Help needed - Subwoofer loundness good but not tight

[JStewart]

I’ve been preaching this for decades now, but that wouldn’t give any advantage that couldn’t be accomplished by simply boosting the low end to get the 20 Hz. Either way, the sub drivers and amplifiers will need enough headroom to accomplish it.

Myths and Legends Concerning Positive-Gain Filters when Equalizing Subwoofers

Myths and Legends Regarding Positive-Gain Filters when Equalizing Subwoofers Disclaimer: The following relates to the use of outboard, stand-alone equalizers such as the miniDSP and other equalizers and digital speaker processors from the pro audio field. Reduced Dynamic Range and...

www.avnirvana.com

Regards,
Wayne

Isn’t there a potential for clipping in the digital domain with boosting filters, though? Even more so with a subwoofer where digital signals are combined and sent to it with bass management?
I’m not an expert in this regard, but I have observed that many digital equalizers reserve headroom for their boosting filters. Dirac for example reserves 10dB of digital headroom and will apply no boost filter above 10dB.
The problem is we typically have no way to know when digital clipping is occurring except perhaps some ugly noises, so safest is probably not using or minimizing boost filters. Use cut filters and turn up the sub amp gain. The headroom loss would be the same... I think?

 

[Todd Anderson]

You can definitely clip in the digital domain… depending on how aggressive the boost is

 

[Todd Anderson]

Isn’t there a potential for clipping in the digital domain with boosting filters, though? Even more so with a subwoofer where digital signals are combined and sent to it with bass management?
I’m not an expert in this regard, but I have observed that many digital equalizers reserve headroom for their boosting filters. Dirac for example reserves 10dB of digital headroom and will apply no boost filter above 10dB.
The problem is we typically have no way to know when digital clipping is occurring except perhaps some ugly noises, so safest is probably not using or minimizing boost filters. Use cut filters and turn up the sub amp gain. The headroom loss would be the same... I think?

I think you’re right….

This calls for an experiment!

 

[Wayne A. Pflughaupt]

Isn’t there a potential for clipping in the digital domain with boosting filters, though? Even more so with a subwoofer where digital signals are combined and sent to it with bass management?
I’m not an expert in this regard, but I have observed that many digital equalizers reserve headroom for their boosting filters. Dirac for example reserves 10dB of digital headroom and will apply no boost filter above 10dB.
The problem is we typically have no way to know when digital clipping is occurring except perhaps some ugly noises, so safest is probably not using or minimizing boost filters. Use cut filters and turn up the sub amp gain. The headroom loss would be the same... I think?

Yes, the headroom loss is the same either way. The way to avoid internal clipping in the equalizer is to reduce the input signal appropriately. You can make it up on the back end (sub amp gain setting).

Regards,
Wayne

 

[EGR64]

Tweak City Audio CS18.2 ... are these from Craig Chase? I would say they are quite capable.

Why are you using 200Hz as your crossover? Your RF7 fronts should easily play down to 70-80Hz... I would set the crossover at 80Hz.

You should also be able to run a sweep on the left front speaker, and as long as your Pioneer AVR is set to Stereo, it will capture your sub and front left speaker together, which will show the frequency integration between the subs and front left speaker in one sweep.

Agreed

 

[ddude003]

Boosting digital or analogue EQ more than a few dB (+3 to +5) can cause issues which may include clipping, phase and/or cause the sound to become "muddy" or "smeared" as some would say... Consider frequency and time domain... There are a few exceptions to be had with some professional studio quality gear and/or software DSP...

You might read this post by @Mitchco https://www.audiosciencereview.com/...what-is-your-favorite-house-curve.2382/page-2

Sorry, I did not see this until now...
"@ddude003 How can I find out about this time alignment issue? Is there something I can do in REW to check this out?"

This is one of the best tutorials I am aware of...

And for those that want a deep dive down the rabbit hole there is my all-time favorite by Mitch Barnett.

 

[Tim Hector]

I just breezed through this post, I I missed someone already saying this, I apologize. Add the second sub. Multiple subs very much help smooth out those peaks and valleys, then back them off a bit. Do not have them near each other. That should help. Depending on room nodes, you will need to set crossover anywhere from 60 to 80, but adding the second sub will definitely help with the room issue. Good luck.

