|
|
Post by msimanyi on Feb 21, 2022 14:08:04 GMT -5
I found a very good video for "aligning" multiple subs together using REW. For "normal" setups with multiple subs connected to a miniDSP, one input from the processor, multiple outputs to all the subs, the procedure is very well explained. It looks to be quite a bit easier than how I've been doing things, so I'll try it out this weekend. I finally got around to setting up my miniDSP and REW, and I started with this version of aligning my three subs. I'm not done yet - haven't run Dirac after the sub alignment - but I decided to test the estimated summed combinations against the actual. I won't say they were perfect, but they were reasonably close. (I tested the first two summed against their actual, as well as the complete summed against actual.) The beauty of the process shown in the video is it's easy to understand and implement, even for a newbie like me. And it's effective. I'm glad ttocs posted it. Next up, Dirac. Then my big move will likely be gathering all the samples for Multi-Sub Optimizer later this week, and re-running Dirac after that optimization. This is all still a learning process for me, so while it's a little tedious, it's also quite interesting to see what happens. Question: I realize now that I may have messed up. I'm feeding a single sub-out to the miniDSP, but I used the center sub and have it set to Mono. I *think* I should have used the left sub out, set to Mono. Can anyone tell me definitively if that makes a difference? |
|
ttocs
Global Moderator 
I always have a wonderful time, wherever I am, whomever I'm with. (Elwood P Dowd)
Posts: 8,162
|
Post by ttocs on Feb 21, 2022 14:37:09 GMT -5
I found a very good video for "aligning" multiple subs together using REW. For "normal" setups with multiple subs connected to a miniDSP, one input from the processor, multiple outputs to all the subs, the procedure is very well explained. It looks to be quite a bit easier than how I've been doing things, so I'll try it out this weekend. I finally got around to setting up my miniDSP and REW, and I started with this version of aligning my three subs. I'm not done yet - haven't run Dirac after the sub alignment - but I decided to test the estimated summed combinations against the actual. I won't say they were perfect, but they were reasonably close. (I tested the first two summed against their actual, as well as the complete summed against actual.) Next up, Dirac. Then my big move will likely be gathering all the samples for Multi-Sub Optimizer later this week, and re-running Dirac after that optimization. This is all still a learning process for me, so while it's a little tedious, it's also quite interesting to see what happens. Question: I realize now that I may have messed up. I'm feeding a single sub-out to the miniDSP, but I used the center sub and have it set to Mono. I *think* I should have used the left sub out, set to Mono. Can anyone tell me definitively if that makes a difference? Doesn't matter. The settings in the subs and miniDSP are what matter. To get more flexibility, connect the Left Sub Output and Center Sub Output to miniDSP Inputs 1 and 2. It's just one extra cable. Then, when using correction software, both sub outputs will each get a correction and you can later choose to use both or only one, like, Left Sub Mono and Center Sub LFE, or, just Left Sub Mono. This doesn't give you more subwoofers, but it does give you more channels to use for specific reasons. |
|
|
|
Post by leonski on Feb 21, 2022 14:54:47 GMT -5
45hz has what wavelength? Maybe 23 or 24 feet? A single room dimension or combination which measures that can result in the condition you note.
As for bass traps?
I visitied a friend, along with a couple other persons, who said he had a bass problem. We TOOK TURNS standing in corners. In every case the listeners noted an improvement in Bass.
Corners appear to be where the bass 'loads up' so a little Fat-Mass helped. I have designs for corner traps made from 4" thick OC703 or similar which will have the same effect.....
'Sub Match' for Magges CAN be a PIA. The front / back radiation results in difficulty with sub phase while the inherent quickness of hte panel means SOME subs can sound ...maybe.....sluggish?
