I have a 4-way digitally active system with two 12" sealed (Rythmik kits), and even though I use Acourate for room correction the room seems to be just awful and response in the lower octaves is a challenge. I'm considering adding another 2 similar subs, a miniDSP 2x4HD, and use MSO to get a more even response. I'm using MSO to simulate the corrected response, yet I wonder how accurate the simulation might be vs reality. Getting input from those who have done this, who can maybe confirm the simulation is representative of what I will get in reality, would be great.
Today I have the two subs placed on the front corners. The system is in the living room so aesthetics and ergonomics limit placement options. Room is 13' long x 15' wide x 8' high, with a large opening to the left to another room of similar size (dining room). I have a fair amount of bass traps on front and back wall, and on the ceiling.
Against the back wall there is a 3-person couch. For "critical" listening I bring a chair in front of this couch, which I call the main listening position (head about 4' from back wall). The back left and right positions are on the couch, which is where I sit with my wife/kids to listen to music/watch movies (it's a music system, though).
I took measurements from these 3 positions of the front right and front left subs, then moved the subs where the new ones would be located and took the next 6 measurements.
The 4 measurements of the subs at the main listening position.
Subs are labeled Front Right, Front Left, Back Right, Back Left.
See how the FL sub drops above 40Hz. FR makes it to 70Hz and drops off a cliff. The room mode at 35Hz is quite noticeable.
BTW, these measurements were made without windowing.
Note the hole for FL and FR in the 45 to 65Hz region.
So I bring the above into MSO and get this initial summation at each listening position, uncorrected:
Don't know that it sounds as bad as it looks, but it does look ugly.
I ran an initial correction, and after 30min:
This is looking much better. Still issues in the 65 to 90Hz region, but much smaller issues than before. Maybe crossing this at 65Hz will do the trick (to open baffles). I was estimating a xo at 80hz or below anyways. Would need to experiment.
What has been your experience with MSO modelling/predicting capabilities?
Feedback super welcomed.
Today I have the two subs placed on the front corners. The system is in the living room so aesthetics and ergonomics limit placement options. Room is 13' long x 15' wide x 8' high, with a large opening to the left to another room of similar size (dining room). I have a fair amount of bass traps on front and back wall, and on the ceiling.
Against the back wall there is a 3-person couch. For "critical" listening I bring a chair in front of this couch, which I call the main listening position (head about 4' from back wall). The back left and right positions are on the couch, which is where I sit with my wife/kids to listen to music/watch movies (it's a music system, though).
I took measurements from these 3 positions of the front right and front left subs, then moved the subs where the new ones would be located and took the next 6 measurements.
The 4 measurements of the subs at the main listening position.
Subs are labeled Front Right, Front Left, Back Right, Back Left.
See how the FL sub drops above 40Hz. FR makes it to 70Hz and drops off a cliff. The room mode at 35Hz is quite noticeable.
BTW, these measurements were made without windowing.
Note the hole for FL and FR in the 45 to 65Hz region.
So I bring the above into MSO and get this initial summation at each listening position, uncorrected:
Don't know that it sounds as bad as it looks, but it does look ugly.
I ran an initial correction, and after 30min:
This is looking much better. Still issues in the 65 to 90Hz region, but much smaller issues than before. Maybe crossing this at 65Hz will do the trick (to open baffles). I was estimating a xo at 80hz or below anyways. Would need to experiment.
What has been your experience with MSO modelling/predicting capabilities?
Feedback super welcomed.
If you spend some time reading the MSO thread at avsforum, you'll see many people amazed by MSO's predictions. And the people that aren't amazed usually just have configuration issues with their AVR, for example. For me, the prediction was spot on and the improvement was the best thing I've done for my 2.2 setup.
Before you get another two subs I'd try running MSO with your current setup, if you haven't already. As you mention, the front position responses seem to drop pretty fast, but if you position the subs at FL and BR (for example), I think you could get a pretty good combined response. You might want to measure some more listening positions for best results. You could, for example, measure the left and right side of your listening chair instead of a single position at the middle. This way you also increase the weighting on the chair a bit without making the software overfit things. Also make sure to make a few copies of your configuration in MSO and run the optimization separately for each. The results can vary quite a bit, as the optimization can get stuck in local minima.