 

[Sdk]

even the freq response already looks good (inside 6dB band flatness, cmiiw), pull the main speaker a bit away from the wall will help to get better sound, and can increase the bass detail, and check it with stereo music you are familiar with (that contains quite full audio spectrum, bass mid, high)

 

[achugh]

You can definitely clip in the digital domain… depending on how aggressive the boost is

Yes, the headroom loss is the same either way. The way to avoid internal clipping in the equalizer is to reduce the input signal appropriately. You can make it up on the back end (sub amp gain setting).

Regards,
Wayne

Agreed

Boosting digital or analogue EQ more than a few dB (+3 to +5) can cause issues which may include clipping, phase and/or cause the sound to become "muddy" or "smeared" as some would say... Consider frequency and time domain... There are a few exceptions to be had with some professional studio quality gear and/or software DSP...

@Todd Anderson / @Wayne A. Pflughaupt / @EGR64 / @ddude003 Sorry about this delayed response. It took me a while to do a full manual calibration again. I used OCA Manual Calbiration Video (see below) for the most part plus some of his other videos as needed. I also used @HD MOVIE SOURCE video to use REW RTA with S&M Disc to set the SPL Levels of all my speakers. He talks about it in the following video around 35 min mark.

For this new calibration, my AVR LFE is set at -4.5db to reduce the clipping which goes though AntiMode 8033 that does mostly nothing which then enters miniDSP 2x2 HD where I set the INPUT GAIN to -20db.

I believe all this reduction should take care of this clipping situation as there is no easy way to see this digital clipping like everyone has already mentioned here. I am also not trying to listen at reference levels so I hope that is also helping me avoid this clipping from happening in my case.

I just breezed through this post, I I missed someone already saying this, I apologize. Add the second sub. Multiple subs very much help smooth out those peaks and valleys, then back them off a bit. Do not have them near each other. That should help. Depending on room nodes, you will need to set crossover anywhere from 60 to 80, but adding the second sub will definitely help with the room issue. Good luck.

@Tim Hector I have 2 subs at the corner of the room. In the picture below they are on my left and right side of the sofa and are getting the corner boost as well.

View attachment 69088

even the freq response already looks good (inside 6dB band flatness, cmiiw), pull the main speaker a bit away from the wall will help to get better sound, and can increase the bass detail, and check it with stereo music you are familiar with (that contains quite full audio spectrum, bass mid, high)

@Sdk I will try moving these speaker in my next attempt of calibration.

@ManCave11 I did the complete remeasurements and this time I used 1 db flatness curve in REW instead of 3db for miniDSP to get even a flatter response. I also used the LowShelf filter to apply the 8db boost on the Input side so that I can keep AntiMode 8033 out of picture as is suggested here. It is still connected in the pipeline but I am not using it's house curve which is a 7db boost at 35 Hz, I believe.

My LFE measurements now look like below for all the crossovers (80, 100, 150, 200) that my AVR allows. It allows some other values below 80 but I did not test/try those.

And here's the LEFT speaker sweeps for all these crossovers with speaker set to SMALL. I kept LFE 80 Hz crossover ONLY in this picture for reference.

Here we can see for some weird reason my LEFT channel is not hitting the same levels as my LFE. I believe this is because the 10db LFE boost that happens in decoding the LFE signal is no longer happening hence this lines are exactly 10 db below the LFE.

With all the manual calibration I did using the OCA method, my surround panning and my center speaker dialogue clarity has improved but it has not made much of a difference in the chest bump.

The other struggling part is that my MUSIC listening (or playing games on Nintendo Switch) like everyone else is without the 10 db LFE boost so it is definitely sounding worse than watching a movie. How did you all solve for it? Did you add 10 db in the miniDSP on a different profile? As I mentioned in this post, I cracked up my sub amps and have now set the miniDSP input gain to -20 db because that is the only thing that comes to mind for solving the non-movie mode setup.

 

[HD MOVIE SOURCE]

@Todd Anderson / @Wayne A. Pflughaupt / @EGR64 / @ddude003 Sorry about this delayed response. It took me a while to do a full manual calibration again. I used OCA Manual Calbiration Video (see below) for the most part plus some of his other videos as needed. I also used @HD MOVIE SOURCE video to use REW RTA with S&M Disc to set the SPL Levels of all my speakers. He talks about it in the following video around 35 min mark.