As for WHERE to place multiple subs? Even numbers like 1/2 or 1/4 are, IMO to be avoided. I'd instead try a 62/38 split which is PHI. This is 1.618 and the 'magic number' of the universe.
|
|
|
|
Post by msimanyi on Feb 21, 2022 17:22:34 GMT -5
To get more flexibility, connect the Left Sub Output and Center Sub Output to miniDSP Inputs 1 and 2. It's just one extra cable. Then, when using correction software, both sub outputs will each get a correction and you can later choose to use both or only one, like, Left Sub Mono and Center Sub LFE, or, just Left Sub Mono. This doesn't give you more subwoofers, but it does give you more channels to use for specific reasons. So I have a couple questions about that. First, do you then "mix" them in the miniDSP unit so both are joined in the output to the subs, or do you maintain separate outputs (LFE feed out to LFE inputs, Mono feed to the general subwoofer inputs)? Second, does the mono/lfe output on the processor make Dirac treat that as two subs for their scans? It seems like it does, and there's an added complication with MSO in the mix. Perhaps I can do the MSO work with the single mono setup, and add the separate inputs after. Third, do you have any comments on the mono+lfe inputs approach versus setting my fronts to large, using speaker-level connections to the two front subs, and letting bass management feed to the large fronts, with LFE to all three subs? I can see how the latter loses my third (rear) subwoofer for bass management signals... |
|
|
|
Post by p4t on Feb 21, 2022 20:53:11 GMT -5
I did try that, also does not make any significant different. I pretty much try every location in the room for sub position, but always have dip in 45Hz region. I think it might be cause the floor to ceiling reflection, as I do not have any absortion at all on the ceiling. Walk around the room with a mic in your hand while playing a 45Hz tone with REW and watching the trace in REW's RTA. If you haven't done this, you might be surprised about what you see. I would try with all four subs, then one sub, then two, then move the third sub into a triangle config and measure three subs. All while seeing if 45Hz will raise up higher than now. Start at the MLP, then move the mic forward and back, then left and right of MLP. This will help to map in your mind where weak and strong spots are. You may find that the mic doesn't need to move very far. I will try it. Thank you for your suggestion. |
|
|
|
Post by p4t on Feb 21, 2022 20:55:54 GMT -5
This sub remind me when Ken Kreisel still in M&K. Once in 1997 I listened their top of the line sub backthen MX5000 and the sound just awesome. Now, their “design” is use by perlisten. And according the audioholics review, it is the most accurate subs they have tested so far. Home theater HiFi indicated the Kreisel’s were the most accurate, lowest distortion subs they had tested at the time of the review. They are exceptional with music and, for lack of a better term, they sound incredibly quick. I think they will be a good match with my new Magneplanar speakers. I did not know Perlisten used the same design. I must say Ken Kreisel is the most helpful person/manufacturer I have ever dealt with. I hope Emotiva comes through with Dirac’s bass control soon so I can maximize the DXD-12012’s. The best of all subwoofer world’s. Try to checkout sparechange channel on youtube. He just review perlisten sub 2 12” few days ago. |
|
|
|
Post by p4t on Feb 21, 2022 21:00:23 GMT -5
45hz has what wavelength? Maybe 23 or 24 feet? A single room dimension or combination which measures that can result in the condition you note. As for bass traps? I visitied a friend, along with a couple other persons, who said he had a bass problem. We TOOK TURNS standing in corners. In every case the listeners noted an improvement in Bass. Corners appear to be where the bass 'loads up' so a little Fat-Mass helped. I have designs for corner traps made from 4" thick OC703 or similar which will have the same effect..... 'Sub Match' for Magges CAN be a PIA. The front / back radiation results in difficulty with sub phase while the inherent quickness of hte panel means SOME subs can sound ...maybe.....sluggish? As for WHERE to place multiple subs? Even numbers like 1/2 or 1/4 are, IMO to be avoided. I'd instead try a 62/38 split which is PHI. This is 1.618 and the 'magic number' of the universe. I am sorry I do not understand about how to use 1.618 or 62/38 split. My room dimension is 18ft X 15ft, where should I place my sub using this method? Thanks. |
|
|
|
Post by leonski on Feb 21, 2022 22:05:45 GMT -5
