Readjusting the Bass EQ

I was hoping I'd just leave things be after my big EQ event on Father's Day when I added the chairside EQ and found that a bit of bass boosting helped.

But it did not in fact help universally, and it's always worth remembering why.

Boost is the enemy of dynamic range.

Often in those areas where boost is needed, the speakers etc are already struggling.  Adding boost can too easily make a marginal situation, where there is marginally enough headroom, into a terribly bad one with rattling, arcing, and buzzing.

This phenomenon reared it's head when I was playing one of my own experimental improvisations, an as-yet-never-released album of improvisations I call The Ways Around.  Many of the tracks on this album will challenge any audio system, by design or accident.  A particular track, Slide, caused out-and-out buzzing.  Well I could play it at lower level, and just a couple of dB below my previous maximum level setting it was fine.  But it bugged me.

I promised myself I would just "take a look" at my old EQ settings and do nothing.  And I kept that promise for about 24 hours.  Finally I decided "just a little change."  That's how it starts...

I think I had documented the previous settings in both channels, but only re-recorded it on one channel this time.

It was not fully "starting over" but most everything was thought through, and swept with my Krohn Hite oscillator, to be sure I wasn't making things worse.

I decided fairly quickly not to boost the highs in the bass or the deep bass in the panels, so that means I'd bypass my new chairside EQ and dial in new per channel graphic EQ's for boost separately in the Bass and Midrange EQ units.

The per-channel part is important in compensating for speakers, or complicated systems of speakers like mine.  They will always need per-channel adjustments.  Doing as I was doing before, just applying boost of up to 6dB in key areas of deep and mid bass was just asking for dynamic range trouble.

Finally, I created a buzz doing oscillator sweeps that sounded a lot like the buzz I heard in the left channel playing "Slide".  With the crossover still switched in, I had to crank up the level quite a bit to make it loud enough to even happen.  Then I discovered something.

I was overloading the input A-D of the Tact preamp with the oscillator.  THAT was what was causing this buzz.

So then I thought a bit...if I'm adding up to 6dB boost, and the signal feeding it has peaks greater than -5dB, even if I'm giving the chairside EQ a 5dB cut, it's still going to clip.

So now it looks like the buzzing distortion that inspired this deep dive may have been simple clipping all along, not an acoustical buzz.

But a re-EQ was long overdue anyway.  Not done in two or more years IIRC. 

So I continued my  deep dive into re-adjusting the bass EQ all afternoon Friday.  I always intend just to try one thing, but soon the entire set of adjustments is up for grab.

I never actually eliminated all the parametric EQ's (PEQ's) and started from scratch.  There are really two sets of cuts, typically 39-40 and 45.  Those correspond to the most major lower nodes, then 70-125 (left sub seems to go crazy 100-120 Hz so I cut it back, but other channel too).  I did each of those sets without bothering much with the other which are fairly independent.

In the left I came up with a greatly simplified set of EQ's which corrects sub and panel issues, boosting panels 100-125 (below 125 crossover) and meanwhile cutting bass.  This seemed to relieve a kind of congested sound, while maintaining gradually downward sloping response through the bass, the target.

But sadly, not as nicely flat in stereo as I could get it to look without bothering with the niceties, which despite this one buzz actually having been a non-issue, was still apparent on other recordings.  

So individual channels looked better, but the stereo response (to"correlated" pink noise) looked far worse!

Stereo Response after new bass EQ

The stereo response shows a bass rising from the lowest frequencies up to a major resonance at 71 Hz, then becoming significantly notched out 90-125 Hz.  But neither channel shows these features!

Right Channel after new bass EQ

Left Channel after new bass EQ

Both the individual channels show an almost monotonically rising response going downwards from 250 Hz.  So how does that get inverted in the stereo response to a major peak at 71 Hz ???

There is a tiny peak in the left channel at 71 Hz.  I might add that I worked mightily to keep that peak tiny and I could not eliminate it completely.

Both individual channels show the response rising all the way to 20 Hz, but the stereo response shows response falling from 71 Hz down with a tiny rise just at 20 Hz.

So, the plan now is to remember these new settings (they were photographed and saved to preset "X"), but go back to the same old, using the chairside EQ and everything.  And be sure I can duplicate the original wonderful looking stereo response.  Then try to improve that per-channel as before, BUT always checking the stereo response afterwards, to see where the stereo response gets messed up.

I tried the simple test of turning off the graphic EQ boost in both ways and channels.  I still see the bass sloping on either side of a 71 Hz peak, but now falling even more on the low side.

