• 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!

MSO - Integrating subs in stereo setup (unusual - high level connections)

Bobesku

Member
Joined
Dec 14, 2022
Messages
20
Likes
12
Location
Europe, Romania
Do you know if it's a good ideea or what's the best way to tweak MSO and or actual REW measurements in order to integrate subs with a stereo amp without any means of using a dedicated processor like minDSP? But instead using the DSP included in the Subwoofers?

What I want to achieve is having a better curve in the sub-bass and lower bass region, then let the towers lead the way. So a nice addition/blend of the subs in the 25Hz to 150-200Hz region.

My setup is based on 2 tower speakers (KEF Q750), 2 subs (SVS SB1000 Pro) that are linked using their high level input directly from the Integrated AMP (Arcam SA20) L+R speakers connectors.
Subs are placed in a non-symmetrical way across the room. Subs have some DSP capabilities (gains, phase 0-180, 3 PEQ settings, Room Gain Compensation, Crossover setting etc).

Music source flow: piCoreLms on rpi4 -> Topping E50 -> Arcam SA20 -> Speakers & Subs.


I have already used REW with UMIK-1 and gathered measurements: L+R full range, Sub1 and Sub2 crossed at 60Hz with 24db slope from 3 positions across my couch and already been playing with MSO in the last couple of days.

I am a bit frustrated that almost every time I run an Optimize, MSO is more or less neutering my subs with max cuts on the PEQs. Sometimes just repeats the same Parametric EQ 2-3 times in a row for same sub.

So what I am asking is more like a sanity check of what I want to do with the above.

Any help, remark, dismantling my quest.... etc.... I am looking forward and appreciate as this is my first post here as well.
 

DaveBoswell

Member
Joined
Dec 15, 2022
Messages
29
Likes
25
Location
Toronto, Canada
You should be able to do this and get quite decent results. I did this with my setup for a while until I dusted off my old Behringer Feedback Destroyer and played around with it to very good effect - saved me from going for more expensive new DSP as I wanted to keep the balanced connections to subs.

I have 2 SVS Sealed 13" Ultras that work great - one in front right of room, and one right behind the couch facing the couch.

Here's what worked:
- reversed the polarity of the Back Sub - try this in MSO to see the effect - it was equivalent to adding about 6-8ms of delay, or introducing a second-order all pass filter with low Q and 180 reversal at around 50hz - just doing this step alone made an immediate impact to the seat to seat variation - you could easily do this with your high level inputs by swapping red/black on the speaker wires for one sub
- in MSO, make sure to use the shared gain filter to avoid having the PEQs on the SVS have to double up to meet your target reference level (I use +/- 24db on shared filter and typically target 70db)
- make sure to select ranges for frequency that align with the SVS settings (31.5 to 125 Hz, Q from 2 to 9.6, and Gain range of 0 to -12 for my Ultra 13 subs). When you get the results in MSO, just pick the filter frequency and Q that is closest to the optimized results in MSO - it will not be precise but decent enough - also this means that you should not try to sue high Q filters as the frequency accuracy is not there - keep Q to maybe 8 or lower I would suggest (thanks to Jeff aka fattire for that suggestion)
- I add the filter channels to my measurement outputs offset by the same 70db so I can see them live with the results and to make sure that one sub is not too far out of balance with the other for gain/filter etc

Will see if I can grab some screenshots to show results.
 
Last edited:

DaveBoswell

Member
Joined
Dec 15, 2022
Messages
29
Likes
25
Location
Toronto, Canada
This is the before with raw data measured from 5 positions. Green line is Main Listening Position, Magenta and Red are Left and Rights seats along a single wide couch, and the RMID, LMID are a bit in front of the couch. You can see the very very large seat to seat variation - ouch!
1671073626204.png




This is the result of only adding a single Polarity Inversion to the Back Center sub - what a huge improvement in seat to seat variation from 15 to 65Hz! If any of you have subs in the front and back the room, give this a shot first - MSO will not automatically try flagging this on and off in the optimization routine so it's a manual step but and easy one.
1671073895377.png