Good Question, P4T. For completeness? How TALL is the room? I phrased is exactly and it is part test. The number? Known as the Golden Mean or Golden Section. But for audiophiles? A neat way to divide up space. The 62 / 38 part? Well......your long way in your room is 18 feet...... Multiply by .618 (more exact number) and you get 133.5 INCHES which is where to test your sub along the wall. Doing so will automatically SPLIT the wall in the ratio I list in my first post. If you are curious? Well, this is a VERY important number in history, going back about 5000 years or more....to the construction of the Great Pyramid and more recently in middle ages ART and today in the construction of the United Nations building in NYC. For Audiophiles? www.cardas.com/room_setup_fibonacci.phpAs it turns out, they use the Fibonacci Series which is another way to phrase the Golden Section. It turns into an EASY math problem. But if you are curious, I've given MORE than enough clues for you to get your feet wet to just above your KNEES. This math is good to use to design anything from buildings / to speaker enclosures...... |
|
ttocs
Global Moderator
I always have a wonderful time, wherever I am, whomever I'm with. (Elwood P Dowd)
Posts: 8,162
|
Post by ttocs on Feb 21, 2022 23:32:27 GMT -5
To get more flexibility, connect the Left Sub Output and Center Sub Output to miniDSP Inputs 1 and 2. It's just one extra cable. Then, when using correction software, both sub outputs will each get a correction and you can later choose to use both or only one, like, Left Sub Mono and Center Sub LFE, or, just Left Sub Mono. This doesn't give you more subwoofers, but it does give you more channels to use for specific reasons. So I have a couple questions about that. 1. First, do you then "mix" them in the miniDSP unit so both are joined in the output to the subs, or do you maintain separate outputs (LFE feed out to LFE inputs, Mono feed to the general subwoofer inputs)? 2. Second, does the mono/lfe output on the processor make Dirac treat that as two subs for their scans? It seems like it does, and there's an added complication with MSO in the mix. Perhaps I can do the MSO work with the single mono setup, and add the separate inputs after. 3. Third, do you have any comments on the mono+lfe inputs approach versus setting my fronts to large, using speaker-level connections to the two front subs, and letting bass management feed to the large fronts, with LFE to all three subs? I can see how the latter loses my third (rear) subwoofer for bass management signals... 1. Both Left Sub and Center Sub are input into the miniDSP, then three outputs from the miniDSP to each of the three subs. So yes, the BM and LFE are "mixed" in the miniDSP, then output together to the subs. 2. I'm not sure how MSO works. I don't currently have a Win computer in operation right now, so cannot try MSO. The Left Sub and Center Sub outputs are each seen by Dirac and each are corrected individually, but Dirac is unaware how many subwoofers are involved, and we don't care. The subwoofers are collectively treated as a single virtual subwoofer by Dirac. It's up to me to make sure I've done the best I can with aligning the subwoofers in the group, then let Dirac do what it does. 3. I was using the Large Fronts for BM for quite a while. It sounded very good. The sound quality was better than just using subwoofers setup as Mono and Dual Mono. It was around this time that I discovered that BM wasn't working like it should've been. I wasn't able to adjust BM separately when it was sent to Large Fronts, but I could adjust BM separately if I used the Left Sub Output as long as I was also using the Center Sub Output for only LFE. This developed into how my system is now configured. But the biggest improvement in sound quality and clarity was when I setup my Center Speaker as Large and used a dedicated subwoofer and active crossover. Being able to separate LFE into its own channel gives us a great amount of control and flexibility. |
|
|
|
Post by donh50 on Feb 21, 2022 23:42:16 GMT -5
MSO: www.andyc.diy-audio-engineering.org/mso/html/Starting from a mono sub feed, it optimizes delays and filters for two or more subwoofers. The usual implementation is a miniDSP driven by the AVR/AVP sub output, load MSO filter coefficients, and connect subs to the outputs of the miniDSP unit. HTH - Don |
|
|
|
Post by msimanyi on Feb 22, 2022 0:20:52 GMT -5
2. I'm not sure how MSO works. I don't currently have a Win computer in operation right now, so cannot try MSO. The Left Sub and Center Sub outputs are each seen by Dirac and each are corrected individually, but Dirac is unaware how many subwoofers are involved, and we don't care. The subwoofers are collectively treated as a single virtual subwoofer by Dirac. It's up to me to make sure I've done the best I can with aligning the subwoofers in the group, then let Dirac do what it does. I just had a (probably bad) thought. Seems to me that aligned subs are aligned subs. I don't know if MSO does more than that - I expect it does - but again, once they're optimized, they're optimized. If that's true, I can run MSO against the current configuration and use it to set the miniDSP, then separate the bass management and LFE signals into the miniDSP, then run Dirac. If anyone sees a flaw in that plan, please chime in. |
|
KeithL
Administrator
Posts: 10,273
|
Post by KeithL on Feb 22, 2022 10:50:41 GMT -5
It's not exactly a bad thought... but it is somewhat of an oversimplification.
