, is the sound of the two 750 subs separated into left and right channels ? I am planning to purchase a pair of KH 120 II monitors and connect them with two KH 750 subs using coaxial cables.

 

[Hexspa]

Not sure if this is an accurate or useful categorization but here is how I see the different flavors of subwoofer integrations:

1. no filtering of the speakers, adjust low pass and volume of the sub to blend
2. high pass filtering of speakers, sub is like an extra way with proper crossover, and EQ is stereo full range, not 3 channels
3. more than one sub, but each receive the same signal and EQ, just with different delays
4. each sub is EQ’d to flat separately
5. each sub has separate EQ but the target isn’t to be individually flat but for the sum to be flat.

For 5. I guess this is because of an asymmetry in how peaks and nulls are considered when creating a filter, so 4 could be suboptimal.

I only have one sub and I get the best response with (1.). But, having looked into the matter, any of these options might be best. (4.) is not necessarily suboptimal because if each sub is flat at a given point, the only thing that would cause non-flatness is a timing issue.

 

[Hexspa]

What’s the consensus here with MA-1? Do you all trust Neumann, this product and their level of service? Watching from the sidelines, it seems less reliable and recommended than GLM with the caveat that it’s a newer system.

While heartening to see an update, I can’t help but feel dismayed this is an additional charge where GLM has the functions included for the roughly equivalent base price.

 

[DJBonoBobo]

What’s the consensus here with MA-1? Do you all trust Neumann, this product and their level of service? Watching from the sidelines, it seems less reliable and recommended than GLM with the caveat that it’s a newer system.

While heartening to see an update, I can’t help but feel dismayed this is an additional charge where GLM has the functions included for the roughly equivalent base price.

Difficult to answer. I trust Neumann to be an excellent hardware manufacturer with good people. But it seems they have underestimated the software side. So, on the one hand it is a good sign they made new software versions just because i told them about bugs - this happened a few times already. On the other hand i think, why do i have to be the one who sees the bugs and why aren't they able to test everything thoroughly themselves?
Also, they tried building a support team (external, guess), but they don't have the same level of expertise as the core guys.
All in all, afaik it is still a small team with limited ressources.
I have always found a solution to work around the quirks, or they fixed it. So while sometimes frustrating, it has been ok for me, because - last point - i trust the end results to be as good as it gets.
I never tried Genelec/GLM.

 

[DJBonoBobo]

Excuse me for my silly question. After calibrating with MA-1, is the sound of the two 750 subs separated into left and right channels ? I am planning to purchase a pair of KH 120 II monitors and connect them with two KH 750 subs using coaxial cables.

I think you want to know how to connect everything? I saw a diagram somewhere, it is probably in the manual. There are different ways, but i think most common is a daisy chain digital out -> sub1 -> kh120ii 1 -> kh120ii 2 -> sub2. Not sure though, i don't have digital speakers.

 

[palm]

I only have one sub and I get the best response with (1.). But, having looked into the matter, any of these options might be best. (4.) is not necessarily suboptimal because if each sub is flat at a given point, the only thing that would cause non-flatness is a timing issue.

If we consider a theoretical system with 2 subs, each one being flat at listening position beside a severe drop around 2 different frequencies each f1 and f2.
Equalization of both subs separately (4.) doesn't help, the EQ system correctly avoiding to fill those nulls. The sum of responses shows two drops.
On the other hand, if the EQ software is also considering the complex sum to calculate correction filters (5.), it might be possible for sub1 to moderately compensate for f2, and sub2 for f1?

 

[Hexspa]

If we consider a theoretical system with 2 subs, each one being flat at listening position beside a severe drop around 2 different frequencies each f1 and f2.
Equalization of both subs separately (4.) doesn't help, the EQ system correctly avoiding to fill those nulls. The sum of responses shows two drops.
On the other hand, if the EQ software is also considering the complex sum to calculate correction filters (5.), it might be possible for sub1 to moderately compensate for f2, and sub2 for f1?

I do not know. If I had two different subs and was getting some kind of phasing issue due to different extensions I might just high pass the more performant sub to match. Given an equal slope, this should fix the problem but at the expense of less bass. An all pass filter might also work. These are just guesses.

 

[McGlear]

(4.) is not necessarily suboptimal because if each sub is flat at a given point, the only thing that would cause non-flatness is a timing issue.

This assumes you can get individual sub curves flat within a non-optimal room, right? Seems quite unlikely even for a single listening position.
The beauty of (5.) - see dlbc, mso, geddes etc. - is that you can fill in one sub's null with another sub at another position in the room, then correct the sum (which unlike the individual sub's response can be flattened, because ideally it does not have any nulls caused by room modes anymore).

 

[Hexspa]

This assumes you can get individual sub curves flat within a non-optimal room, right? Seems quite unlikely even for a single listening position.
The beauty of (5.) - see dlbc, mso, geddes etc. - is that you can fill in one sub's null with another sub at another position in the room, then correct the sum (which unlike the individual sub's response can be flattened, because ideally it does not have any nulls caused by room modes anymore).

It would have been better if I said “flat at a given frequency range at a given point”. Any given frequency may have a null but it may be shallower or further outside your target band than others. For me, a hypothetical narrow 6dB null at 40Hz is less important than a wider 15dB null at 100Hz.

