I happened to come across an RCA Y adapter laying on top of the living room power conditioner as I was cleaning the right subwoofer and environs on Tuesday. I was carefully vacuuming cobwebs and dust bunnies from between the woofer and the metal subwoofer grille, then I used a dry microfiber cloth to lightly brush off the remaining dust from the SVS woofer cone which appears to be covered in a felt-like substance (I think it might be a metal cone also). I'm not sure if water could be used for this cleaning or not, or even if microfiber is all that safe, but I hoped it would be ok for a light dusting. I was puzzling the 120 Hz "notch" which actually seemed to have a strong buzzing quality, even more like a strong modal resonance in the alcove between the back of the right Acoustat 2+2 and the right SVS PB13 Ultra. I was also neatening the junk on top of the power conditioner in that area. My cleaning did seem to help the buzzing, actually, but not eliminate it, and I think I'm going to have to remove the now heavily damped metal fireplace and rebuild the front wall to fix it, something I've been thinking about doing for years (but where to move the stereo in the meantime??? The organ and junk had to be removed from the Gym, to permit the Gym to be used as temporary holding space for the living room stereo. The old organ and junk were finally removed last year, after 6 years of procrastination.)
I must have put that Y adapter there precisely for use with the oscillator, for stereo sweeping, which I had not done yet. Ahah, one more kind of sweeping I can do! I figured right away that as soon as I started stereo sweeping I'd identify issues that required mono sweeping again. I was right.
Immediately I was back examining the 120 Hz suckout in the right channel. It was quite noticeable in stereo (as in the stereo correlated pink noise response). But then I also noticed that the bass was seeming quite light. Had the subwoofers died? Ooops, I had the subwoofer EQ turned to a different input, which is my tricky way of turning the subs off for testing the panels by themselves, which I must have been doing the day before. But since then I'd spent a day or so listening to Marie-Claire Allain play the Complete Bach Organ Works, in background mostly, and marveling how good it sounded. With no subwoofer and the panels rolled off at 100 Hz. We audiophiles make these mistakes all the time, and in my case I'm all for pointing it out how unreliable our hearing is that major issues can go unnoticed for days (weeks, years, or even decades, I can show in my own amazing audiophile history), while we nit pick over small stuff that may be essentially impossible to hear.
With the subwoofers back on, it was now clear that the subs didn't suck out as much as the panels. Something weird seems to be going on in and in back of the right panel at 119 Hz. It's buzzy and loud, and apparently phase reversed from the subs. If the subs are on without the panels, and then I turn the panels on, the level of 120 Hz at the listening position goes down to nothing, over 20dB down. Maybe 30dB. Now that's cancellation!
I could (and did try) to fix the problem by boosting the subs at 119 Hz. But that didn't go so far, and made nodal resonances at other points in the room, such as in the doorway to the kitchen (where 119 Hz gets VERY loud) even louder, by the same amount as the boost (or seemingly more).
But what actually seemed MORE effective, was notching out the panels at 119 Hz. And curiously the more I notched out 119 Hz in the panels the louder it became at the listening position, apparently because of reducing the cancellation of the more effective subwoofer drive. It also seemed to reduce the buzzing, as if I was suppressing an anti-phase resonance in the panels (that's not my theory, exactly, but there may be some issue like that involved).
Dialing in the full -15dB available in a 1/10 octave notch at 119 Hz reduced the 30dB of cancellation to a mere 3dB of cancellation. And it didn't seem noticeable when sweeping the speaker. Weird stuff goes on at 119 Hz in any case, but with the -15dB notch it sounded only better than at lesser levels, not worse.
Still, I'm disinclined to apply such radical EQ. I looked for the level where I'd get "most of the benefit" without having to go all the way. And I found I could get within 3dB of the best available performance (at -15dB) with a mere -8dB, so I decided to go with that.
This is very strange. I'm fixing a suckout at 119 Hz in the right channel by appling a notch of -8dB at 119 Hz to the right channel panels. Strange but it seems to work and I get even better elimination of the notch with -15dB.