The catch is that, with multiple subs, there's no such thing as "just aligned" or "just optimized". Because the wavelength of low bass frequencies is similar to the dimensions of a typical room each sub creates what you can think of as "standing waves". And, even for one sub, you're going to get a complex three-dimensional pattern of "nodes and nulls", which will change, depending on where the sub is located, and where the listener is located. And, when you have multiple subs, the waves from each sub interact with both the room, and all of the other subs. And, once you think you've got a picture in your head of these 3D wave interactions, consider that they're going to be different at different frequencies.
Because of this complexity there is simply no such thing as "fully optimized". Do you want the flattest response, in your favorite chair, from 20 Hz to 150 Hz... even if the guy sitting on the couch gets no bass at all? Or would you prefer a slightly less flat response at your chair... but a better sound over a larger area? And would you prefer to have response that goes lower... or response that's smoother through the upper bass? And are you more worried about explosions that rock the walls... or having a bass cello sound just right? And, by the way, are you making compromises because it's not convenient to put your sub where it sounds the best?
MSO is an excellent tool to help you figure out the best way to set up your subs to do what you want... And to help you experiment with various solutions to help you approach that goal...
But it is NOT a "push the button and everything will be perfect" sort of solution... Because of little details like the laws of physics no such absolute optimal solution is even theoretically possible. 2. I'm not sure how MSO works. I don't currently have a Win computer in operation right now, so cannot try MSO. The Left Sub and Center Sub outputs are each seen by Dirac and each are corrected individually, but Dirac is unaware how many subwoofers are involved, and we don't care. The subwoofers are collectively treated as a single virtual subwoofer by Dirac. It's up to me to make sure I've done the best I can with aligning the subwoofers in the group, then let Dirac do what it does. I just had a (probably bad) thought. Seems to me that aligned subs are aligned subs. I don't know if MSO does more than that - I expect it does - but again, once they're optimized, they're optimized. If that's true, I can run MSO against the current configuration and use it to set the miniDSP, then separate the bass management and LFE signals into the miniDSP, then run Dirac. If anyone sees a flaw in that plan, please chime in. |
|
KeithL
Administrator
Posts: 10,273
|
Post by KeithL on Feb 22, 2022 10:53:40 GMT -5
If you're only using a single output...
We recommend using the Center Sub... mostly because it's easier. And, IF YOU USE THE L/R SUB OUTS, and you're only using one, then you should set it to SINGLE, and use LEFT...
But, if you're only using one sub, and have no plans to "split LFE and bass management", then it really doesn't matter.