****

Following the plan I developed on Friday Night, I proceeded to restore, to the best of my ability, the exact Bass EQ settings I had been using merely a week ago when I first dialed in the new boost into the chairside EQ and was pleased by the new response curve.  Sadly, the most recently photographed bass EQ's were from December, but I didn't think there had been any changes since then.  (Knowing what a can of worms adjusting the Bass EQ is, I generally leave it alone, but sometimes make a 0.5dB change here or there, or adjust the relative subwoofer levels.)

The result, now stored in the memory named RESET, was disappointing.  When I applied the new bass boost through the chairside EQ, I did not restore the flat response gradually rising to the deep bass.  Subjectively the flatness was no better than with the quick re-adjust I had done on Friday.

I decided to go back to my newest bass settings, and continue further adjustments based on per-channel and per-side sweeps.  I also tried substituting some of the earlier adjustments individually or in blocks.  Some of the old adjustments were actually pretty close to or the same as new adjustments.  I'm just beginning to think that I had finer resolution with the old General Radio oscillator I had been using a couple of years ago compared to the newer (but still 1960's) Krohn Hite.

After hours of messing around with these settings, I was finally getting something that looked almost as good as last weekend (with a similar extra boost in the chairside EQ):

In this response, I was able to tame the bulge from 71-85 Hz using far bigger cuts, up to 12dB, than I had used before.  My December EQ had two nearly overlapping 9dB cuts in 70's and others nearby.  To achieve flatness in that region I'm now using fewer but bigger cuts.

I tried mightily to depress the 80 Hz micro peak, but it just wasn't going away quietly in the night without creating bigger problems around it.  Likewise the micro dip at 112 Hz.

I'm still not entirely happy with this, but it's far better than the stereo response shown above.  Unlike some intermediate versions, I am NOT doing any cuts in the chairside EQ, only boosts, and I think it's useful to keep them separate.

This "stereo" response also includes an attenuation of the right channel by 1.1dB, added to correct an apparent channel imbalance measured by C weighted SPL app.  (That is set in the Tact.)

I have saved these settings as "SUN" in bass, midrange, and chairside EQ's.

(Lesson One: when you find settings that work pretty well, save them !!!)

I then started looking at the 4 lowest bands, hoping to eliminate the slight depression around 28 Hz.

I tried reducing the 20Hz boost, and it appears that I can reduce the 20Hz boost to nearly zero and still have a slight rise at 20Hz.  That might be preferable as 20Hz boost is throwing away dynamic  range right where it is needed.  But it still looked depressed at 28 Hz.

But it seemed MASSIVE amounts of boost, 12-15dB would be needed to correct the shortfall at 28 Hz.  Massive boost is not a good idea so I gave that up.

But then I decided to check the settings on the right subwoofer, which did not appear to be putting much out at 28 Hz.

I discovered that the right sub had it's mode control set to "Extended" which corresponds to having one of three ports sealed.  Meanwhile, all 3 ports were sealed.

(Herein lies a long story.  The right and left subs are not in symmetrical locations.  The left sub is in a corner, and the right sub is next to the door and a hallway.  I determined long ago I liked having the doorway sub on the right "sealed" and the sub in the corner in "extended" mode.  The corner sub gets extra 20Hz frequency boost that makes up for the lack, or so I thought, of the lowest frequencies.  The sub on the right sounds too honky unless I seal all the ports.  I thought to myself "the best of both worlds" as the ported sub gives me endless bass dynamic range, and the sealed sub gives me deep bass...or so I thought.)

Was this wrong or was this a design choice I made earlier?  Back when PB13's had 4 mode settings (which I miss btw) you had 4 filter mode choices

20 Hz tune (all ports open)

15 Hz tune (two port open)

one port open (not supported)

Sealed (I would have believed this to be the "lowest" frequency, sacrificing some headroom)

The manual had a warning about selecting a "lower" tune than would correspond to your number of ports:

NOTE: It is strongly recommended you do NOT run a lower subsonic filter point than your port plug configuration would dictate. Failure to properly align Subwoofer Tune and port plug configuration can potentially result in serious damage to your subwoofer, especially if run at higher sound pressure levels.

Since it didn't say "higher" I thought it was OK to have, for example, two ports open with the 20 Hz tune, or one port open with the 15 Hz tune.  In fact the latter was what I tried for several years before going to completely sealed, which seemed to reduce all sorts of bad sounds possibly resulting from reflections down the hallway.