There is obviously something (or many somethings) very weird going on here, possibly including the acoustic crossover not matching Linkwitz-Riley LR4 alignment very well because of driver resonances AND other EQ corrections! Also imperfect time alignment.
LR4 itself is a little weird. It's not as "perfect" as LR2 perhaps, though it doesn't require polarity inversion on one side. LR4 is strange because you can use it either with or without polarity inversion. I think I like it better without the inversion, and it's one of the reasons I like LR4 (no polarity inversion required). But Linkwitz liked it better WITH the polarity inversion, because it has less group delay that way. Hmmn. I have always wondered if no-inversion LR4 has regions where the drivers are not always in-polarity as they are with LR2 at ALL Frequencies. At distant extremes, the drivers are in polarity, but near the crossover the relative phase twists around in opposite directions for each driver, further complicated by non-coincidence and non-uniformity-of-radiation.
These issues are worthy of serious investigation! However, for now, I'm using the -8dB notch on the panels, which increases the near total suckout at 119 Hz at the listening position to a fluctuation.
Sadly, I do not believe I can add this -8dB notch merely to the right panels in my current configuration, which is stereo-linked for the panels. I'm not sure I need to do stereo-linked anyway, but it has always seemed convenient for adjusting the panel issues, which aren't as position dependant. There had not appeared to be a serious suckout in the left channel at 119 Hz that I remembered. For now, I swept the left channel with the -8dB notch in place and surprisingly it sounded fine. Perhaps the notch makes it better too. Obviously I should re-check this also.
And I should remember I'm applying such a hack and re-check if it continues to help after future changes and adjustments.
I ought to try LR4 with polarity inversion. I don't recall even trying it, I just didn't like the idea of polarity inversion. But it seems to me now that only LR4 with polarity inversion yields the central benefit as LR2 which must have a polarity inverted...both sides remain in polarity at all times. Without that critical inversion, the relative phases must be twisting around, and frankly right now I can't even fathom how it's even supposed to add up to flat response.
Perhaps much of the last 10 years of fiddling with EQ's around 100 Hz is largely to compensate for using a defective crossover alignment...
Using the Y adapter is nice, because I can switch channels using the remote (through the Tact Level Menu, reducing alternate channels to -99.9db) and compare channels. It was clear the right channel was somewhat lacking around 32 Hz, as I remembered. All the boost shown in the stereo response graphs was from the left channel response nearly exploding around 32 Hz and down to below 20 Hz, and not sounding all that good either. I dialed in a -4dB 3/4 octave valley around 28 Hz to tamp this down. It could be tamped down more. This allowed me to reduce the notch at 40 Hz to -5dB.
More issues in this channel as well. Is the subwoofer tuning EQ correct for the current one-port-open? Is one-port-open a good idea in this channel? I already went to sealed in the other channel, but I sorta liked the combination of the two in different channels. I thought I liked the left channel often filling in extra whoomph otherwise unavailable from the sealed configuration. Was I mistaken about the effect of that overall?
I'm noticing a lot more whoomp in recordings like Grouse We want to be loved. How much this is due to the very last change, the 119 Hz notch which seems to boost overall response, I don't actually know, but it made me decide to increase that notch to -12dB (I know it continues to measure better, in the right channel at least, all the way out to -15dB, and that gives the flattest listening position response to 119 Hz, only about 3dB below baseline, which is what you get cancelling out one part of the signal--the panels--and leave the remainder).
In addition to a curious property of the in-polarity LR4 connection, another concern is the damping in the back of the Acoutstat panels, which sometimes becomes loose with age. That could be the very element which begins to vibrate out-of-polarity at 119 Hz because of a resonance at precisely that frequency in the panels themselves. Chalk this up as one prototype theory in which a panel notch at 120 Hz improves the response greatly at that frequency, at least in the right channel.
It may in fact be that resonance at 119 Hz and/or others is partly reponsible for the rattling/grating side I hear mostly in the back of the panel on certain pieces of music, notably a track from Grouse. I can eliminate that rattling with a wide notch of 1 octave centered at 250 Hz, or better yet with -15dB sliders from 120Hz to 630Hz, possibly with some middle sliders back at zero. The rattling is inconsistent and history dependent, making it hard to figure exactly. But some reduction at 120 Hz always seems necessary, and I'm now already reducing the panel response by -12dB by default, which may reduct the rattling somewhat by default.
Sadly oscillator sweeping has not yet revealed a single frequency for the rattling, which sounds like the rattling of stator wires, possibly because an attachment has come loose.
I must have put that Y adapter there precisely for use with the oscillator, for stereo sweeping, which I had not done yet. Ahah, one more kind of sweeping I can do! I figured right away that as soon as I started stereo sweeping I'd identify issues that required mono sweeping again. I was right.
Immediately I was back examining the 120 Hz suckout in the right channel. It was quite noticeable in stereo (as in the stereo correlated pink noise response). But then I also noticed that the bass was seeming quite light. Had the subwoofers died? Ooops, I had the subwoofer EQ turned to a different input, which is my tricky way of turning the subs off for testing the panels by themselves, which I must have been doing the day before. But since then I'd spent a day or so listening to Marie-Claire Allain play the Complete Bach Organ Works, in background mostly, and marveling how good it sounded. With no subwoofer and the panels rolled off at 100 Hz. We audiophiles make these mistakes all the time, and in my case I'm all for pointing it out how unreliable our hearing is that major issues can go unnoticed for days (weeks, years, or even decades, I can show in my own amazing audiophile history), while we nit pick over small stuff that may be essentially impossible to hear.
With the subwoofers back on, it was now clear that the subs didn't suck out as much as the panels. Something weird seems to be going on in and in back of the right panel at 119 Hz. It's buzzy and loud, and apparently phase reversed from the subs. If the subs are on without the panels, and then I turn the panels on, the level of 120 Hz at the listening position goes down to nothing, over 20dB down. Maybe 30dB. Now that's cancellation!
I could (and did try) to fix the problem by boosting the subs at 119 Hz. But that didn't go so far, and made nodal resonances at other points in the room, such as in the doorway to the kitchen (where 119 Hz gets VERY loud) even louder, by the same amount as the boost (or seemingly more).
But what actually seemed MORE effective, was notching out the panels at 119 Hz. And curiously the more I notched out 119 Hz in the panels the louder it became at the listening position, apparently because of reducing the cancellation of the more effective subwoofer drive. It also seemed to reduce the buzzing, as if I was suppressing an anti-phase resonance in the panels (that's not my theory, exactly, but there may be some issue like that involved).
Dialing in the full -15dB available in a 1/10 octave notch at 119 Hz reduced the 30dB of cancellation to a mere 3dB of cancellation. And it didn't seem noticeable when sweeping the speaker. Weird stuff goes on at 119 Hz in any case, but with the -15dB notch it sounded only better than at lesser levels, not worse.
Still, I'm disinclined to apply such radical EQ. I looked for the level where I'd get "most of the benefit" without having to go all the way. And I found I could get within 3dB of the best available performance (at -15dB) with a mere -8dB, so I decided to go with that.
This is very strange. I'm fixing a suckout at 119 Hz in the right channel by appling a notch of -8dB at 119 Hz to the right channel panels. Strange but it seems to work and I get even better elimination of the notch with -15dB.
There is obviously something (or many somethings) very weird going on here, possibly including the acoustic crossover not matching Linkwitz-Riley LR4 alignment very well because of driver resonances AND other EQ corrections! Also imperfect time alignment.
LR4 itself is a little weird. It's not as "perfect" as LR2 perhaps, though it doesn't require polarity inversion on one side. LR4 is strange because you can use it either with or without polarity inversion. I think I like it better without the inversion, and it's one of the reasons I like LR4 (no polarity inversion required). But Linkwitz liked it better WITH the polarity inversion, because it has less group delay that way. Hmmn. I have always wondered if no-inversion LR4 has regions where the drivers are not always in-polarity as they are with LR2 at ALL Frequencies. At distant extremes, the drivers are in polarity, but near the crossover the relative phase twists around in opposite directions for each driver, further complicated by non-coincidence and non-uniformity-of-radiation.