I found a very good video for "aligning" multiple subs together using REW. For "normal" setups with multiple subs connected to a miniDSP, one input from the processor, multiple outputs to all the subs, the procedure is very well explained. It looks to be quite a bit easier than how I've been doing things, so I'll try it out this weekend. I finally got around to setting up my miniDSP and REW, and I started with this version of aligning my three subs. I'm not done yet - haven't run Dirac after the sub alignment - but I decided to test the estimated summed combinations against the actual. I won't say they were perfect, but they were reasonably close. (I tested the first two summed against their actual, as well as the complete summed against actual.) Next up, Dirac. Then my big move will likely be gathering all the samples for Multi-Sub Optimizer later this week, and re-running Dirac after that optimization. This is all still a learning process for me, so while it's a little tedious, it's also quite interesting to see what happens. Question: I realize now that I may have messed up. I'm feeding a single sub-out to the miniDSP, but I used the center sub and have it set to Mono. I *think* I should have used the left sub out, set to Mono. Can anyone tell me definitively if that makes a difference? |
|
|
|
Post by p4t on Feb 22, 2022 12:06:55 GMT -5
Good Question, P4T. For completeness? How TALL is the room? I phrased is exactly and it is part test. The number? Known as the Golden Mean or Golden Section. But for audiophiles? A neat way to divide up space. The 62 / 38 part? Well......your long way in your room is 18 feet...... Multiply by .618 (more exact number) and you get 133.5 INCHES which is where to test your sub along the wall. Doing so will automatically SPLIT the wall in the ratio I list in my first post. If you are curious? Well, this is a VERY important number in history, going back about 5000 years or more....to the construction of the Great Pyramid and more recently in middle ages ART and today in the construction of the United Nations building in NYC. For Audiophiles? www.cardas.com/room_setup_fibonacci.phpAs it turns out, they use the Fibonacci Series which is another way to phrase the Golden Section. It turns into an EASY math problem. But if you are curious, I've given MORE than enough clues for you to get your feet wet to just above your KNEES. This math is good to use to design anything from buildings / to speaker enclosures...... Ahh.... I get it now. By the way my ceiling height is 10.5 feet. |
|
|
|
Post by leonski on Feb 22, 2022 20:51:09 GMT -5
Math Notes: Phi, or 1.618 and so on is a unique number. When squared? Itself PLUS 1. Square Root? Itself MINUS 1 It is an Irrational number, like Pi and e........ But can be found thru the Fibonacci Series. Divide any 2 of the series after the first 4 or 5 entries IN the series and you'll get an increasingly 'good' value for phi. Very early? 13/8 is good, but not great value. Later? By the time you get to 610/377 you get quite a reasonable value. www.cuemath.com/numbers/fibonacci-series/This is Geek Math and useful for those like me who are curious about the Great Pyramid. Many classic buildings and paintings used this 'ratio' to help arrange elements and provide a pleasing effect. Nature uses this in hundreds of ways. As for your ROOM? Real size is not 'square feet' of lengthXWidth but must include Height and become CUBIC FEET. This value will give you a clue as to how much subwoofer you may need to 'excite' that VOLUME of space.... A 15x20 foot room is 300 sq/ft and with an 8 foot ceiling is 2400 cubic feet. the same room with a 10 foot ceiling? Much larger @3000 cubic feet. Same smaller sub might NOT work in both rooms.....adequately for a REAL movie lover! By the time your room gets to 4000 to 5000 cubic feet or MORE? Multiple subs are almost mandatory..... |
|
|
|
Post by marcl on Feb 23, 2022 11:28:06 GMT -5
I found a very good video for "aligning" multiple subs together using REW. For "normal" setups with multiple subs connected to a miniDSP, one input from the processor, multiple outputs to all the subs, the procedure is very well explained. It looks to be quite a bit easier than how I've been doing things, so I'll try it out this weekend. I spent most of yesterday and this morning using this method to align my subs to each other, and also to align my front L/R 3.7 speakers to their corresponding DWM woofer modules and summed sub output. I had previously tried to use the Alignment Tool's automated functions to align phase at crossover points. I really didn't have much success with that in my configuration. But the method described in this video that uses the simulation of the combined response and the slider to introduce delay to one speaker worked really well! I was able to realize a significant improvement to the response of the subs alone playing LFE, as well as my front L/R playing through each 3.7, DWM and summed sub output. Then this morning I took a second step and tried repositioning the DWMs to see if I could improve response further and eliminate a cancellation at 90Hz. The DWMs play 40-300Hz and are only 2x2ft so they are easily repositioned to fill in nulls. This was also successful, and I repeated the measurements and adjustments with the Alignment Tool and realized a bit more improvement on the left channel. In addition to the frequency response improvements, I also found that phase was much better aligned between the L/R channels and Group Delay for each was much improved, eliminating some significant peaks below 100Hz. I expect all of this will correlate to better imaging. Excess Group Delay in each channel correlates exactly with the 60Hz dip ... but I have no idea how to correct that since it can't be EQ'ed. Here's the response of each channel before the Alignment Tool optimization, after initial optimization, and after DWM reposition and final optimization. The 60Hz dip is impossible to eliminate with these methods but after a Dirac calibration its effect will be minimized.  |
|
|
|
Post by leonski on Feb 23, 2022 15:34:37 GMT -5
Real crossovers CONSTANTLY shift phase thru the passband. Just look at ANY review from Stereophile. They are very good at measuring phase.