SINCE those days, I have moved the Bag End E-Trap to the corner of the hallway 10 feet past the subwoofer.  This means the sub is firing into something virtually like a damped trap, so the reflection issues may be much lower than before.

ALSO eq and many other things have changed many times over since those days.

Anyway, I might have done this tuning mismatch deliberately, and in fact I might continue doing it (or just re-open the two ports corresponding to EXTENDED mode).

THE problem with having all the ports sealed and using SEALED mode is that the lowest frequencies, like 20 Hz, just vanish.  Sure if you got down to 8 Hz or so there might be more output from the sealed, but you lose 10dB at 20 Hz.  But if I dial in the EXTENDED mode, I get that deep bass, but at the cost of the 25-28 Hz bass.  THAT is what I had been doing, and that is what mainly explains the dip I had at 28Hz on the 1/6 octave RTA from Analyzer app.

If I choose to do things the correct way for SEALED I get 10dB loss at 20 Hz.  (I would not have expected it to be so bad, but perhaps it's accentuated because the phase mismatches the other subwoofer more.)

Seeing the much better response immediately below 20 Hz of the Extended mode...I'm not so sure it was safe to run it EXTENDED with all the ports sealed...the paragraph above also says "failure to align" and doesn't say what's safe for having all the ports sealed.

Do I want a 10dB depression at 20 Hz or at 28 Hz?  Well, neither, and to get that I'm going to have to go fully into "Extended" mode and open two ports.

The plate amplifier on the Right sub lasted the longest time, with all the settings I'm describing.  But it eventually burned out, and the replacement plate amp had to be replaced already.  Perhaps my mis-tuning is not such a good idea.  I'm worried about that now too.

IMPORTANT NOTE.  All of the following responses have the chairside EQ enabled with 6dB bass bost in 20, 25, and 31 Hz bands, except the last one I removed the 20 Hz boost...obviously no longer needed!

Stereo Response Right Sealed/Sealed

Stereo Response Right Extended/Sealed

Right Channel Extended/Sealed

Right Channel Extended/Extended

Stereo Response with both subs Extended

No Boost at 20 Hz Stereo response

Pretty much as expected.  Going to Sealed/Sealed was a disaster to the sub 20 Hz response.  What there is seems to cancel the other channel.

Sealed/Extended sort of works (if dangerous) but has limited LF output and dip around 28 Hz that's hard to correct.

Extended/Extended has full response.  Boost at 20 Hz no longer needed...and still more could be cut.

Beyond the measurements, the sound is, even at the listening position, no longer devoid of the punch-you-in-the-gut power that you'd think my systems should have had for the last decade but were somehow missing.

I'd say Extended/Extended is definitely the way to go in both subs.  Now it's a matter of dealing with the opposite conundrum...too much response at 20 Hz.  But that's easy, and 10dB of notch filtering will give an added 10dB of dynamic range.  A win-win!

I tried using different Q's in a notch filter at 20Hz.  Seemed hardly to make a difference, and it always caused reduction nearby (perhaps because I was only doing the right channel).

So I just decided to dial in 15dB (!) of reduction at 20Hz in the chairside EQ.  It's still tipped up at 20Hz, but pretty nicely anyway.  If I ever need more 20 Hz punch, I can just raise this slider.

Note that if I try messing with the 25 Hz, it just messes everything above all up.  Just sticking to the 20 Hz slider I'm OK.

This is the best looking, the flattest, and the closest to what I believe is the correct room curve for a line source dipolar speaker at short distances I've ever achieved.  And it comes with an extra 21dB dynamic range at 20 Hz so I can play those bass tracks without fear.  (No 6dB boost at 20 Hz anymore, now a 15dB cut!)

Stereo Response with -15dB at 20 Hz

It's easy to see why I overlooked having the right sub in sealed mode.  Somehow, though it's not a corner, it couples to a 20 Hz resonance including the entire house, which makes it awfully boosted.  To make it sound right, one needs to cut 15dB at 20 Hz, as I am now doing.  AND in both channels, so the phase between the two channels is constant.  Then one gets powerful response all the way to sounding nearly effortless.  (But I still need to boost 31 Hz, also in both channels.)

If I had thought about the measurements presented in the Audioholics review, I would have never selected anything but the extended mode.  The sealed mode starts rolling off at 40 Hz.  What kind of a subwoofer is that?  No wonder I kept needing to add 31 Hz boost to the right channel.  Meanwhile the Extended (15Hz) response does go into the subsonic even without room gain.