Other than that, this is abstract to me since I’ve only ever used one sub and non-high passed Yamaha HS50s.

 

[Lion♡]

Yes, but only for a KH310 without sub. Here it is with an without FDW, left channel only.
The "pre-ringing" (is this the correct term?) happens only with the MA1.

View attachment 273374
View attachment 273375

Additionally with uncorrected subs:
View attachment 273378


And phase uncorrected vs. corrected:
View attachment 273376

(First of all, I'm not a DLBC user)
I saw some DLBM and DLBC measurements on another Korean community, which raised some questions for me, so I started looking into various things.
And I also reviewed another user’s REW Mdat data and noticed something unusual. I'm curious about the opinions of the OP and other Dirac users on this

First, he measured DLBM and DLBC separately with the subwoofer alone and with the subwoofer + speakers. The start timing of the impulse was correctly aligned.

This is the comparison of DLBM vs. DLBC for Sub 1+2. There isn’t a significant difference visible.
But, when i look Sub + Speaker

Yes, the frequency response (FR) is slightly different. However, I suspect this is due to the impact of timing delay. I intended to extract the excess phase to compare this aspect.

DLBM and DLBC show very different results when the speakers are integrated. To gain a clearer view, I applied a 1/3 octave 100 Hz IR filter to observe only the low-frequency range, with the timing already aligned. My focus was specifically on the delay effects, so I examined the excess phase to exclude frequency response differences.

Hmm... Very Weird.
Green is DLBM , Red is DLBC.
The low-frequency impulse of DLBC definitely appears delayed. This can also be observed in the group delay and spectrogram.

I then time-aligned it and found that it was delayed by approximately 20 ms, with the polarity even reversed.
Why would DLBC suddenly invert the polarity and add a 20 ms delay when integrating with the speakers?
I recall seeing similar issues in DLBC graphs not only in this measurement but also in other communities, and even on ASR. This phenomenon is also visible in the screenshots shared by the thread's OP.

The person who took these measurements also mentioned that, with DLBC enabled, the sound actually seemed to resonate more (not in a good reverb way, but rather as if it were echoing). I’ve been wondering for months why there would be a 20 ms delay and polarity inversion, but I still haven’t found a clear answer, even after reviewing other people’s measurements occasionally.

I’d appreciate input from ASR users on this issue.

////////////////////////

Never mind, I just checked the known issues tab.

 

[Digital_Thor]

I manually tune my 4 subwoofers, so I fully know exactly what filter, slope, gain and delay, that is used for each subwoofer.
When using an automated system - can we then be sure that the system did not use a filter or slope, that maybe "twisted" the phase, so that the response did something - unexpected? Even though that is a thing that should have been thought of?

 

[Lion♡]

I manually tune my 4 subwoofers, so I fully know exactly what filter, slope, gain and delay, that is used for each subwoofer.
When using an automated system - can we then be sure that the system did not use a filter or slope, that maybe "twisted" the phase, so that the response did something - unexpected? Even though that is a thing that should have been thought of?

I don't know if this is a post about me or a question about other users or OPs.
If you're talking to me, maybe you can check.
I remember glimpses of avs threads, but it worked out by leaving subwoofers at opposite polarities (I remember seeing the nuance of Dirac just doing it, not calculating and optimizing the results for bipolar, reverse polarities, but just doing it, I'm not sure).
So it's not polar, and DLBC is trying to compensate for that with delay, and I think these results are combined with inaccurate impulse peak detection compared to REW, but I'm not sure.

 

[Lion♡]

Although it's a post from quite a while ago, I carefully asked the OP, and he kindly shared the measurements. Thank you once again. You had already posted many graphs, but I checked them more neatly through IR timing alignment and looked at a few additional graphs.

(DLBC)

(MA1)

The MA1 seems to have a generally slower late decay attenuation, and the shape of the DLBC and full-range might appear different, but this seems to be due to the influence of the noise floor in the image attached below.

(DLBC noisefloor)

(MA1 noise floor)

Therefore, the overall attenuation pattern across the entire range (especially above the midrange) is likely to be similar, and the difference observed in the measurements may be due to measurement variations.

(DLBC)

(MA1)

I have received and reviewed the mdat files from quite a few DLBC users, but to be honest, I still don't fully understand the intention behind DLBC's phase wrap.

When looking at the Excess Group Delay, the effect of the bug reported on the Dirac side (increase in group delay of the subwoofers) is evident.

this is original step response.(no dlbc, no ma1)

(DLBC)

(ma1)

The 30ms section of the DLBC is unusual. To make this easier to visualize, let's compare the graphs of the minimum phase responses of each DLBC and MA1.

(DLBC compare Minimumphase)

(MA1 compare Minimumphase)

The results will vary depending on the procedures such as pre-ringing cutoff and subsonic filters for each DRC, but the difference between MA1 and DLBC appears to be quite large.

(Original FR)

(Top: DLBC / Bottom: MA1)

As mentioned earlier, it seems that the OP's MA1 measurements, including the noise floor, were not measured under the same conditions as the DLBC. Therefore, I don't think the objective characteristics of each DRC are fully accurately reflected in the comparison, making it somewhat unstable.
However, despite that, the difference between the two DRCs (DLBC and MA1) is still quite significant.

What are your thoughts on this?