Seperate drivers for LF and ULF or one driver? - Page 11

"...where I seem at odds with the theories being thrown around in this thread, is that I don't believe the loading is variably proportional to wavelength."

you are at odds with much more than that. :-)

as far as the claim that loading is not variably proportional to wavelength, there is plenty of data to suggest that it is. as you are an empiricist, the simplest experiment is to plop a sealed sub in a car, then in a home room, and then outdoors. choose some frequency below the modal region not influenced by boundary resonances and it is clear that the smaller that you go, the higher in frequency where the "gain" kicks in, ergo loading is variably proportional to wavelength.

You're confusing several issues.

I'm not saying that loading isn't different in different environments, I'm saying that loading isn't different across the bandwidth in one environment.

Q: If the pressure goes up and down everywhere in the room below the modal region, what happens to that theory when the mic is placed in the dormer chute?

Let me get this straight - are you actually suggesting that acoustic space is a good environment for studying pressure gain effects?

I see that this subject is still alive, albeit somewhat bloody and maimed at this point...





Gotcha...

All of the measurements that I have previously posted in this thread should be without adjusting the levels relative to each other.

FWIW my new space is 8.5 x 14.5 x 25.5 or 3,150 cubes versus about 4,100 at the old place due to a shorter ceiling height and a narrower width. Plus it is a cement slab with three brick exterior walls including a brick fireplace and a wide hall way off of the front left corner area. There is one small window and a pair of double doors out to the back patio at the back of the room. I expect it to behave very differently from the old room below 100Hz. Whenever I get a chance to get some measurements in there I will post them.

Ill be very interested with your measurements Josh, as I have a similar type room and wall situation. I'd love to see if your response matches mine in some areas, but I honestly will hope you have a little better luck than I have

I feel like Chief Brody slinging chum over the side of the Orca, but oh well!

http://www.powersoft-audio.com/en/products/amp-modules/ipal.html

Pressure sensors, feedback loops, zero lag DSPs...there must be SOMETHING in that white paper that applies here.

I'm really interested in this as well and have been doing some reading on it when I can. The white papers share some comments and various ideas behind it in there. The patents are good reading as well. One thing I do know is that the Powersoft amps are the most amazing amplifiers that I have ever used. I really want to get one of the B&C Ipal 21's developed for this. I love the 21sw152's drivers they do just about everything right : Uber strong motor, crazy thermal power handling, extremely low distortion, rugged, very low inductance, smooth top end...They just don't have the raw displacement I like for the sub bass. The IPAL version has a voice coil with >50% longer wind and even more thermal mass and the geometric xmax calculation is 15mm vs 9mm which is a 67% increase. The BL^2/RE goes up from an already ridiculous 311 to 521 which is the most of any production driver that I can think of. Energy conversion levels should be tremendous compared with most other drivers. Part of the IPAL system is to shape the drivers parameters to fit an ideal model that works with your enclosure design. That is why the driver has to be so tremendously powerful and overbuilt. Really cool stuff. I just can't justify the expense for the driver at this time and the amp module would be very expensive as well. I'm not sure whether the amps I have will deal with the 1 ohm minimum load at full tilt.

Took some very rough first measurements last night. Still have a lot of work to do to even get things to the point that I can watch a movie but...Some observations that I found.

The new room seems to have a LOT more gain. I don't have any proof I'm just going with my gut and the drive levels producing powerful bass already. The drivers are barely even moving at all and it is room filling loud. In retrospect 8 drivers may be completely overkill in the new space.

I had a vicious peak at 46Hz in the old room. I apparently will have one very similar at about 42Hz in this one. Both rooms were longer than they are wide and the listening position will end up at around the halfway point of the length near the center of the room. Very similar seating placement in both. Looks like this peak is caused by the half wave corresponding to the increased length of the room as the ceiling height is reduced and the width is greatly reduced as well yet the peak has dropped a few Hz in frequency.

I do not have a notch in the bass below 20Hz. No evidence of anything like that in the new room.