****

On Tuesday (June 28) I tried replacing the -15dB graphic EQ (GEQ) adjustment with a parametric EQ in the chairside EQ.  Basically I could not do any better with PEQ than with GEQ, and both are limited to a reduction of -15dB.  So then I tried augmenting the GEQ cut, which is 1/3 octave, with an even steeper PEQ notch at -15dB and other levels.  That seemed to help some of the very deep bass "boom" which has arisen using 2 ported subs.

The RTA looks little different with the added 1/7 octave cut at 20 Hz.  If I cut much wider, a gap starts appearing in 25 Hz and 28 Hz on the RTA despite the 6dB boost being added at 25 and 31 Hz.

It seems that a total cut of at least -22dB is required somehow, and the DEQ only allows maximum cuts of -15dB (which is stupid...long ago Dick Burwen showed the need for -48dB and larger cuts).

I was testing using the Bass Connection CD, which I listened to over and over, and a few others.  Sadly it seems like my Krohn Hite 4200A oscillator has gone dead below 40 Hz, it can do higher frequencies but if I do lower it takes some time at higher frequencies before it recovers.  I've ordered a manual-on-CD but meanwhile hope to get out my General Radio oscillator from storage, which was possibility was blocked by heavy thunderstorm on Tuesday evening.

So this does need better testing, but for now I have -10dB cut at 20 Hz in the graphic EQ, and -12dB cut at 20 Hz with bandwidth of 1/7 octave in the parametric EQ, giving a total cut of -22dB at 20 Hz.

The RTA response looks little different.  There's still a rise at 20 and 25 Hz.  If I cut any more, 25 Hz starts showing a dip and possibly 20 Hz as well.

It occurs to me that if you put two notch filters in sequence, the resulting notch filter will have half the Q, just as with high pass and low pass filters.  So, two identical 1/3 octave filters in series would give you something like a 2/3 octave filter.  To maintain the 1/3 octave bandwidth of the graphic EQ, the Q of the parametric would need to be infinite.  By choosing a 1/7 octave filter (which btw seems recently to work best for a lot of things) I'm getting a resulting filter of just over 1/3 octave, perhaps closer to 1/2 octave.

When I say "at least -22dB" I mean I don't really know how deep the cut should be...because there's still a trace of that awful whole house 20 Hz boom that only the Extended mode of the SVS seems to excite, and most readily from the right subwoofer which is near entry and the main bedroom hallway.

I spent some time Tuesday trying to tweak the Bag End E-Trap in the corner a bit better.  It was now greatly over-reacting to some things, so I dialed down the 40 Hz-ish notch.  I noticed that the 20 Hz notch had been turned off (probably didn't matter when I was using the right sub as "sealed") so I turned that on with a pretty conservative setting too.   With the Bag End, the short hallway to the corner is fairly uniform in level, otherwise you get incredible BOOM in that corner, though I don't have measurements proving this (and certainly not with dead oscillator).

The first priority of the day had been setting the subwoofer levels, which had never been done since flipping the subwoofer mode.  Using pink noise and measuring with A-weighted meter, it seemed like the panels needed somewhere between 0.5 and 1.5dB of attenuation in the right, and the sub needed about 2.5-3.5dB of attenuation.

I attenuated the right sub by 2dB (from -7dB to -9dB) using the SVS app.  I don't know what the other sub is set to and it's hard to find out, it's probably about the same...I had previously given the sealed side a slight boost to compensate for its lowered output (but still using the Extended mode, which may not have been a good idea).

And I'm adjusting the relative channel level in the Tact (maddengly the Behringer's don't have this feature).  I will just have to remember that I do this.  I'm not sure why it's necessary on the panels as well, I think probably because of the difference in acoustical loading and coupling.  The left channel is in a corner and the right channel is in a hallway.

And I sometimes think that balance should be adjusted using a mono test.  (BTW, my pink noise IS a mono test.)  Mono should appear to be dead center in phantom stereo.

Over the last 5 or so years, I've gotten much better at doing phantom stereo WHILE being pretty close to large speakers for the maximum stereo impact.  Many with a room like mine would listen in the back of the room, which has stronger bass among other advantages.  But a friend convinced me that the widest stereo possible is best...you hear far more and better separated.   The widest angle that still allows for a solid phantom center image is best...and it IS about 30 degrees.  I've seen a lot of people with much narrower speakers, I think 25 degrees is more typical, especially with back-of-the-room listeners.

But I think image centering is as much or more a factor of the sonic arrival time.  So I am very very sensitive to both channels arriving at precisely the same time for a phantom center.  So much so I am often moving my head very small distances to get in that precise spot.  But I believe it may not have much to do with center frequency levels.