And here is what you can get with 2 subs, one polarity inversion and only 2 PEQ filters per sub. Not too shabby at all! The remaining peaks at 43Hz and 92 Hz can be tamed easily with Audyssy or in my case ARC as I have an Anthem 50v. BTW I love the trick of adding the filter channel traces to the results chart (you can see them here light blue and black) - a great way to dynamically see what is happening as you are optimizing. The goal should always be to have the filter lines be as close together as possible such that each sub is sharing the load as equally as possible giving you as much headroom as possible.
1671074071001.png



The results from this approach have been startling improvements to say the absolute least, and the rich depth and tightness of the bass in the room simply blows me away and is something I had thought not possible inside a smallish room. Thanks to Andy for this amazing tool for us to use to so rapidly experiment with different approaches.

Dave
 
Last edited:

DaveBoswell

Member
Joined
Dec 15, 2022
Messages
29
Likes
25
Location
Toronto, Canada
And if you haven't tried plotting the filter channels, here's how I do it - right click on the graph and click on Graph Properties and then Filter Channels then turn on the filter channels for each sub for the relevant Configuration, then go back to the chart and right click again and click on the Trace Properties, and adjust the Magnitude Display Offset to be the same db as you are targeting for (so in my case I target to 70db, so I set the offset to 70db). I also add the target Trace to the chart and this way you can easily see what is happening in the filters and can directly compare to the target curve.

1671075538053.png


1671075565875.png
 
Last edited:
OP
Bobesku

Bobesku

Member
Joined
Dec 14, 2022
Messages
20
Likes
12
Location
Europe, Romania
Thank you for your input. Awesome stuff. I was about to give MSO a no-go and to focus more on REW Alignment tool results which is what my subs are using now.
I will fiddle and implement your tips today and get back with my findings.
 

kiwifi

Active Member
Joined
Apr 21, 2021
Messages
235
Likes
187
Do you know if it's a good ideea or what's the best way to tweak MSO and or actual REW measurements in order to integrate subs with a stereo amp without any means of using a dedicated processor like minDSP? But instead using the DSP included in the Subwoofers?
One of the most confusing things if you're not using MiniDSP, is the knowing which definition of Q to use when programming an external DSP. The best way to confirm that the DSP is programed correctly is to measure the combined sub output after the PEQ filters have been programed. The response should be very close to what MSO predicted. If not, it is possible that you need to use RBJ (Q) filters in MSO but program your DSP using the classic Q values.
What I want to achieve is having a better curve in the sub-bass and lower bass region, then let the towers lead the way. So a nice addition/blend of the subs in the 25Hz to 150-200Hz region.
Is there a way for you to high pass the signal to the towers to match the low pass on the subs? You don't want bass from the towers competing with the subs.
I am a bit frustrated that almost every time I run an Optimize, MSO is more or less neutering my subs with max cuts on the PEQs. Sometimes just repeats the same Parametric EQ 2-3 times in a row for same sub.
You can limt the stacking of PEQ filters using the constraints: total PEQ boost & cut limits.
 
OP
Bobesku

Bobesku

Member
Joined
Dec 14, 2022
Messages
20
Likes
12
Location
Europe, Romania
One of the most confusing things if you're not using MiniDSP, is the knowing which definition of Q to use when programming an external DSP.
Yes, that confusion still lingers in my mind, what Q does the SVS DSP use? I will start tweaking and once I'm satisfied with the results will measure them with REW to get a complete feedback on what has actually changed.

Is there a way for you to high pass the signal to the towers to match the low pass on the subs? You don't want bass from the towers competing with the subs.
No way to high-pass the towers, my amp does not have such capabilities. And that's one of the biggest items to overcome, actually was hoping to smooth the towers bass with the subs in this process as well. Been juggling with MSO with either Subs-Only or Mains+Subs templates in order to have an clue of how the overall bass output would come up.

You can limt the stacking of PEQ filters using the constraints: total PEQ boost & cut limits.
I have also played with the limits, but then I was thinking that maybe by putting up constraints I would actually skew the MSO capabilities. So I only used the max limits of my subs gains listed in the built-in DSP they have.
 