You can also play around with the crossover frequency limits depending on how much your mains require help and how much you can sacrifice imaging. Try different PEQ boost limits too, depending on how much headroom you have in the subs. I also had better results by allowing different SPLs at different positions and by including my rather bass-heavy room curve. The latter is not recommended by the author because it might make it difficult to apply a new room curve without running a new optimization, but I take a moving mic measurement of my combined response after applying MSO to apply a room curve anyway so it's not a problem for me and reduces the total amount of filters, for what it's worth.
Before you get another two subs I'd try running MSO with your current setup, if you haven't already. As you mention, the front position responses seem to drop pretty fast, but if you position the subs at FL and BR (for example), I think you could get a pretty good combined response. You might want to measure some more listening positions for best results. You could, for example, measure the left and right side of your listening chair instead of a single position at the middle. This way you also increase the weighting on the chair a bit without making the software overfit things. Also make sure to make a few copies of your configuration in MSO and run the optimization separately for each. The results can vary quite a bit, as the optimization can get stuck in local minima.
You can also play around with the crossover frequency limits depending on how much your mains require help and how much you can sacrifice imaging. Try different PEQ boost limits too, depending on how much headroom you have in the subs. I also had better results by allowing different SPLs at different positions and by including my rather bass-heavy room curve. The latter is not recommended by the author because it might make it difficult to apply a new room curve without running a new optimization, but I take a moving mic measurement of my combined response after applying MSO to apply a room curve anyway so it's not a problem for me and reduces the total amount of filters, for what it's worth.
Great to know!! I wasn't aware of AVS MSO forum. Will certainly go and read. Great to know MSO predictions were spot on.
Indeed, I was thinking along the same lines. The BR sub would be most controversial aesthetically, so I tried simulating 3 subs without BR. Will post it in a minute. Was planning to run 2 subs too. Will try your suggested FL+BR too.
Measure more listening positions even if those aren't really positions where people will sit? My MLP is a chair I place in front of the couch when I listen alone, so measuring to its right and left would be easy yet I wouldn't really have someone experiencing the response in that location.
Interesting. Play with crossover limits as in raising them beyond 80Hz? Could do and see how this might impact imaging, but not clear how this would improve things?
Boosting PEQ have been something I interpreted wasn't a good idea, but it's easy to test here so will give it a try too.
Thanks for the feedback!!
Here's the simulation optimizing for 3 subs: front right and left, plus back left. It's an improvement over the baseline, but quite a bit worse than 4 subs. I was kind of expecting less difference between optimized 3 vs. 4 subs, mostly from comments by Earl Geddes and others. Maybe more tweaking of MSO parameters is in order?
Yeah. I'm not an expert, but for example the most comprehensive Dirac setup has you basically measure the corners of your listening area (e.g. 8 points) which I believe form the "extremes" of response variation, and then a few measurements right next to the listening positions. If you don't move or slouch a lot in your chair, I would say two measurements around your head should do, but if you move a lot, you could do like 4-5 in a wider area.
Keep in mind that if things are bad, even the 20cm distance between your ears can mean a significant change in response. With one sub, I had this super sharp 20x20x20cm area where around 105Hz would drop completely. It was really funky listening to a tone that you could only hear with one ear, when the other ear was in the cancellation. Obviously a null this sharp is not a huge impact on actual music, but the principle still applies. If you only measure one position for the chair, you might end up "overfitting" for a sharp null that doesn't exist elsewhere.
The idea is that the subs can fill some of the dips in your mains, rather than only fixing the subs' response. But going past 100Hz or so in the XO in my experience brings a rather negligible benefit and you might lose some of the improvement on sub-bass, as you have a fixed amount of PEQs. If you're strict on imaging, you could also try lowering the XO. I just let MSO find the XO for me between 40Hz and 100Hz for my front sub, and between 40Hz and 60Hz for the sub to my left, as it would muddy the imaging a bit with a higher XO.