For this new calibration, my AVR LFE is set at -4.5db to reduce the clipping which goes though AntiMode 8033 that does mostly nothing which then enters miniDSP 2x2 HD where I set the INPUT GAIN to -20db.

I believe all this reduction should take care of this clipping situation as there is no easy way to see this digital clipping like everyone has already mentioned here. I am also not trying to listen at reference levels so I hope that is also helping me avoid this clipping from happening in my case.

@Tim Hector I have 2 subs at the corner of the room. In the picture below they are on my left and right side of the sofa and are getting the corner boost as well.


@Sdk I will try moving these speaker in my next attempt of calibration.

@ManCave11 I did the complete remeasurements and this time I used 1 db flatness curve in REW instead of 3db for miniDSP to get even a flatter response. I also used the LowShelf filter to apply the 8db boost on the Input side so that I can keep AntiMode 8033 out of picture as is suggested here. It is still connected in the pipeline but I am not using it's house curve which is a 7db boost at 35 Hz, I believe.

My LFE measurements now look like below for all the crossovers (80, 100, 150, 200) that my AVR allows. It allows some other values below 80 but I did not test/try those.

View attachment 69246

And here's the LEFT speaker sweeps for all these crossovers with speaker set to SMALL. I kept LFE 80 Hz crossover ONLY in this picture for reference.
View attachment 69247

Here we can see for some weird reason my LEFT channel is not hitting the same levels as my LFE. I believe this is because the 10db LFE boost that happens in decoding the LFE signal is no longer happening hence this lines are exactly 10 db below the LFE.

With all the manual calibration I did using the OCA method, my surround panning and my center speaker dialogue clarity has improved but it has not made much of a difference in the chest bump.

The other struggling part is that my MUSIC listening (or playing games on Nintendo Switch) like everyone else is without the 10 db LFE boost so it is definitely sounding worse than watching a movie. How did you all solve for it? Did you add 10 db in the miniDSP on a different profile? As I mentioned in this post, I cracked up my sub amps and have now set the miniDSP input gain to -20 db because that is the only thing that comes to mind for solving the non-movie mode setup.

I counter the level issue, I went into the 2 channel settings, set it to manual, and on my Denon. Lowered my left and right channel by -10db. This only applied to 2 channel music, and fixes the issue.
The issue is caused by using a specific target curve for your LFE channel, and a different target for the main speakers. One change you can also make is lower the subwoofer target curve in Mini-DSP by exactly 50%. So my sub room curve is 7.4db. So, use a 3.7db house curve. Start the house curve at 80hz, never boost above that. I use a Q of 0.9. This allows for smoother integration with levels. But to be fair I still like running my subs hot still for 2 channel music. Then remeasure the channels at 75db and let me know what you think.

Unless you are using the exact same target curve for your sub and mains, the levels will always be slightly different. I found that lowering my room curve by 50% seems like a nice workaround and it still has killer bass.

 

[Tim Hector]

I believe that you should at least try moving your woofers out of the corners. You are looking at a 10db boost from 60hz & down. I suspect moving the woofers out of the corners will bring that down at least 6db. You should end up with a much smoother more balanced transition.

 

[JStewart]

The other struggling part is that my MUSIC listening (or playing games on Nintendo Switch) like everyone else is without the 10 db LFE boost so it is definitely sounding worse than watching a movie. How did you all solve for it?

I’m not sure that you are looking at the LFE channel and the “10dB boost” correctly.
On a movie track the LFE channel is 10dB lower than all the other channels, so when the AVR applies a 10dB “boost” it is boosted back to the same volume of all the other channels.
Or perhaps I’m just totally misunderstanding what you’re getting at?

With respect to chest bump, this article may hold some clues.

Bass: the Physical Sensation of Sound

Ever wonder why you can feel bass? We take a look behind the science of bass waves and the tactile sensations that they cause. Our bass experiments were quite revealing but don't try them at home.

www.audioholics.com

 

[achugh]

I counter the level issue, I went into the 2 channel settings, set it to manual, and on my Denon. Lowered my left and right channel by -10db. This only applied to 2 channel music, and fixes the issue.
The issue is caused by using a specific target curve for your LFE channel, and a different target for the main speakers. One change you can also make is lower the subwoofer target curve in Mini-DSP by exactly 50%. So my sub room curve is 7.4db. So, use a 3.7db house curve. Start the house curve at 80hz, never boost above that. I use a Q of 0.9. This allows for smoother integration with levels. But to be fair I still like running my subs hot still for 2 channel music. Then remeasure the channels at 75db and let me know what you think.

