Does anyone else phase a sub to cancel a room resonance?

[ebslo]

Sub placed just left of the right main, room eq not yet applied. Sub-to-mains time offset was twiddled (after setting crossovers) to get "best" balance of cancellation at 38Hz and support at 49 and 81 Hz. Each trace is a single measurement at the main listening position.

I then measured response over the full listening area using Moving Mic method, generated a single equalization filter set from the total response, and applied it to all channels (mains & sub) so the filters do not affect the phase relation between any of the drivers.

I tried a couple of other methods based on various guides first, but got subjectively the best result this way. Their are many guides for room equalization and many more on sub integration, but I could not find much on how doing both together and the seem tightly coupled. Posting to:

• Share experience
• Hear if anyone else has tried this and how their results compared with other procedures
• Attempt to stimulate discussion of various procedures for combined sub integration and room equalization
• Find out why this is a terrible idea and that i'm an idiot for thinking it sort of worked

 

[sigbergaudio]

Looks like you would benefit from some kind of EQ capabilities in your system.

EDIT: I may have misunderstood this post. All of this is pre-eq, but you do have EQ? Do you have a post-EQ measurement as well?

 

[ebslo]

Do you have a post-EQ measurement as well?

Here is pre-eq vs. post-eq. The measurements are not comparable to the above, as these are Moving Mic over whole listening area.

 

[sigbergaudio]

So, if I understand your initial post you're adjusting delay to the subwoofer to experiment with how this affects the frequency response? This is certainly a viable method to somewhat improve the response. Post-EQ curve looking good too!

 

[ebslo]

So, if I understand your initial post you're adjusting delay to the subwoofer to experiment with how this affects the frequency response?

Exactly, with lots of overlap; sub LP is BW24 @ 100Hz, mains HP is BW24 @ 24Hz (just to cut stuff below port resonance). My goal was to get the sub to help with the null's and reduce decay time at the main resonance in addition to adding extension. The big downside seems to be that the sub integration is now tightly coupled with the room EQ; once the phase is set I can't change anything in the DSP that affects the phase relation at any frequency or the whole thing falls down.

 

[ernestcarl]

Exactly, with lots of overlap; sub LP is BW24 @ 100Hz, mains HP is BW24 @ 24Hz (just to cut stuff below port resonance). My goal was to get the sub to help with the null's and reduce decay time at the main resonance in addition to adding extension. The big downside seems to be that the sub integration is now tightly coupled with the room EQ; once the phase is set I can't change anything in the DSP that affects the phase relation at any frequency or the whole thing falls down.

While I don't see anything obviously wrong here, I don't *think this would work well if you use the sub for x.1 LFE duty.

 

[Juhazi]

While I don't see anything obviously wrong here, I don't this would work well if you use the sub for x.1 LFE duty.

I suppose all AVRs have option to send LFE .1 signal to FL/R mains. Then you naturally miss high-pass of mains and might not get any signal from subwoofer out RCA.

This kind of no-HP assisting "multi-subwoofer" system has problems considering timing/delay. When spl response is smoothed by introducing phase/delay change, transient quality suffers. As well traditional EQ doesn't work any more because of cancellations.

 

[ernestcarl]

The difference in phase and magnitude of the inserted LFE channel will cause cancellations when combined with the summed main & sub response. I will do a quick check on how the LFE sums with my bass managed multichannel setup -- that is when all channels are playing pink noise simultaneously at the same time there should be no visible cancellation.

 

[Juhazi]

ernestcal, remember that subwoofer's plate amp will add some delay to it's signal! Measuring phase effects from acoustic output (sound) with microphone sums all kind of delays and phase shifts in the chain, including distance from loudspeaker to mic - even reflections in the room.

The real problem with these measured bass responses is always the room! Low frequencies need long gating with FFT analysis. Nearfield measurements from all bass drivers will miss air travel time....

Been there, done this kind of experiments and measurements in my HT room. The result was total confusion, there was no logic in changes I made and measured. I had max three subwoofers playing same signal, one above the bookshelf. Didn't have dsp then, only turned phase knob of subs and tried different xo frequencies in AVR and sub positions, also tried with no HP for mains. Ended with making a single 2x15" closed box sub with Hypex DS2.0. I use AVR 5.1 for movies, but for stereo music chain is different - pc - minidsp 2x4HD - sub and only front mains. I set individually eq, then xo and delay based on nearfield measurements. No LP equalization at all, but this is best bass sound for music despite of room modes.

 

[ernestcarl]

Nearfield measurements from all bass drivers will miss air travel time....

I meant at the actual listening position(s) -- so we're also taking into account the room.