DaveBoswell

Member
Joined
Dec 15, 2022
Messages
29
Likes
25
Location
Toronto, Canada
No way to high-pass the towers, my amp does not have such capabilities. And that's one of the biggest items to overcome, actually was hoping to smooth the towers bass with the subs in this process as well. Been juggling with MSO with either Subs-Only or Mains+Subs templates in order to have an clue of how the overall bass output would come up.

This is a tricky one - if you cannot bypass the towers, then suggest you measure them as separate "sub woofers" to be included in your overall calc in MSO just like you do with your subs one by one for each seat, and simply don't include any filters in MSO on these tower speaker "subs" in your optimization runs.

So your MSO optimization runs would have 4 subs: SVS 1, SVS 2, Tower L and Tower R and the towers would have no filters assigned in MSO, and the SVS subs would have 3 PEQ filters each. I think this is a reasonable approach even though there may be some variance in bass signals between the left speaker and the right, but this is unlikely IMO to be that relevant for frequencies below 100Hz which is where the smoothing of the combined bass will have the biggest impact.

I used the non-RBJ filters in MSO and these align well to SVS built in filters. To be honest not quite sure what the differences between them but the output measured in REW aligned well with predicted in MSO using this approach.
 
OP
Bobesku

Bobesku

Member
Joined
Dec 14, 2022
Messages
20
Likes
12
Location
Europe, Romania
Thank you! You have made my day with this brilliant idea. Also, I was about to ask SVS team about the Q type.
Will get back with results at some point soon. Thank you once again!
 

DaveBoswell

Member
Joined
Dec 15, 2022
Messages
29
Likes
25
Location
Toronto, Canada
Hey no worries. I'm actually quite curious to see how this approach works as I don't remember seeing this in any other threads as just about every one assumes a cross over between subs and satellites. Other than dealing with level matching and potentially wasting some amp power on the full range bass to towers, not really sure why this should not work. Good luck!
 
OP
Bobesku

Bobesku

Member
Joined
Dec 14, 2022
Messages
20
Likes
12
Location
Europe, Romania
Well, it's starting to look nice.

This is the starting base, see below

Baseline.png



After some tweaks, using Audiofrog House Curve

Latest_opt.png




And my latest run, with a flat target (no house curve)

Latest_opt_conf10.png



Now, all I have to do is implement them into the Subs DSP, have a good listen and REW measurements to remove human bias, placebo effects etc
But it's late here, I have to call it a day and get back tomorrow.

Thank you for all your help, and I'll keep you posted.



 

DaveBoswell

Member
Joined
Dec 15, 2022
Messages
29
Likes
25
Location
Toronto, Canada
Looks promising! Red seat is +/- around 3db from 24 to 140Hz except for small dip at 44 Hz which is really good.

Suggest you remove the Filter Measurements and replace with Filter Channels (one for each of the SVS subs) so you can see the filters being applied.
 
Last edited:
OP
Bobesku

Bobesku

Member
Joined
Dec 14, 2022
Messages
20
Likes
12
Location
Europe, Romania
Couldn't wait, so I've put the settings in the Subs. Along with that, I have added LPF to the subs and MSO tuned them one at 83Hz and the other at 150Hz.
Now enjoying a movie session at low volume, around 30-35db in-room noise. I can hear some new sounds in the lower register. But that's all I can do as it's night time.

The chart used

Screenshot_20221216_034108.png


I will ramp up the volume tomorrow, pull out REW and all the tools in order to get to a conclusion.


See below the MSO reported filters and parameters:


Filters by channel:

Individual sub channels:

Channel: "Left Main"
FL1: Parametric EQ
Parameter "Center freq (Hz)" = 80
Parameter "Boost (dB)" = 0
Parameter "Q" = 2
Q (RBJ) = 2

FL2: Parametric EQ
Parameter "Center freq (Hz)" = 80
Parameter "Boost (dB)" = 0
Parameter "Q" = 2
Q (RBJ) = 2

FL3: Parametric EQ
Parameter "Center freq (Hz)" = 80
Parameter "Boost (dB)" = 0
Parameter "Q" = 2
Q (RBJ) = 2

FL5: Delay Block
Parameter "Delay (msec)" = 0
This delay value is an intermediate calculation for reference only.
For final delay values, see "Final gain and delay/distance settings" at end of report.