These issues are worthy of serious investigation! However, for now, I'm using the -8dB notch on the panels, which increases the near total suckout at 119 Hz at the listening position to a fluctuation.
Sadly, I do not believe I can add this -8dB notch merely to the right panels in my current configuration, which is stereo-linked for the panels. I'm not sure I need to do stereo-linked anyway, but it has always seemed convenient for adjusting the panel issues, which aren't as position dependant. There had not appeared to be a serious suckout in the left channel at 119 Hz that I remembered. For now, I swept the left channel with the -8dB notch in place and surprisingly it sounded fine. Perhaps the notch makes it better too. Obviously I should re-check this also.
And I should remember I'm applying such a hack and re-check if it continues to help after future changes and adjustments.
I ought to try LR4 with polarity inversion. I don't recall even trying it, I just didn't like the idea of polarity inversion. But it seems to me now that only LR4 with polarity inversion yields the central benefit as LR2 which must have a polarity inverted...both sides remain in polarity at all times. Without that critical inversion, the relative phases must be twisting around, and frankly right now I can't even fathom how it's even supposed to add up to flat response.
Perhaps much of the last 10 years of fiddling with EQ's around 100 Hz is largely to compensate for using a defective crossover alignment...
Using the Y adapter is nice, because I can switch channels using the remote (through the Tact Level Menu, reducing alternate channels to -99.9db) and compare channels. It was clear the right channel was somewhat lacking around 32 Hz, as I remembered. All the boost shown in the stereo response graphs was from the left channel response nearly exploding around 32 Hz and down to below 20 Hz, and not sounding all that good either. I dialed in a -4dB 3/4 octave valley around 28 Hz to tamp this down. It could be tamped down more. This allowed me to reduce the notch at 40 Hz to -5dB.
More issues in this channel as well. Is the subwoofer tuning EQ correct for the current one-port-open? Is one-port-open a good idea in this channel? I already went to sealed in the other channel, but I sorta liked the combination of the two in different channels. I thought I liked the left channel often filling in extra whoomph otherwise unavailable from the sealed configuration. Was I mistaken about the effect of that overall?
I'm noticing a lot more whoomp in recordings like Grouse We want to be loved. How much this is due to the very last change, the 119 Hz notch which seems to boost overall response, I don't actually know, but it made me decide to increase that notch to -12dB (I know it continues to measure better, in the right channel at least, all the way out to -15dB, and that gives the flattest listening position response to 119 Hz, only about 3dB below baseline, which is what you get cancelling out one part of the signal--the panels--and leave the remainder).
In addition to a curious property of the in-polarity LR4 connection, another concern is the damping in the back of the Acoutstat panels, which sometimes becomes loose with age. That could be the very element which begins to vibrate out-of-polarity at 119 Hz because of a resonance at precisely that frequency in the panels themselves. Chalk this up as one prototype theory in which a panel notch at 120 Hz improves the response greatly at that frequency, at least in the right channel.
It may in fact be that resonance at 119 Hz and/or others is partly reponsible for the rattling/grating side I hear mostly in the back of the panel on certain pieces of music, notably a track from Grouse. I can eliminate that rattling with a wide notch of 1 octave centered at 250 Hz, or better yet with -15dB sliders from 120Hz to 630Hz, possibly with some middle sliders back at zero. The rattling is inconsistent and history dependent, making it hard to figure exactly. But some reduction at 120 Hz always seems necessary, and I'm now already reducing the panel response by -12dB by default, which may reduct the rattling somewhat by default.
Sadly oscillator sweeping has not yet revealed a single frequency for the rattling, which sounds like the rattling of stator wires, possibly because an attachment has come loose.
No comments:
Post a Comment