I initially placed two subs in the front of the room, left and right and one sub in the back right corner. I have some cabinets in the back left side of the room that preclude placement there so I decided to try the 4th sub placed midwall on the left side. After playing with a few different configurations of different subs, it is apparent that running the left side mid wall sub causes a major cancellation or reduction of energy below 20Hz all of the way down into the ULF range when measured in the middle of the room, no matter which other combo of subs it is running with. I never saw this dramatic of an effect in any of my tests in the old room regardless of microphone or sub placement. I am about 97% sure that the polarity is matching the rest of the subs and that the drivers are wired in phase in that cab but there is a possibility that this is the cause. I'm going to double check that and check the cable wiring at the sub and amp. If all of that is good then the placement of that sub is having a huge negative effect on the bass under 20Hz in the middle of the room. This sub is about 6 or 7 feet from the microphone where the others are roughly 14ft or so. I believe that one measurement showed a drop of almost 9dB in the ULF just by plugging in that one sub. If it is the placement causing this and not a wiring or polarity snafu I am going to stack it in the back right corner with the other rear sub and call it good.
Edited by Ricci - 8/28/12 at 9:43am

Josh, when I am tweaking this week, will post some bare graphs of the RE's native response in several positions. They are on my PC at home so I just need to dig them up.

If you do make sure that you note what you have going on for the measurements like subs + mains or subs alone, with or without a low pass from your reciever, if so what frequency, is there Audyssey or other EQ involved? Mic position in the room, sub placement, room size and shape, etc. If you ever do the measurements again try to get the subs as raw as possible, no low pass or as high as possible, no EQ, etc. It's better to start from that as it makes it more difficult to tell what is the room and what is the signal manipulation if that other stuff is involved.

ive got all that, from previous measurements, but im going to be doing a whole new slew of measurements where I can really dial it down. Right now I have aud, no aud, aud+DCX, etc. but ill try and get the stuff all posted in my thread soon as I have more to post.

Resurrecting this...


After spending many hours over the weekend working on things it turns out that I did have the SW cab that was placed mid wall wired incorrectly after all. Instead of +1 and -1 terminals for driver #1 and the two #2 terminals for the second driver, I inadvertently had +1 of driver #1 wired with the negative of the second driver etc in this cab. Somehow the amp still ran it but it caused a major response cancellation. Went through and checked everything with a 9 volt battery to ferret this out. In the meanwhile I also discovered that I had an intermittent short developing in the front right SW where the cables all connect at the speakon terminal, two of the connectors were just barely touching. The amplifiers ran through that as well without issue and the drivers are wired in parallel for 2 ohms each cab. Fixed that as well. After fixing the mis-wired cab there is no longer a ULF cancellation when adding in this subwoofer with the others.

Here is what starting with one sub and adding an additional one for each measurement until all 4 are operating looks like at the main position. Note the gigantic 43Hz peak. This is VERY similar to what i dealt with before. Placement of the LP in the new place is similar to what i had before so this is unsurprising.





After EQing down the mountain of energy buildup at 45Hz I'm left with this before versus after response at the main position. Not too bad.




Here is how the average response from 6 different listening positions changes after Eq.






I'm not done yet and I've only watched a couple of movies so far but there is much less tactile feeling in this room than my old one. Especially at very low freq's Overall the impression is that the bass is a little drier with less overhang but the low stuff doesn't have the sense of power it did when I could feel the individual pulses through the structure of the old room. The low bass is still there as the measurements show but is much more subtle compared with the upper bass now. I may have to start watching movies at a slightly louder playback level to get some of the lost feeling back. I'm going to miss my suspended floor I think. Anyway if there are any measurements in particular that anyone wants to see, let me know.

So this is a new room (new house?) now? I see that hole in the lower teens is now gone completely. Woot! Sounds like you're experiencing some of what I have got going in my room but I have quite a null from 40-60hz.

What sweep length and window width are you using for these plots?

What are the room dimensions vs. seating location?

Can you overlay each location's response curve along with the summed response?

Glad you got some of the problems worked out Ricci.

Suspended floors are unfortunately like a peaky freq response, even though they are fun. I have a concrete slab, but my main LP is a futon that is suspended at the ends. 17-22Hz energy makes it move quite a lot. Sounds like you should look at Crowson's tech and build a riser suspended on the linear actuators to get all the old fun plus more....they are pretty pricey, though.

JSS

This ^ ^ ^ ^ ^

Yes...Completely different place. No hole below 20Hz. Actually no bad nulls to speak of until up past 70Hz. Looks like I have some issues at 73, 118, 150, 230Hz though.





For bass measurements I typically use the maximum 1M (1024K) sweep length in REW, a 48khz sampling rate and a range of either 0-120Hz, 0-200Hz or 0-300Hz depending on what I am looking at. I usually leave the window for bass measurements set like this. Tukey 0.25 for both windows with a 125ms open and 500ms close. I can increase the close to the maximum of >1700ms and it does change the resollution in the ULF range a little but it has very little effect until below 8Hz so I usually just leave it set to 500ms. A comparison of that is below.