End Channel: "Left Main"

Channel: "Right Main"
FL6: Parametric EQ
Parameter "Center freq (Hz)" = 80
Parameter "Boost (dB)" = 0
Parameter "Q" = 2
Q (RBJ) = 2

FL7: Parametric EQ
Parameter "Center freq (Hz)" = 80
Parameter "Boost (dB)" = 0
Parameter "Q" = 2
Q (RBJ) = 2

FL8: Parametric EQ
Parameter "Center freq (Hz)" = 80
Parameter "Boost (dB)" = 0
Parameter "Q" = 2
Q (RBJ) = 2

FL9: Gain Block
Parameter "Gain (dB)" = 0
This gain value is an intermediate calculation for reference only.
For final gain values, see "Final gain and delay/distance settings" at end of report.

FL10: Delay Block
Parameter "Delay (msec)" = 0
This delay value is an intermediate calculation for reference only.
For final delay values, see "Final gain and delay/distance settings" at end of report.

End Channel: "Right Main"

Channel: "Sub 1"
FL11: Parametric EQ
Parameter "Center freq (Hz)" = 20.2437
Parameter "Boost (dB)" = 6
Parameter "Q" = 4.11743
Q (RBJ) = 2.91491

FL12: Parametric EQ
Parameter "Center freq (Hz)" = 46.1053
Parameter "Boost (dB)" = -9.41534
Parameter "Q" = 10
Q (RBJ) = 5.81589

FL13: Parametric EQ
Parameter "Center freq (Hz)" = 22.8428
Parameter "Boost (dB)" = 6
Parameter "Q" = 0.915799
Q (RBJ) = 0.648336

FL14: Gain Block
Parameter "Gain (dB)" = 6
This gain value is an intermediate calculation for reference only.
For final gain values, see "Final gain and delay/distance settings" at end of report.

FL15: Delay Block
Parameter "Delay (msec)" = 2.29256
This delay value is an intermediate calculation for reference only.
For final delay values, see "Final gain and delay/distance settings" at end of report.

FL20: LPF Butterworth 24 dB/oct
Parameter "Cutoff freq (Hz)" = 83.3889

End Channel: "Sub 1"

Channel: "Sub 2"
FL16: Parametric EQ
Parameter "Center freq (Hz)" = 97.6483
Parameter "Boost (dB)" = -10.6279
Parameter "Q" = 10
Q (RBJ) = 5.42379

FL17: Parametric EQ
Parameter "Center freq (Hz)" = 40.8959
Parameter "Boost (dB)" = -12
Parameter "Q" = 10
Q (RBJ) = 5.01187

FL18: Parametric EQ
Parameter "Center freq (Hz)" = 60.2858
Parameter "Boost (dB)" = 6
Parameter "Q" = 10
Q (RBJ) = 7.07946

FL19: Gain Block
Parameter "Gain (dB)" = -5.83172
This gain value is an intermediate calculation for reference only.
For final gain values, see "Final gain and delay/distance settings" at end of report.

FL21: LPF Butterworth 24 dB/oct
Parameter "Cutoff freq (Hz)" = 150

FL22: Polarity Inversion

End Channel: "Sub 2"

Individual mains channels:

No mains channels defined

Shared sub channel:

FL4: Gain Block
Parameter "Gain (dB)" = 6.29167
This gain value is an intermediate calculation for reference only.
For final gain values, see "Final gain and delay/distance settings" at end of report.

End shared sub channel

Shared mains channel:

No filters in this channel.


Raw uncorrected gain and delay values:
These gain and delay values are for reference only.
Unadjusted gain values of all gain blocks:
FL4 (shared, gain block) gain: 6.29 dB
FL9 (Right Main, gain block) gain: 0.00 dB
FL14 (Sub 1, gain block) gain: 6.00 dB
FL19 (Sub 2, gain block) gain: -5.83 dB
Unadjusted delay values of all delay blocks:
FL5 (Left Main, delay block) delay: 0.00 msec
FL10 (Right Main, delay block) delay: 0.00 msec
FL15 (Sub 1, delay block) delay: 2.29 msec

For final gain and delay values, see
"Final gain and delay/distance settings" at end of report.

Final gain and delay/distance settings:

Minimal gain settings:
Left Main gain: -6.00 dB
Right Main gain: -6.00 dB
Sub 1 gain: 0.00 dB
Sub 2 gain: -11.83 dB
Delay settings:
Left Main delay: 0.00 msec
Right Main delay: 0.00 msec
Sub 1 delay: 2.29 msec
Sub 2 delay: 0.00 msec

Channel inversions:
Sub 2: Invert
 

Attachments

  • Screenshot_20221216_032702.png
    Screenshot_20221216_032702.png
    90.3 KB · Views: 46
Last edited:

DaveBoswell

Member
Joined
Dec 15, 2022
Messages
29
Likes
25
Location
Toronto, Canada
LOL - been there late nights many times. We can all be a tad obsessive about trying new optimizations once we have a new idea!

One observation from your filter charts is that the subs are quite far apart in gain. Assuming your measurements had your subs at the same gain setting when you measured, your Sub 1 is taking the majority of the load, and sub 2 is running at 10 to 20 db lower in vol. You may want to limit the gain settings of the subs in the optimization to force them to be closer in gain to each other - to do this set the gain max and min for each sub to be in a smaller range.

You are trying to avoid the louder sub from reaching its upper limit and compressing while the other sub is only using a small percent of its capacity.
 
OP
Bobesku

Bobesku

Member
Joined
Dec 14, 2022
Messages
20
Likes
12
Location
Europe, Romania
OK, so I have confirmation from SVS Support (Ed Mullen representing SVS) that my sealed subs SB1000 Pro are using LPF of type Butterworth and the filter Q type used is Classic (not RBJ).

Going forward, and after using a couple of MSO optimize runs with various results and crossovers, I have listened and the sound was not that great, although the curve was looking nice. It's like all the lost frequencies were there, but the subs would take over and mask the signature of my speakers... Some kind of overlap. Maybe some people would like that, I didn't. The sound was a bit too full-bodied for my taste, and not that I like brighter sounding, but still... to heavy and full on the base line. For movies could be an option. This was with the subs measurements without LPF and playing with proposed MSO LPFs on the subs.

So, instead of modelling the LPF within the MSO, I thought I would better take advantage of the already made in-room measurements which included the subs in all 3 positions using several crossovers so that I could also let the main speakers keep their mark on the sound. The latest good result I could get in these conditions was using an already predefined measured LPF (in my case 60Hz) for both subs, reversed polarity on the closest one, and other various PEQ and delay plotted by MSO. So I am very happy that I spent more than 3 hours to measure all kind of variations. And I think that this is dependent on the room, so others may need to fiddle with several settings and measurement types until finding the best result possible.

See below the latest graph

Screenshot_20221217_030754.png


Filters by channel:

Individual sub channels:

Channel: "Left"
FL1: Parametric EQ
Parameter "Center freq (Hz)" = 80
Parameter "Boost (dB)" = 0
Parameter "Q" = 2
Q (RBJ) = 2

FL2: Parametric EQ
Parameter "Center freq (Hz)" = 80
Parameter "Boost (dB)" = 0
Parameter "Q" = 2
Q (RBJ) = 2

FL3: Parametric EQ
Parameter "Center freq (Hz)" = 80
Parameter "Boost (dB)" = 0
Parameter "Q" = 2
Q (RBJ) = 2

FL5: Delay Block
Parameter "Delay (msec)" = 0
This delay value is an intermediate calculation for reference only.
For final delay values, see "Final gain and delay/distance settings" at end of report.

End Channel: "Left"

Channel: "Right"
FL6: Parametric EQ
Parameter "Center freq (Hz)" = 80
Parameter "Boost (dB)" = 0
Parameter "Q" = 2
Q (RBJ) = 2

FL7: Parametric EQ
Parameter "Center freq (Hz)" = 80
Parameter "Boost (dB)" = 0
Parameter "Q" = 2
Q (RBJ) = 2

FL8: Parametric EQ
Parameter "Center freq (Hz)" = 80
Parameter "Boost (dB)" = 0
Parameter "Q" = 2
Q (RBJ) = 2

FL9: Gain Block
Parameter "Gain (dB)" = 0
This gain value is an intermediate calculation for reference only.
For final gain values, see "Final gain and delay/distance settings" at end of report.

FL10: Delay Block
Parameter "Delay (msec)" = 0
This delay value is an intermediate calculation for reference only.
For final delay values, see "Final gain and delay/distance settings" at end of report.

End Channel: "Right"

Channel: "Sub 1"
FL11: Parametric EQ
Parameter "Center freq (Hz)" = 45.8907
Parameter "Boost (dB)" = -11.6831
Parameter "Q" = 10
Q (RBJ) = 5.10414

FL12: Parametric EQ
Parameter "Center freq (Hz)" = 122.887
Parameter "Boost (dB)" = 6
Parameter "Q" = 3.1855
Q (RBJ) = 2.25516

FL13: Parametric EQ
Parameter "Center freq (Hz)" = 30.3086
Parameter "Boost (dB)" = 6
Parameter "Q" = 1.63032
Q (RBJ) = 1.15418

FL14: Gain Block
Parameter "Gain (dB)" = 6
This gain value is an intermediate calculation for reference only.
For final gain values, see "Final gain and delay/distance settings" at end of report.

FL15: Delay Block
Parameter "Delay (msec)" = 8.33
This delay value is an intermediate calculation for reference only.
For final delay values, see "Final gain and delay/distance settings" at end of report.

End Channel: "Sub 1"

Channel: "Sub 2"
FL16: Parametric EQ
Parameter "Center freq (Hz)" = 83.6154
Parameter "Boost (dB)" = 5.45117
Parameter "Q" = 10
Q (RBJ) = 7.30669

FL17: Parametric EQ
Parameter "Center freq (Hz)" = 89.9879
Parameter "Boost (dB)" = 6
Parameter "Q" = 2.14582
Q (RBJ) = 1.51912

FL18: Parametric EQ
Parameter "Center freq (Hz)" = 38.9669
Parameter "Boost (dB)" = -12
Parameter "Q" = 10
Q (RBJ) = 5.01187

FL19: Gain Block
Parameter "Gain (dB)" = 5.08579
This gain value is an intermediate calculation for reference only.
For final gain values, see "Final gain and delay/distance settings" at end of report.

FL20: Polarity Inversion

End Channel: "Sub 2"

Individual mains channels:

No mains channels defined

Shared sub channel:

FL4: Gain Block
Parameter "Gain (dB)" = 6.73786
This gain value is an intermediate calculation for reference only.
For final gain values, see "Final gain and delay/distance settings" at end of report.

End shared sub channel

Shared mains channel:

No filters in this channel.


Raw uncorrected gain and delay values:
These gain and delay values are for reference only.
Unadjusted gain values of all gain blocks:
FL4 (shared, gain block) gain: 6.74 dB
FL9 (Right, gain block) gain: 0.00 dB
FL14 (Sub 1, gain block) gain: 6.00 dB
FL19 (Sub 2, gain block) gain: 5.09 dB
Unadjusted delay values of all delay blocks:
FL5 (Left, delay block) delay: 0.00 msec
FL10 (Right, delay block) delay: 0.00 msec
FL15 (Sub 1, delay block) delay: 8.33 msec

For final gain and delay values, see
"Final gain and delay/distance settings" at end of report.