Unless you are using the exact same target curve for your sub and mains, the levels will always be slightly different. I found that lowering my room curve by 50% seems like a nice workaround and it still has killer bass.

@HD MOVIE SOURCE Sorry my really old legacy AVR (Pioneer Elite SC-37) does not have any such configuration. I have channel level controls and the closest thing would be for me to create a new non-Movie profile (I have 6 profiles I can use with some limitations around certain things shared across profiles) and try what you are describing. I have crated the 8db house curve in miniDSP as INPUT gain using Q 0.9 at 50Hz which give 8db exactly at 30Hz they way you talked about in your designing a subwoofer house curve video. Since I am on the newest miniDSP firmware and control app, using 63hz with Q 0.7 like you talked about did not workout as it was hitting 8 db before 30Hz. If I take out this house curve (like disable these filters) my miniDSP is set to flat 75db response to match the rest of the speakers.

FYI, I also like my subs a little hot compared to the MAINS. I have played with adding 10db boost to the sub just for music instead of lowering the MAINS signal and that definitely sounds much better. As you know miniSDP has 4 Presets. On 1 present I have the 8 db house curve with -20 db Input Gain. On another preset I have the 8 db House Curve with -10 db Input gain to recover the lost 10 db.

I believe that you should at least try moving your woofers out of the corners. You are looking at a 10db boost from 60hz & down. I suspect moving the woofers out of the corners will bring that down at least 6db. You should end up with a much smoother more balanced transition.

@Tim Hector The boost you are seeing is because of the 8db house curve added on the INPUT GAIN side of miniDSP, if I disable, it is flat like all the other speakers. I do not believe my issue is with subwoofer placement in this particular situation.

I’m not sure that you are looking at the LFE channel and the “10dB boost” correctly.
On a movie track the LFE channel is 10dB lower than all the other channels, so when the AVR applies a 10dB “boost” it is boosted back to the same volume of all the other channels.
Or perhaps I’m just totally misunderstanding what you’re getting at?

With respect to chest bump, this article may hold some clues.

Bass: the Physical Sensation of Sound

Ever wonder why you can feel bass? We take a look behind the science of bass waves and the tactile sensations that they cause. Our bass experiments were quite revealing but don't try them at home.

www.audioholics.com

@JStewart I am also not sure hence I am here looking for help. My expectation was the AVR will send the same single to LFE that it would have sent to MAINS (Left or Right) channel. Since LFE has a 8db House Curve, it should show up exactly like the normal LFE signal. In other words, I am not sure where I am losing exactly 10 db before the crossover point when the speakers are set to SMALL.

 

[ddude003]

It might be interesting to see an .mdat file to see the alignment of the impulse responses between the various speakers... Especially the L&R mains with the subs...

It might not be easy to get that "chest thump" when your subs are pointed at the back of the couch... Are you driving your L&R mains as full range (large) down as low as they will go (30Hz)... And if not, why not... They are made for that...

 

[JStewart]

My expectation was the AVR will send the same single to LFE that it would have sent to MAINS (Left or Right) channel.

There is a difference between the LFE channel and bass sent to the subwoofer from a bass managed speaker (i.e. a speaker set to small and an XO set in the AVR).

When you use REW to send a signal to the LFE Channel (typically #4 if using HDMI) it will be 10dB SPL higher than any other channel. The LFE channel is boosted by the AVR to account for the fact that when you play a movie that has an LFE channel track (the .1 in a 5.1 or 7.1 movie) the LFE signal is 10dB lower on the source, so the AVR has to boost it back to be equal to all the other speakers.

 

[achugh]

It might be interesting to see an .mdat file to see the alignment of the impulse responses between the various speakers... Especially the L&R mains with the subs...

It might not be easy to get that "chest thump" when your subs are pointed at the back of the couch... Are you driving your L&R mains as full range (large) down as low as they will go (30Hz)... And if not, why not... They are made for that...

@ddude003 The MDAT files are very big and I am not able to attach them. I have ZIPPED them all up and put them on my Google Drive for download. I hope you all can access them using https://drive.google.com/drive/folders/1msyWvIzToZxCBbU-tHufjNXdSeAWu0YI?usp=sharing link. If you have any trouble downloading them, please let me know. These were created using REW V5.40 Beta 19 and Beta 20 versions of the software.