 

[ernestcarl]

This is what I mean about simply checking what the total effect is if all channels were summed using pink noise:

MMM RTA

As we introduce the independent LFE signal to the bass-managed system, we should ideally not see unwanted discrepancies in any of the summed responses because of some hidden phase mismatch.

Bass managed crossover is 80Hz. But the completely separate LFE channel itself extends all the way to 120Hz -- I've cut off all of the frequencies above 120Hz to avoid cancellation with the satellites due to a phase mismatch. BTW, that dip around 90Hz is not because of a sub-satellite phase cancellation, but rather caused by the room -- and it's already present were the LFE channel playing on it's own in isolation.

Of course, my own setup is relatively simple where there's just a single sub to manage.

 

[ebslo]

While I don't see anything obviously wrong here, I don't *think this would work well if you use the sub for x.1 LFE duty.

Agreed, control of the resonance relies on the sub channel coming from summed L+R. If I switched it to an LFE the resonance would return full force from the mains. That's ok for me though, I'm stereo only with just L and R into a miniDSP 2x4HD which does xover, eq, and generation of the sub channel.

This kind of no-HP assisting "multi-subwoofer" system has problems considering timing/delay. When spl response is smoothed by introducing phase/delay change, transient quality suffers. As well traditional EQ doesn't work any more because of cancellations.

Interesting. Of all the sub integration strategies I tried, this one subjectively seemed to have the sharpest bass attack and decay at the problem notes. The others all seemed to smear certain bass notes together at one or more of the seats. I don't have any data to back this up with though because it seems difficult to get a reliable measurement. But my setup may be a bit different from what you are talking about, as I only have a single sub.

As we introduce the independent LFE signal to the bass-managed system, we should ideally not see unwanted discrepancies in any of the summed responses because of some hidden phase mismatch.

Bass managed crossover is 80Hz.

So you test this with the mains crossover disabled? Do you not get phase issues near the crossover point due to the phase change caused by the crossover filter? Or did I misunderstand your configuration?

*Edit: Just noticed on the plot "23: LR" extends to 10 Hz so that's clearly not just your mains, unless you have some really hardcore mains.

 

[ernestcarl]

So you test this with the mains crossover disabled? Do you not get phase issues near the crossover point due to the phase change caused by the crossover filter? Or did I misunderstand your configuration?

Yep, all channels are bass managed. The LFE is playing pink noise simultaneously using some manual re-routing. No phase issues with fronts (with sub) + LFE or surrounds (with sub) + LFE since I've gone out of my way to ensure the that LFE's phase response is fully compliant with the bass managed satellites -- multiple ways of doing this such as FIR or using notch filters to remove non-phase matching frequencies from the sub i.e. above 120Hz. I can play bass managed stereo or multichannel x.1 using the same system, and there's really no need to switch around. It was such a pain to manually configure, though. It is also completely possible to do this with an LFE present even if your mains extend down all the way to 40Hz instead of 80Hz etc.

 

[Helicopter]

Gloves off Amir... dirty dog.

I have trouble going after Paul, because, though a fool in the Biblical sense, he is such a charismatic stupid old grandfather.

When you said the weak part of his case was 'folklore,' i thougt you were about to say F#$*ing BS, in reference to the whole 'BS Audio' thing. Nice save... using inoffensive words and sticking to the points of the argument.

I am looking forward to his reply. I bet he describes you as a 'little guy who has his own website' or something like that. I bet he also criticizes your use of good science to refute his more personal gut instinct. That is, if he isn't smart enough to keep his mouth shut. He is such a DB, but he knows his decreasingly ingorant market. Paul won't describe ASR in terms of its relative audience size. That would be too couragously honest.

 

[Smutna]

Sub placed just left of the right main, room eq not yet applied. Sub-to-mains time offset was twiddled (after setting crossovers) to get "best" balance of cancellation at 38Hz and support at 49 and 81 Hz. Each trace is a single measurement at the main listening position.
View attachment 118227
I then measured response over the full listening area using Moving Mic method, generated a single equalization filter set from the total response, and applied it to all channels (mains & sub) so the filters do not affect the phase relation between any of the drivers.

I tried a couple of other methods based on various guides first, but got subjectively the best result this way. Their are many guides for room equalization and many more on sub integration, but I could not find much on how doing both together and the seem tightly coupled. Posting to:

• Share experience
• Hear if anyone else has tried this and how their results compared with other procedures
• Attempt to stimulate discussion of various procedures for combined sub integration and room equalization
• Find out why this is a terrible idea and that i'm an idiot for thinking it sort of worked

Hi,


I have had some trouble with adding a sub to my hifi system. I have a Yamaha S-S1200 integrated amp, KEF LS50s and a REL T5x sub. Sub is connected to Yamaha via pre-outs. My room dimensions are L4,3m/14ft X W5,7m/18,5ft X H2,6m/8,5ft. Before I go into this, I want to say that I am not a professional and learning as I go along.

I have been testing with REW and found out that I have a 40Hz room mode. Adding a sub, made that room mode even worse.

1. First I tried to fix the problem with sub placement and settings. I had only few places to put my sub because of practical reasons. It is now placed between front right speaker and the front right corner of the room. The only thing I could do to control the 40Hz peak, was to turn the sub volume down. And because the 40Hz peak was already so high without the sub, I had to leave the sub volume so low, that the sub became basically useless all together
2. Then I tried plugging my speakers. That did bring the bass of my speakers down, so I could bring sub volume up, but not by much. Some song were still unbearable. (pictures L+R_P+sub_phase_180 and L+R_P+sub_phase_0)
3. Then I basically gave up, and tried what would happen, if I turned my Node HPF on, so it cut everything under 50Hz. It actually sounded pretty good. Because of how my setup was connected, it also cut away the sub, but for now it was fine. It did sound the best so far, because now response was pretty flat down to 40 Hz. I know I could use my Node as a "preamp" and the Node's sub connection which could give better control to this problem, but I wanted to use my Yamaha's volume control. My Node is just a streamer.
4. After a while, once I got the energy to tackle this issue again. I read the REL instructions again, and again noticed the statement that the right position for phase is where the bass sounds the loudest. And I thought that’s the way to go so no point on fiddling with that. But I still wanted to try, with my speakers, because the last time I tried the phase switch was with only the sub turned on, and I noticed that the 40Hz peak went away. It was funny because I happened to have the RTA in REW ON and I literally saw in front of my eyes when the peak disappeared. I started to listen and I thought that it sounded best so far. Then I took some measurements and I thought they looked pretty good. At least the best they had so far. (picture L+R+sub_phase_180)

So now that I have found a way to control to room mode peak, what is the downside of this method. I know that the sub is now sucking out bass, but isn’t it sucking out bass only in the areas where the sub and the mains are overlapping? And in this case the 40 Hz is too high, so the sub takes that down a bit, and vise versa, in areas that the main speaker response that are a bit shallow (in my room), the sub is boosting. Does this mean that my mains and my sub are not completely out of phase, but just enough that the peaks and dips are smoothed over.

I have to comment REL’s instruction. I do appreciate their customer care for helping me out on several issues. For example this room mode issue and my high level connection issue with my Yamaha. But still I think that they give the impression that the setup process is as easy as 1,2,3. And it isn’t. After experienced all of this, all subs should have an App or a remote, that you can use to dial in the settings.

As you can see from the pictures, the L+R+sub_phase_180 looks the best. And also sounds the best.

Comments. Tips.

 

[Fredygump]

Sub placed just left of the right main, room eq not yet applied. Sub-to-mains time offset was twiddled (after setting crossovers) to get "best" balance of cancellation at 38Hz and support at 49 and 81 Hz. Each trace is a single measurement at the main listening position.
View attachment 118227
I then measured response over the full listening area using Moving Mic method, generated a single equalization filter set from the total response, and applied it to all channels (mains & sub) so the filters do not affect the phase relation between any of the drivers.

I tried a couple of other methods based on various guides first, but got subjectively the best result this way. Their are many guides for room equalization and many more on sub integration, but I could not find much on how doing both together and the seem tightly coupled. Posting to:

• Share experience
• Hear if anyone else has tried this and how their results compared with other procedures
• Attempt to stimulate discussion of various procedures for combined sub integration and room equalization
• Find out why this is a terrible idea and that i'm an idiot for thinking it sort of worked

My custom speaker with co-located multiple subwoofers depends on using delays to correct room modes. In my case, carefully selected delays allow co-located drivers to interact in a room as individual drivers in different locations. I have been using delays only. The only EQ I am using is to correct the overall response of each driver.

One trick I have found is start with a delay on all channels, so I can use a negative delay.

I posted a thread here awhile ago, but didn't get much interest. https://www.audiosciencereview.com/...on-of-w371a-8381a-style-speaker-system.45223/

I am planning to run an experiment with a worst case scenario of 4 co-located subwoofers. I want to see how far I can push the idea, and I hope to present my findings at some point in as scientific a way as I can.

More details of the construction: (I thought this was going to be final, but I've put off building the second "finished" cabinet, because I'm not 100% happy with the aesthetics.)

Build Thread of DIY Genelec 8381A/ W371A (Evolution of Integrating Multiple Subwoofers)

Officially introducing my new speaker build. This one is "mostly" done, but unfortunately I won't be done with the second cabinet in time to make it to the Parts Express speaker competition. But I am far enough along that I have something to show. My goal is to introduce an idea that is...

www.diyaudio.com