Hey Josh thanks for sharing


What type, if any, and how much bass trapping do you have employed?

Can you explain, with as much detail as you care to offer, how all your boundaries are constructed? (ie., typical stud, sheetrock, etc, first floor of multi story, main-floor of ranch, basement, any walls have perpendicular bracing from adjacent rooms, etc.) I know, ate up

Any time domain measurements? Klipsch mains, right?

Thanks

Space is 8.5 x 14.5 x 25.5' or 3,150 cubes. Cement slab with three brick exterior walls including a brick fireplace and a wide hall way off of the front left corner area that opens up to a stairwell. There is no door at the top of the stairwell or in the entryway to the room. There is one small window in the back right and a pair of double doors out to the back patio off the back of the room. There is also a bar area with cabinets in the back left of the room. A very rough room sketch and layout is below. The scale of some things is off but you get the idea. The seating position is slightly over half way back into the room. Not ideal but it's what is available to work with.






Here are the measurements. Each of the 4 SW's are shown and numbered in the measurements and sketch posted above, as is listening position #1 where they were taken. These are just the measurements of each of the individual subs raw and all 4 together with no processing other than a 200Hz low pass from the pre-pro. The electrical signal is thrown in as well. I assume this is what you wanted.

FOH,

No treatments or bass traps of any kind. I am way way off from getting into that. There is still a ton left to do to even get the room fully together, hiding wires and lots of equipment tweaking. I have to work on my gain stage, making a mount for the CC, new mains all around are coming at some point, etc...

I already went through the room construction a couple of times.It's in the last 5 or 6 of my posts. I justed wanted to add that this is never going to be a dedicated, blacked out, highly modified and treated theater room. That is what I wanted originally but the space I ended up with just does not lend itself to it. It's more of a social entertainment area and functional living room space and will always be rather humble compared to what some of the guys here have. To tell you the truth that is more my style anyway.
Edited by Ricci - 9/5/12 at 9:21am

Cool, I know the feeling, my room's in transition as well.

Well, if you capture a time domain plot now, then you can do a comparo before and after wrt LF damping. Just took mine out to perform some before tests, so it was on my mind.


Thanks

All the right info.

I would experiment with adding some delay to sub 3 & 4, maybe individually, probably combined. Ideally you would want to find some adjustment that will pull down the 43Hz peak seen in the summation, while not creating any big holes elsewhere. I would suggest trying 2ms increments up to no more than 10ms while watching the <20Hz summation and any holes created in the passband.

There might be too many other casualties with that adjustment before you knock down the peak much, but even 6dB off the peak reduction would be very beneficial.

If straight delay causes too many issues above 50Hz, try using a low pass filter on the rear 2 subwoofers, starting with 2nd order BW. Lowering the low pass frequency adds delay, as does going to a steeper slope. You can even use a combination of low pass and straight delay.

Finally, I'd experiment with a little shape to the sub frequency response dialing in a shelf which rises 3-6dB from the 60-100Hz range down into the 25-35Hz range, flattening again below there. Obvious stuff, but that should help greatly with the perception. It might take even a bit more tilt. Note that the Harman guys involved with the Synthesis systems had found VLF decay times (related to room size and construction) played greatly into equal perceived bass levels. They found as much as a 10dB frequency response variation in what sounded similar to listeners in different spaces. Knowing that our hearing mechanism is power related, not just intensity, so understandably, duration matters.

You crammed a lot in there. I will see if I can play around with some of it when I get a chance. Adding an extra goose down into the deep bass shouldn't be a problem. I have the headroom for it.

About adding delays to the rear subs to try and reduce the energy buildup... Why not just cut it back with EQ and leave the headroom like I'm currently doing? I suspect that if I delay the rear subs it may cause further problems elsewhere in the bass. What is the advantage of the delay method over plain old EQ? Despite the size of the peak I don't recall it ringing in the group delay or waterfall plots. It seems to decay rapidly. My problems with ringing are close to 73 and 118Hz if I recall. I will have to double check.

Another thing that is probably useful to know is that I have the drivers crammed in a much smaller airspace than ideal, about 4.5cubes per driver, so they are exhibiting a peaky response that probably centers near to 40Hz. It is not all room in this case. Part of of that peak could just be the raw response of the subs.

Good and important to note the driver response possibility. I would suggest a 1/2" measurement from one of the dustcaps and first address any peaking in the sealed box response, as you know this will be consistent through the room and this can be precisely EQ'd in a manner that won't change with listener/measurement location. That would give you a better idea of what is being contributed by the room/locations and give an idea of what range of adjustment might be possible.

I'd later try and take a look at the decay around that peak in the room, and measure at a few more locations around to see how strong it remains. While EQ will pull the response down, and the added efficiency is generally welcome, I'd liken it to an athlete putting heavy pads over a bruise. In a simple straight on test you might not feel much, but if you get hit or move in a particular way, you will still feel the bruise. That strong peak will amplify any signals at those frequencies, distortion included. There is also a consideration for the average response over a general listening area, where just as with a peaky outdoor subwoofer response which is smoothed only at the measurement positions, you will typically still hear the underlying peaks in real use.

Some will assert that I'm full of it and it doesn't matter where you smooth the response, just that it does get smoothed at the listening position, but that doesn't correlate to my own observations and the realities of a finite number of measurement locations. My 2c.

Ill agree with you Mark, as eqing to a flat response AT the speaker is obviously paramount to whatever the room does to the response by the time it reaches the listener. Obviously if there is already a peaky response, this would in most cases just get worse with distance. With so many rooms just having such a crazy response between the listening positions, I feel unfortunately that often, they may start with a peaky response at the speaker, and from that point it is just damage control.

I think delaying some of the subs adds another variable and the possibility for improvement, akin to changing room dimensions and/or distance between subs.

Some valid points but by adding delay to help cancel out the peak at 40Hz I have to wonder how much would be needed to make a notable difference. A 40Hz cycle is about 25-30ms in duration. 10-15ms would be as much as needed at maximum but what will this do in other parts of the FR and the phase relationship? Obviously the left sub on the mid wall is physically closer to the majority of the listening area so it will arrive a little sooner at the ear than the rest. I may mess with delaying that sub first to see what it does relative to the others.

I care about what is delivered to the listening area not what happens right at the speaker because that is not where I listen to them from. That is why I don't bother to EQ for a perfectly flat response at the speaker first. It just seems like a wasted step to me since your room may be counteracting those response features and/or exacerbating them. Either way you may undo what you spent time dialing in at the speaker. If you are selling a finished good with no way of knowing what room it will end up in then yeah I would tailor the speakers response. When I say that the response may be a little peaky I don't want it to get blown out of proportion we are talking maybe 6dB at 40Hz relative to 100Hz at most. So there is still some 10dB or so attributable to the LP and subs placement. My main speakers which are placed on top of the front left and right subs also exhibit a big 40Hz peak at the listening area when measured alone and they actually have a shallow roll off down towards 30Hz.

I do have the same measurements that were posted above from each of the other listening locations as well. Mark has a good point that all of that boost at 40-45Hz will pump up any distortion harmonics though so H2 of 20Hz and H3 of 15Hz will be boosted up what looks like 15dB or so.

Ricci,

I know this may be very simplistic of me, but looking at your room, I come to the following conclusion:

Room:


Your room mode profile:


You can see that your 2nd length mode is at 44Hz, and cannot be canceled by sub placement at the opposite ends of the room. You have a problem at 42Hz, I call that close enough. If you have a 42 Hz problem at seats 1,2 and 3, that is the problem; the 2nd order length mode. Your room also has a tangential mode at 44Hz per the calculations, harder to mess with. Unless sub placement is negotiable (1/4 the room dimensions along the sidewalls), you just have to try using delays or just EQ to get that peak out.

I know some people will throw out the fact that there are passageways connected to the room that will make the theoretical mode study above inaccurate, and they are correct. But I have also found that up to 3rd order modes in smaller rooms bear out in rooms that are mostly closed off, like yours.

SW3 in your setup is the only one that may be close to not exciting the 42Hz mode (near the 1/4 length dimension), and lo and behold, it is ~6dB down compared to the others. That is proof enough for me that you have a 2nd order length mode problem at 42Hz, that can be fixed with EQ for seats 1,2,3. Other seats may see a dip at 42Hz, but you can only do so much if sub and seating positions are limited.

Hope this helps,

JSS

Interesting discussion.

I'm a strong advocate of optimization in the time domain, ..bearing fruit in the frequency domain.

There's so much emphasis in analyzing and manipulating the freq domain, however typically things tend to clean up nicely (both objectively and subjectively) in freq by pursuing signal alignment in the time domain.

IMO, It's all about the room, the room's all about the time domain. Sure, that's way over simplified and somewhat exaggerated



Thanks