When you get both the time and the levels correct, as most high end digital correction systems do, THEN you have a centered phantom image AND a solid one.

I think that's what I have now but will need to do more testing.  I still have the chair positioned so that it just covers up the tape marking where it's supposed to be

****

Finally I decided to re-examine the Behringer Dynamic EQ (DEQ) I have set up for the bass in the bass equalizer, but it needs to be re-set every time I change levels.  For now I've moved it to the chairside EQ by BYPASSing the one in the Bass digital equalizer (DEQ).

Overbearing and rattling bass was an issue with Bass Connection even being played back at -10dB now.  What's needed is a good limiter and the Dynamic EQ is one of the best...though it could have been better.

One thing I noticed this time is that I needed to set the corner frequency higher than expected.  Excess bass starts becoming an issue around the 40 Hz region and lower, but if a corner frequency is set to 40 Hz it does very little, because all it has there is about 3dB of rolloff from the dynamic filter.  What you need to do is set the corner frequency an octave higher, around 80 Hz, to start getting real attenuation at 40 Hz.  And that is WITH the 12dB slope, so you'd be getting 15dB attenuation by that point.

And sadly -15dB is the absolute maximum attenuation the Dynamic EQ can provide, just as for Parametric and Graphic equalizer functions.  That's a terrible limitation here, it means that just as you're starting to get serious limiting it ends and you can't get any more.  And sometimes you need more, especially going to very low frequencies where the subwoofer is getting closer and closer to its limits.

Anyway, it's still useful, and I've dialed in a Dynamic EQ which seems to make Bass Connection playable (though perhaps could be made better).  I've chosen 30 msec attack and 300 msec release, to allow bass transients and expression to come through.  I cannot hear it working except by the relative absense of bass booming, still not perfect until I dial in complete dullness by lowering the threshold.

It appears now however that I really have to implement this Dynamic EQ for the subwoofers in the digital equalizer dedicated to the subs.  Because the it is triggered by the signal level, and there appears to be no way to change that.  I don't want my bass rolling off because cymbals are crashing, etc.  I only want my bass being rolled off when there is too much BASS.  It would be better yet if the Behringer allowed you to filter the control signal for the DEQ to frequencies of interest...I had previously assumed that it did so corresponding to the corner frequency of the DEQ, but it now appears that it does not.  So for use in the chairside EQ I must set the Threshold to -26 or so, but if I did so in the bass EQ I could set the threshold lower, maybe to -20dB, and be unaffected by the levels of higher frequencies.  That would permit more thorough filtering.

I also need to move the narrow parametric cut at -20dB to the subwoofer equalizer so that it applies to bass.  In fact much of the EQ should ultimately be moved to the per-way equalizers.  Until I do that, it does not get applied when my system is set to the FM radio through my "Background Music" control.  Right now, the lack of 20 Hz attenuation is becoming an issue even for FM radio.

*****

Doing measurements before, with the 20 Hz notch bypassed, and then moved to the subwoofer equalizer showed very little difference.  You'd think that notching out 20 Hz with a -12dB  1/7 octave notch would cause "20 Hz" on the iphone RTA unequivocally go way down, but it seems to make only a few dB difference amidst much larger variability and warm-up (the RTA doesn't stabilize on 20 Hz right away, but slowly climbs for about 1 minute).   (The RTA 20 Hz is actually 22 Hz, but the lower 1/6 octave of the band should capture at least half of the effect of a 20 Hz notch, I would think.)

Anyway, it seems to show about the same 2dB or so loss at 20 Hz regardless of which EQ unit I use to perform the notch, so I now consider the notch "moved" to the subwoofer EQ unit if not necessarily perfected.  I needed to move it to take the 20 Hz "boom" out of listening to FM radio on the Background Music selection of my system, which bypasses the Tact and the chairside EQ.

I can only make the notch easy to see by increasing the width to 3/4 octave or more, by which point all the bass below 40 Hz is being lowered.

I'm still dialing in a 10dB cut at 20 Hz in the chairside graphic eq, but only -10dB seems sufficient, and it's even hard to tell if I raise it to 0dB.

If I pull down the lever enough to bring 20 Hz down to flat, it wipes out the 22-40 Hz response.  These things are far more complicated than they look.  You can't just set the EQ to the inverse of the error response.  Notches affect other bands and they often seem to do so in non-linear ways.  Some bands seem "weak" and are easily affected by other bands.  Other bands seem "determined" and refuse to give up their peaks no matter how much notching you throw at them, and when they finally do, they bring down all the other nearby "weak" bands with them.

I also noticed that fairly small changes in the position of the phone can make the apparent peak at 20z. (even when notched out) go away and be replaced by a dip, while the low peak moves up to 25 or 31 Hz.

I can't explore such frequencies very well because my Krohn Hite 4200A has lost it's lower frequencies.  I've wasted a day looking at both new and used oscillators, but probably the best bet is to fix the 4200A which may only need a bridge adjustment.  I've ordered a manual on CD.  The Krohn Hite company is still around to do refurb/repair also.  I think my General Radio 1300A wasn't working well or at all either.  My B&K has uselessly low precision at low frequencies and only goes to 28 Hz anyway.  I have a bunch of other things that can generate but only in fairly large 1 Hz increments, which is too big at 20 Hz.

****  Update July 2

I got my next working oscillator, a Tektronix CFG 253.  This is probably also nearing the life stage where it needs to be refurbished, but it's still working mostly, and a bargain for $70.  Somewhere below the bottom 0.2 marking on every band it loses output.  I suspect that's not a design feature but a similar kind of ageing that has affected my other oscillators.  But I can measure 210-17 Hz on the 200-20 Hz band, and 21-1.7 Hz on the next band, so I can cover the regions of interest, and the Tek remains flat until the point where it loses output altogether.

This is much cooler than any oscillator I've ever owned.  It's compact yet fully featured (for an analog oscillator).  All plastic so it doesn't scratch anything.  It was possibly the 3rd generation of Taiwan made Tektronix products, so they got the production working OK.  I think it's from the 1980's or 1970's.  It has wider bandwidth and is basically better in every way to the Krohn Hite and General Radio.

I looked at new "arbitrary waveform generators" like the Sigilent and a cheaper one on Amazon.  I'm worried they might not be as nice for manual acoustical sweeping, so I decided to stick with what I'm comfortable using, an analog oscillator.  Neither the Tektronix nor my Krohn Hite have a log sweep knob however like the General Radio does.  But I've decided a linear control knob is OK too (though I liked the log kind better).  Arbitrary Waveform Generators should have a mode to mimic that kind of control.  I looked through the Sigilent manual and there didn't seem to be such a thing, only an ability to vary any particular digit of the frequency (and one Sigilent review showed it not keeping up if the knob is spun too fast...something you wouldn't expect in a Sigilent but likely in the cheap knock offs too).

Even with all the 20 Hz cuts turned off, I hear no actual boom at 20 Hz, both at the listening position and elsewhere.  There may be excess output at 22 Hz (which is the actual center frequency of the lowest band on the iPhone Analyzer app) but that may well be because I boost 25 and 32 Hz, otherwise they're quite depressed.  There is also a bit of chuffing and wall shake in the 17-18 Hz region, but it has already rolled off at that point anyway so there is no boom.  The 18 Hz chuffing and shake was unaffected by my 1/7 bandwidth PEQ but affected by the 1/3 octave GEQ cut.

For now I've turned off all the 20 Hz cuts.  They do not actually appear to be needed now.  I can still dial back in a 20 Hz cut in the chairside EQ when I need it.

Sadly the Behringers give no way to center PEQ's at less than 20 Hz, or I'd fix a bit of the chuffing.

Now however I need to look in the 20-30 Hz range to see what's going on.  I'm still boosting at 25dB, is that a good idea???

******

On Saturday July 2, I did some more adjusting based on per-channel sweeps and pink noise.  There seemed to be a big bulge around 40 Hz that I worked to suppress by adjusting the notches in both channels at around 39 Hz.  I made them as deep as -12dB, much deeper than before.  After response to sound and look pretty good, I discovered I had been applying 40 Hz boost through the chairside EQ.  That must have been an accident, I thought.  Or maybe it wasn't.

Thinking it made little sense to boost 40 Hz by 6dB, and then to cut 39 Hz by 12dB, I removed the boost and tried to roll the cuts back.  First I tried to restore what I thought was the previous EQ, but again the result looked worse.  So I was back to sweeping and pink noising and adjusting just to try to make things look as good as they did before.  But I did not succeed.

After Removing 40Hz boost on July 2

The chairside EQ boost was lowered to 3dB at 28 and 31 Hz as those areas could be boomy.  I also moved the Dynamic EQ to the subwoofer EQ so that it tracks the bass level only, then I was able to turn the limit down to -27dB without fear.  It seemed to be working at least as well as before.

 ****

On July 3, I managed to get the 20-40 Hz looking relatively flat again, with only a slight upturn at the bottom as before.

I reverted the 39 and 45 Hz cuts in the left channel to those used on the previous Sunday.  The left sub hasn't changed at all so there was no reason to change the EQ's.  I dialed back the 8dB cut at 39 Hz in the right channel to 6dB.  The previous Sunday that cut had been 2dB, which was clearly inadequate with the now ported sub.  Strangely, if I now set it to 2dB an apparent dip seemed to appear.  A greater cut yielded flatter response.  This could be a channel matching thing (these pink noise tests were stereo) or perhaps more likely a visual illusion.  But it seemed flattest at 6dB cut.

I rolled back up the boost at 32 Hz to 6dB.

All these settings are now saved as J3 in the chairside EQ, subwoofer EQ, and panels EQ (which hadn't changed since Sunday).

Sweeps are now not showing any need for attenuation at 20 Hz.  (Of course, I'm not boosting 20 Hz anymore either...that led to the need for some of the past cuts.)  Or 22 Hz.  The visible bump up at 20 and 25 Hz on the RTA is not audible on sweeps on either channel.  (It occurs to me now, I should try sweeping both channels at much, which might uncover interactions.)

This has bass like I've never head it before last weekend since converting to Ported/Extended on both subwoofers.  In the listening position, which I previously believed had an unalterable lack of bass and impact, you can now feel the bass when it's loud enough.

I've also reset gain controls on the DAC and elsewhere so I can play about 5dB louder than I could before, which is sometimes helpful.  It's hard to imagine going to 0dB with an ordinary commercial recording.  -5dB on the chairside EQ is the same as 0dB used to be.

Here is what the new J3 (July 3) EQ's look like:

Chairside EQ

Right Channel Sub (above) and Panels

Left Channel Sub (above) and Panels

Update July 5

In decades of audio experiments, it's been hard to keep "notebooks."  I often wrote settings and things I needed to remember down on cards (and still do, prior to writing up in blog).  I sometimes with great fanfare started notebooks and then simply abandoned them for lack of time a week or two later.

But now I have this blog as a public notebook.  And it's clear now that it works best if I follow this plan:

1) Open blog and start writing about today's plan or idea

2) Go off and do the most relevant or important experiment

3) Write up results and next ideas on blog, ASAP, before doing too many more adjustments and experiments to remember (which is about 1 or 2 at most).

Writing is probably the single best way to organize thinking, otherwise it tends to get disorganized, and in the heat of battle chaos ensues, then one finds it's hard or impossible to get back to where one was before which then becomes the order of the day.

This has been apparent in the course of writing the blog post, which I've decided to keep as a continuous stream as long as I'm doing this bass adjustment.  I thought maybe I might end it with the J3 adjustment, detailed above.  But already there are new questions.

One is that contrary to my writing (and memory) of the situation, I do still have a pretty large notch dialed in at 20 Hz, which I may have forgotten about because it's in the Chairside EQ, and I was furiously working the Subwoofer EQ which is currently all parametrics (I bypassed the graphic EQ when I found that my first attempt to "move" the Chairside graphic EQ settings to the Subwoofer and Electrostat EQ's failed or seemed too, in a very very heated "battle" that day.

So when I was sweeping and heard no particular bulge, and even perhaps a depression in the 20 Hz range, that was being affected by this now 10 dB 1/3 octave cut at 20 Hz.  What if I dialed that out?

Well, as I was thinking I should try with and without the cut to see what effect it has (IIRC when I was trying that, my impression was very little or ambigous on the phone RTA, but there was very different boom levels with the oscillator, which then started crapping out at 20 Hz so I bought a "new" ancient analog oscillator), I was listening to an FM radio recording with the announcer very boomy, and the station itself sounding about as heavy in the deepest bass that I'd want, if not a tad more, I'm thinking that 20 Hz cut may in fact be helpful, and perhaps I should increase it to a maximum 15dB cut.  At some time prior I was doing not only that, but also notching it with a 1/7 octave notch with the total attenuation between the two being about 22 dB.  It's almost hard now for me to remember all this, but I remember also that I wrote about it up above in this very blog post.

RTA's with and without the -10dB cut at 20 Hz show the "20 Hz" (actually 22 Hz) band about 5dB lower lower with the cut, with the 25 Hz band looking only slightly changed and nothing above changed.  One could go either way just looking at the picture, and even ask if (with no cut) I should raise 25 Hz instead.