When this ZIP file is extracted, you will see the following files. A brief description of how I organized is given below.

NOTE: You will need to update your REW preferences to be able to work with 100 measurements as some of these files have over 50 measurements. If you do not see all the measurements, please go to Preferences > View Tab > Maximum Measurements and ensure it is set to 100. Changing this value will require you to RESTART REW.

010_SubAlignmentOnly.MDAT - This is the file I used to align the 2 subs I have. AM stands for AntiMode. ByP stands for AntiMode being bypassed. I tried couple of different delays in miniDSP to find out that I get best alignment at 2.2 msec of delays between the two subs which are on the LEFT and RIGHT side of my couch. I set my AVR CrossOver to 200Hz to get the maximum output from LFE for creating a proper LFE EQ later before finding the best crossover.

020_SubEqFlat(20-150,18-200).MDAT - Using the aligned sum, I used this file to create a flat EQ using the REW's EQ panels. You will see I set the Target LFE at 95db to have REW generate the EQ with 1 db flatness. Then I added the 8 db Harmon Curve which is another target in this file. To add this harmon curve, I used the miniDSP Input EQs so that I can easily adjust this house curve to my needs or try few different curves. Previously I would make everything as I curve but I learnt that it is best to separate the two out. The measure that say LFE REQ Filters is a measure I took from within the EQ panel where REW has an option to take a measurements with the calculated filters and it adds these filters into the notes. The LFE EQ measure is taken with the filter loaded into the miniDSP hence the notes does not include the filters being used.

030_SubEqHouseCurve(8db).MDAT - This file has just 2 measurements showing how turning the miniDSP input EQ filters ON and OFF adds the 8 db house curve.

040_MainsAlignmentWithSub.MDAT - This file shows using 80 Hz CrossOver and speakers as Large (you will see SpL for Large Speakers and SpS for Small Speakers in the measurement names) how were the measurements coming when using MCACC default calibration.

050_SpeakerAlignment.MDAT - This file shows that I used the REQ Impulse alignment to find the manual distances so that all my speaker impulses align properly. The manual speaker distances are in the measurement names. All measurements were taken using 80Hz crossover with speakers set to LARGE as I was following the OCA manual home theater calibration setup procedures.

060_SpeakerMeasurements.MDAT - Ignore this file because I took some measurements without resetting the AVR settings so I had to go back to retake the measurements.

061_SpeakerMeasurements.MDAT - I took 6 measurements of each speakers at the 6 spots to cover all of my couch so that I can then create a vector average before creating the EQ for the speakers. This file has 48 measurements (6 positions * 8 speakers [7.1 setup]).

070_SpeakerVectorAverages.MDAT - Ignore this file because I took some measurements without resetting the AVR settings so I had to go back to retake the measurements.

071_SpeakerVectorAverages.MDAT - This file has all the previous 48 measurements from 061_SpeakerMeasurements.MDAT plus the vector averages of each speaker which has been cross correlation aligned. FYI, I did make the 0 position measurement of the respective speaker as my 1st measurement before clicking the cross corrlation align button as was mentioned by OCA in a later video as to how REW does this correlation alignment. Once the Vector Averages are created, I REORDERED the 0 position measurements in the same order so it is easy to see them. Just mentioning here that I did not make any mistake in this process.

081_SpeakerVectorAveragesOnly.MDAT - This file just as the vector averages hence all the traces are shown thick which is the default option in REW to distinguish between vector average trace and a measurement trace. I created this file so that I can quickly see what my vector averages look like without opening really large MDAT files. Also I switched to using 81 as the number to remind me that I made a mistake at 60 and 70 point and hence this is continuation from 61 and 71 measurements.

082_SpeakerFreqDependentWindows-15Cycles.MDAT - This file creates a copy of each vector average and applies the default 15 cycle Frequence Dependent Windowing followed by Psychoacoustic Smoothing so that I can see how the full measurement looks like. I later found that both these settings can be undone very easily so this copy was not needed. I created the copy because of following OCA manual claibration video where I did not realize these can be easily removed at that time.

090_SpeakerEQ.MDAT - This file is a copy of the 082_SpeakerFreqDependentWindows-15Cycles.MDAT file where I started to create the manual EQ for each speaker using REW EQ panel. You will notice that I added a target of 85 db measurement since my AVR volume for everything I have done is -10 db. I consider 0 db AVR as reference of 95 db for speakers and 105db for LFE. If you click on any of the FDW measurement and you open the EQ panel, you will see the target and the EQ that REW calculated for each of my speakers. All speakers as per OCA were calibrated as LARGE (they are truly LARGE but even if they were not according to OCA LARGE should be used so that is what I followed).

100_SpeakerSPLAdjustments.MDAT - I then loaded all the generated EQ from 090_SpeakerEQ.MDAT step into AVR and took fresh measurement readings. Then I applied 15 Cycle Frequency Dependent Windowing to each speaker measuement followed by Psychoacoustic Smoothing to them. At this point, I created a copy of each of these measure (except LFE) and used the SpL Alignment tool to use 1Khz as the alignment point with 4 octaves of range for SPL alignment on the copy. The measurement names shows the setting. At this point in OCA video he talks about manually moving the original speaker measurement to find the offset but I found that using the INFO button I can see the ALIGN OFFSET value very easily so that is what I used to set my Speaker Channel Levels in the AVR.

110_SpeakerValidations.MDAT - In this last file, I took various readings using different CrossOver points (XO) with speakers being Large (SpL) or Small (SpS) and found that I get best smooth response when I set my speakers to SMALL and cross over as 80Hz.

This options of setting speakers as SMALL in the end and finding the best cross over has been mentioned in this thread as well as all over the YouTube videos and various forum threads. My understanding is that we buy Large speakers to get the smoothes crossover as speakers tend to lose their smoothness as they get close to their marketed low-end of the audio spectrum. I am also not able to do Bass Management like miniDSP type to my LEFT & RIGHT and even CENTER channel due to limited EQ options in AVR so I believe using SUB with crossover make it smooth and cleaner. Originally I was under the assumption that I could use my RF7II tower speakers as 2 additional sub to smooth out the bass response. That is proven wrong with these measurements as I cannot control the bass frequency response of them like my subs. Now If I buy another miniDSP and then run my Left and Right speakers via external amp so that I can base manage the Left and Right speaker as well then this discussion changes but that's not an option for me at this time.

There is a difference between the LFE channel and bass sent to the subwoofer from a bass managed speaker (i.e. a speaker set to small and an XO set in the AVR).

When you use REW to send a signal to the LFE Channel (typically #4 if using HDMI) it will be 10dB SPL higher than any other channel. The LFE channel is boosted by the AVR to account for the fact that when you play a movie that has an LFE channel track (the .1 in a 5.1 or 7.1 movie) the LFE signal is 10dB lower on the source, so the AVR has to boost it back to be equal to all the other speakers.

I thought REW sends the same level of signal for LFE as it does for every other channel. Because AVRs assume LFE is always 10db less it does a 10db LFE boost hence in REW we typically see LFE outout as higher.

REW recently added an option to send LFE-10db signal to match the typical Dolby tracks so in REW LFE does not show up boosted.

 

[achugh]

@jtalden Is it possible for you to look into my issue and see if you can spot anything that can be helpful to my issue?

 

[jtalden]

If the basic question is still how to get that slam feeling, my general thoughts are similar to those already offered. I'm not the expert here and have little if anything to add. I have only skimmed the thread and have not looked at your mdat files. That said, here are the key elements in my opinion:

• A smooth SPL response to a well-chosen house curve is a key ingredient. My preference is to a house curve that is close to the B&K curve and would speculate that the very large low frequency boost may actually be working against your concern. You may want to try other curves. A 2-4 dB rise in bass below 200Hz is more to my preference, but there are many factors involved so experimentation is your friend.
• @JStewart provided good advice and a reference to the fact that the chest slam is more closely related to the 100-150 Hz range. Some of the charts you posted show a significant dip in that range. That dip is likely working against you. Since the dips are likely related to room effects, relocation of the mains or LP, or the SWs may be required to address them.
• The timing of the 80 Hz XO may also have some impact in that range as well. If you want me to confirm your XO timing, I can do that if the needed measurements are provided. The 80Hz XO is not necessarily the best choice. For example, If the SWs are smoother through the 100-150 range than the mains are then raise the XO frequency to see if it helps.