Final gain and delay/distance settings:

Minimal gain settings:
Left gain: -6.00 dB
Right gain: -6.00 dB
Sub 1 gain: 0.00 dB
Sub 2 gain: -0.91 dB
Delay settings:
Left delay: 0.00 msec
Right delay: 0.00 msec
Sub 1 delay: 8.33 msec
Sub 2 delay: 0.00 msec

Channel inversions:
Sub 2: Invert


Is it perfect? Of course, not! I think that the limitations of this approach - using just built-in subwoofers DSP shows in the above! But I have made a compromise and been able to keep the well-defined bass notes and frequencies up to 100 Hz but without coloring the sound of the towers.


And, the sound is much better than without these tweaks, see below the baseline chart - starting point.

Screenshot_20221217_032037.png



So I am more than happy with these results at this point and thank you for your help.

Especially @DaveBoswell for the ideea of using the towers as subs in the MSO initial wizard configuration! This should be put in a Manual for using this combo - subwoofers linked to stereo full range speakers using high level connections, given that the subs do have some decent DSP capabilites.


LE: I also found a better looking graph and MSO filter setting while fiddling with it. But that required my Subwoofer DSP to use a delay of 15.93 msec. And that translated in phase accounting for a 60Hz crossover was higher than 180 degrees, which is not possible in the SVS subwoofers DSP.

LE2: If you think I can still improve these, or it's something that does not look that good in the MSO results, please let me know. I am always seeking a way to improve, in all aspects.
 
Last edited:

kiwifi

Active Member
Joined
Apr 21, 2021
Messages
235
Likes
187
OK, so I have confirmation from SVS Support (Ed Mullen representing SVS) that my sealed subs SB1000 Pro are using LPF of type Butterworth and the filter Q type used is Classic (not RBJ).
So my understanding is that if you use RBJ filters in your MSO config, then the filter report will contain both RBJ and classic Q values for each filter, but you will be able to take advantage of a wider range of Classic Q values because RBJ Q values are always smaller than Classic Q.

LE: I also found a better looking graph and MSO filter setting while fiddling with it. But that required my Subwoofer DSP to use a delay of 15.93 msec. And that translated in phase accounting for a 60Hz crossover was higher than 180 degrees, which is not possible in the SVS subwoofers DSP.

LE2: If you think I can still improve these, or it's something that does not look that good in the MSO results, please let me know. I am always seeking a way to improve, in all aspects.
If you see a large delay assigned to a sub channel by MSO, try adding a polarity inversion to that sub's config and running the optimization again.
 
OP
Bobesku

Bobesku

Member
Joined
Dec 14, 2022
Messages
20
Likes
12
Location
Europe, Romania
So my understanding is that if you use RBJ filters in your MSO config, then the filter report will contain both RBJ and classic Q values for each filter, but you will be able to take advantage of a wider range of Classic Q values because RBJ Q values are always smaller than Classic Q.
I don't think it's an issue related to Classic Q, but a maximum wider range for both types - especially for RBJ types.

On MSO - Filter Compatibility With Available Hardware and Software we get this:
Beginning in version 1.23 of MSO, you can specify "Parametric EQ (RBJ)" filters, which use the Q(RBJ) directly in their internal implementation. If you constrain the Q of these filters to a maximum of 10 (the maximum for the iNuke), that will be equivalent to a maximum Q(RBJ) of 10 as well, thus allowing the full range of Q supported in the iNuke to be used. Therefore, if your hardware requires manual PEQ data entry and uses the Q(RBJ) convention as the iNuke does, it's strongly recommended that you use the new "Parametric EQ (RBJ)" filters rather than the standard PEQ filters to allow you to make use of the full range of Q values supported by your hardware.

Where the above iNuke is a RBJ implemented. So my understanding is this: using RBJ filters in the MSO config ensures that both types would run at their full capacity, where the issue has been on RBJ types. So I think it's best to just select RBJ in order to get full compatibility on both types.

For me this was not an issue, as I have a Classic Q on the internal DSP from the built-in subs.


If you see a large delay assigned to a sub channel by MSO, try adding a polarity inversion to that sub's config and running the optimization again.
Of course, that's what I thought and adjusted but I was already having a polarity inversion on the other one and it didn't make things better. Switched those polarity inversions in all combos. So it's not always a guaranteed success. Best way is to test all possible variations as it's room / layout dependent.
 
Top Bottom