But sweeping now shows two fairly small peaks around 22.5 Hz and 18.5 Hz.  The 18.5 Hz one is probably worse because it causes more chuffing and rattling (and obviously driver strain).  I tried notching out the 22.5 Hz one and that had less undesired attenuation at 20 Hz (which becomes a slight depression with the graphic EQ) BUT it had no effect on the 18.5 Hz.

I now find I can simply prop up the phone on the back of the listening chair with the SPL app running which makes it possible to do these measurements better.

The best solution presented to me now is using the graphic EQ at 20 Hz with a cut ranging from 0 to -15dB, as needed.  What I had set it at for J3, -10dB, is a good compromise.

Sadly the "new" oscillator seems to be losing low frequencies more and more, and it doesn't go deep on the 2-20 Hz band either, it just loses it around 12 Hz and you have to dial it way back up to get it going again.  This seems to be the common fault with all analog oscillators and I could probably fix any of them with the appropriate manual.

But now I AM lusting for a new "Arbitrary" signal generator, simply to measure very low frequencies at very finely adjusted and repeatable increments.  It's not easy to do that on the Tek around 20 Hz because of its linear knob.  You have to breath on the knob to get it to do less than 2 Hz on the smallest turn.  The Krohn Hite, as designed, goes down to 10 Hz (but mine dies around 30 Hz now) which is better.  But the similarly now non-functional General Radio with it's big log knob could do much better, when it worked.

This brings to mind one way in which the simple (invert the RTA) approach to setting EQ could not work.  It seems when there is one big peak, it causes nearby bands to rise a little, actually "covering up" what may be going on in there.  And the peaks don't necessarily align with the RTA center frequencies.

After that I decided to check out the apparent peak around 80 Hz that I noticed while sweeping (sitting on the floor btw).  I noticed that most of it seemed to be coming from the right channel, which does NOT have a notch around 80 Hz while the left channel does.  So I added a -4.5dB at 1/7 octave at 83 Hz, which seemed to fully erase the problem in sweeping.  I called this new setting J5 (since it's July 5).

But the RTA now showed a significant depression at 80 Hz.  So I went back to J3.

Meanwhile the RTA was showing an excess at 250 Hz.  Sure enough I was boosting that by 3dB in the chairside EQ.  I removed that boost completely.  That resulted in perhaps the flattest overall response I have ever measured.  I saved this as J5A in the chairside EQ, which should be used in combination with J3 elsewhere at this point.

July 5, 250 Hz Boost removed, now virtually flat to midrange

This actually seems to have more "slam" and "power" in the bass than responses that looked turned up (probably mostly because they had a few boomy frequencies).  And don't forget I couldn't get this without both notches AND boosts.  I'm still boosting between 25 and 31 Hz, and around 250 Hz just not 250 Hz itself.

Remaining issues include:

1) Can I move adjustments out of chairside EQ so they help with FM tuner sound also?

(When playing FM in my "background" mode, the lack of cutting 20 Hz seems to be made up for by the lack of boosts at 25 and 32 Hz.  But the midrange stuff could be moved, maybe.)

2) The Dynamic EQ (dEQ) doesn't seem to completely remove issues when playing Bass Connection.  It's seeming to me that the maximum 15dB attenuation provided by the dEQ isn't sufficient, and I'm wondering if I can stack two identical dEQ's to get better results with up to 30dB attenuation and does the Behringer DEQ even do that???

3) I was playing a digital recording made from FM, and when the announcer speaks there's either horrible POP in the recording, or somehow my adjustments are causing POP.  Or maybe lack of adjustments or additional dEQ's are causing POP.  I'm planning to extract the parts of the program that cause the POP and use them as test material.

4) Would it be better to equalize in the bass or highs using 1/3 octave graphic rather than parametric?  Is there an advantage is using regularly spaced boost/cut as opposed to what just seems to work best?

5) Visually, there's an apparent depression around 500 Hz.  That's one of the bigger flaws in the response at this time, and being closer to the midrange it might be more important.  There's also an apparent depression around 1.5 Hz that is unintended, probably also of considerable importance.

6) I wonder if the famous Room EQ program REQ might come up with a better set of bass correcting curves.  My previous experience with REQ was disappointing.  I felt it didn't probe deep enough in time to find the really bad room nodes.  My feeling is that a short stimulus is simply not going to do this.  A very slow sweep would be best, and no program that I know uses more than a fairly short sweep.  But given the popularity of REQ, perhaps I am wrong.

I think one of the previous problems with moving the EQ's out of the chairside unit was that I didn't apply equally to both channels, and/or to the subwoofer AND the electrostat EQ's.