Indeed, not a good idea in general, but the way I did it was to increase the sub amp volume and decrease the sub output in the MiniDSP. If a sub then needs an overall boost, MSO gives you a total gain which I just reduced from the volume decrease in MiniDSP, so it shouldn't result in clipping. There are still some PEQ boosts but I tried to find an optimization result that didn't have too large or sharp boosts.
I'm not super experienced on MSO or multiple subs in general, but it does look to me like your 3 sub setup should be smoother. Attached is my dual sub setup results in a very untreated room, FWIW. Unfortunately (and, luckily) there are tons of variables you can adjust and it's quite impossible to say which ones you should try
Attachments
Last edited:
Wanted to follow up and thank you for the input. I went to the avsforum MSO thread, and while 170 pages of thread are intimidating, I did some reading and found a great reference for how to go about choosing the number of PEQs. I had been using 4, per the tutorial. I now know I have 12 axial modes below 110Hz, so I tried the 3-sub simulation with 12 PEQs, shown below. I'm including the 4 PEQ optimization as well, for reference. Same 3 subs. Still room to improve, but much better result.
More reading and experimentation in order. Will try with 4 subs and 12 PEQ.
Thank you!
More reading and experimentation in order. Will try with 4 subs and 12 PEQ.
Thank you!
Please share any further improvements you may have achieved or other experiences you've had with optimizing your room's acoustics.
Also, have you had any issues with your pair of Rythmik 12" sealed subs? I am thinking about buying a pair. I've never once read a disparaging word about them, though some here have not been enamored of the servo controlled concept, however well Brian Ding may have applied it. Tonality? Resolution? Sound stage size? Did you later add more than the two subs?
Last edited:
Hello @oltos
I went from a REL Storm III (10" ported) to two Rythmik 12" sealed and have been very happy. Never looked back.
I also actively cross over before the subs so a number of things get in the mix for what I hear now - not just the change in sub. My room has awful modes and I made significant treatments too.
As mentioned above I also implemented miniDSP 2x4HD and MSO for the two subs, which further improved things.
Coincidentally last week I added two additional Rythmik sealed 12" (like F12 but in DIY box) and have now disabled MSO and connected the 4 subs to miniDSP without optimization. I like what I'm hearing! As others have mentioned the variations across the room have greatly diminished, I think.
Next I need to take measurements, run MSO, and enter the filters into miniDSP.
Yesterday I was listening, it wasn't too loud, yet I was thinking I feel the bass in the body more that before. It wasn't SPL as I dialed back the subs, but I was thinking it has gotten more tactile. It might just be in my head. Time will tell!
I went from a REL Storm III (10" ported) to two Rythmik 12" sealed and have been very happy. Never looked back.
I also actively cross over before the subs so a number of things get in the mix for what I hear now - not just the change in sub. My room has awful modes and I made significant treatments too.
As mentioned above I also implemented miniDSP 2x4HD and MSO for the two subs, which further improved things.
Coincidentally last week I added two additional Rythmik sealed 12" (like F12 but in DIY box) and have now disabled MSO and connected the 4 subs to miniDSP without optimization. I like what I'm hearing! As others have mentioned the variations across the room have greatly diminished, I think.
Next I need to take measurements, run MSO, and enter the filters into miniDSP.
Yesterday I was listening, it wasn't too loud, yet I was thinking I feel the bass in the body more that before. It wasn't SPL as I dialed back the subs, but I was thinking it has gotten more tactile. It might just be in my head. Time will tell!
Thank you! Very good report! You're the first I've met to use two pairs of Rythmik sealed 12" subs. That's my plan or my two 12" sealed Rythmiks plus maybe two GR Research OB subs. Please keep us posted on your measurements and listening experiences!