Getting it 'perfect' at one frequency almost automatically means you're OFF at another.....it's a slippery slope!
There IS an out, but it is also a LOT of work. FIR filters, which can be designed and implimented using a MiniDSP system have NO PHASE Shift thru the passband....
And while there are bunches of 'helps' for constructing those filters, the MATH is, for me, a little offputting and I need a simple step-by-step program.....
At that point? I think aligning drivers turns into s less daunting and SIMPLE adjustment of time delay. At 1ms per foot? Small adjustments rule.
|
|
ttocs
Global Moderator
I always have a wonderful time, wherever I am, whomever I'm with. (Elwood P Dowd)
Posts: 8,162
|
Post by ttocs on Feb 23, 2022 16:30:32 GMT -5
At 1ms per foot? Small adjustments rule. Exactly! |
|
|
|
Post by marcl on Feb 23, 2022 16:34:02 GMT -5
Real crossovers CONSTANTLY shift phase thru the passband. Just look at ANY review from Stereophile. They are very good at measuring phase. Getting it 'perfect' at one frequency almost automatically means you're OFF at another.....it's a slippery slope! There IS an out, but it is also a LOT of work. FIR filters, which can be designed and implimented using a MiniDSP system have NO PHASE Shift thru the passband.... And while there are bunches of 'helps' for constructing those filters, the MATH is, for me, a little offputting and I need a simple step-by-step program..... At that point? I think aligning drivers turns into s less daunting and SIMPLE adjustment of time delay. At 1ms per foot? Small adjustments rule. I have tried to figure out how to use the FIR filters in miniDSP HD and have failed so far. I tried to get one of the apps to run and it wouldn't ... another I couldn't figure out. That's just to get the FIR taps to paste into the miniDSP. I'll likely try again someday. But a little explanation of what you're seeing in my graphs. Here's the left channel after I realigned the subs but before I realigned the DWM to the 3.7: 3.7 runs full range with no HPF; Left signal is split and also goes to miniDSP where it feeds the left DWM with level and delay, and HPF at 35Hz (no LPF); miniDSP sums L+R to feed the subs HPF 10Hz, LPF 50Hz. All filters LR 48db/octave. A couple things ... I tried putting a LPF on the DWM and it screwed up the response no matter what I did, so I took it off; Left channel miniDSP HPF is 35Hz, but right channel wants to be 45Hz. So this is not a conventional crossover situation. And add to that the 3.7 is pointed at the left wall at 45 degrees and its DWM is inboard a little forward of the leading edge of the speaker. So this has, of necessity, been a methodical experiment changing each variable in turn and iterating until the response is optimized. The simplest first step is to move speakers with no delays or filters; then add delays; then adjust levels ... lather rinse repeat   |
|
|
|
Post by leonski on Feb 23, 2022 18:47:57 GMT -5
|
|