3ll3d00d, DAC output voltage to offset gain loss when using DIRAC Live 3, Audiolense or whatever Windows room correction software is best for newbies, can be a problem because even the balanced outputs of most audiophile and even many pro audio DACs is usually not much beyond 4 volts. That's why I've kept in mind posts at Gearspace saying that it's important to correct the room acoustically as much as possible before pursuing software based room correction, so gain losses can be minimized, among other reasons.
The one MCH DAC I know of (for my eventual 5.3 surround sound system) with and optional output voltage as much as 12 volts is https://www.oktoresearch.com/dac8pro.htm. Alternately, if much more expensively, for a two channel system and three or four (mono) subs, three of these DACs. https://benchmarkmedia.com/collecti...rk-dac3-hgc-digital-to-analog-audio-converter
Yes, I use Acourate as well. The system is active 4-way stereo and I need 6 DAC channels for midbass, midrange and tweeters, and don't have additional 4 channels for subs. So I take a digital (optical) out and into miniDSP. After running MSO I treat subs as another channel, measure and apply Acourate to the overall response. I'm still iterating for the best setup/settings.
Glad it helped!
FWIW I strongly considered buying the GR sub kit from Rythmik for my 2nd pair. I exchanged with Danny and concluded I did not have enough space to properly position the OB subs. I was thinking of the sealed pair up to 30-40Hz, the OB pair above that. Danny uses this setup at shows in large, hard to fill rooms, but didn't think it was needed in a home setup. I opted for 4 subs to minimize SPL variations across the room. I will be forever intrigued by OB subs though!
@oltos
You seem to be going in the direction of correction software. The maker of Acourate strongly suggests having DAC and ADC done in the same unit to prevent drifting caused by the use of different clocks in the processes of measuring and replaying.
The top units I know that do this are Lynx Hilo (and Aurora, but lesser quality converters), Prism Titan, and Merging Hapi. I have Hilo. At the time was undecided between Hilo and Titan. One day I might switch to Hapi.
Needless to say this is not necessarily "the best" but the best to my knowledge!
You should know this is an endless rabbit hole
Cheers
You seem to be going in the direction of correction software. The maker of Acourate strongly suggests having DAC and ADC done in the same unit to prevent drifting caused by the use of different clocks in the processes of measuring and replaying.
The top units I know that do this are Lynx Hilo (and Aurora, but lesser quality converters), Prism Titan, and Merging Hapi. I have Hilo. At the time was undecided between Hilo and Titan. One day I might switch to Hapi.
Needless to say this is not necessarily "the best" but the best to my knowledge!
You should know this is an endless rabbit hole
Cheers
I have no experience with mso.
However, if i was concerned with response below 30, personally id like to see some boost (not flat response) in the infrabass to compensate for our rapidly decreased hearing sensitivity in that range
Also my experience with eqing room modes has been so so. Refer to "minimum phase" phenomenon
https://www.roomeqwizard.com/help/help_en-GB/html/minimumphase.html#top
However, if i was concerned with response below 30, personally id like to see some boost (not flat response) in the infrabass to compensate for our rapidly decreased hearing sensitivity in that range
Also my experience with eqing room modes has been so so. Refer to "minimum phase" phenomenon
https://www.roomeqwizard.com/help/help_en-GB/html/minimumphase.html#top
Attachments
That's certainly seems a sensible precaution, however, I wonder what ADC (s) Kalman Rubinson used when taking mic measurements to use with his Exasound S88 8-channel DAC. https://www.stereophile.com/content/topping-pre90-line-preamplifier However, as there's no indication that his system uses active crossovers perhaps any ADC vs DAC clock deviations were not audible. https://www.stereophile.com/content/topping-pre90-line-preamplifier-associated-equipment https://www.stereophile.com/content/exasound-s88-multichannel-da-processor
Likewise, both my two-channel system and the 5.3 system that I want to build next year will also have passive crossovers. Indeed, is that likely why Kalman had the nearly 20db system loss problem when using DIRAC Live 3, and had to use those 3 Topping preamps to restore it?? At least that preamp had amazingly low distortion; check out Figs. 3 thru 5 here. https://www.stereophile.com/content/topping-pre90-line-preamplifier-measurements
Last edited: