Bass ringing issues in an odd-shaped room and how I solved it

[fluid]

The dip got way narrower, shallower and moved upwards - really cool!

You can see in that graph that the problem is composed of quite a few separate narrow dips combining to form a bigger problem.

 

[fluid]

Fully agreed, however physical solutions are not an option in my friend's living room hence the band aids

Understood but it could be as simple as moving one speaker relative to the other. I'm only advising to be sure what is happening before deciding which band aid to apply

 

[abdo123]

A very important point, manipulating phase while only looking at graphs is an all too easy way to destroy the sound.

For summation problems like that it is important to look at the cause, why do the two speakers cancel each other out? Looking at the SPL and Phase is one way, the spectrogram can also give a clue as often nulls develop in time. Sometimes it is better to attenuate the problem frequency range in one speaker and boost the other side a little to compensate so they don't fight each other as much. Placement can often solve these sort of things without any processing being needed.

Processing like Dirac needs to be the icing on the cake not a band aid for taming acoustic issues that are easily solvable with physical solutions.

I don’t see how physical solutions can solve this. Unless you’re telling people to live in anechoic chambers. Because at the end of the day in a stereo both speakers should be the same distance away from the listener.

I agree with the fact that everything should be tested by ear, and my corrections are usually not as brute force as OP’s but I don’t see how a physical solution can solve this problem (aside from getting an AVR and a center channel)

 

[sarumbear]

By ringing I mean the sound of decaying bass notes in the room (post ringing, for example after a kick-drum is hit)
As I explained above, my odd shaped living room produces really weird ringing phenomena that was resolved by the M/S EQ-ing method after years of trying many things

You mean the "decay time", which is caused by room's acoustics. Normally the only practical solution is using bass traps, which are absorbers that work on the bass range of the audio spectrum.

 

[fluid]

I don’t see how physical solutions can solve this. Unless you’re telling people to live in anechoic chambers. Because at the end of the day in a stereo both speakers should be the same distance away from the listener.

I agree with the fact that everything should be tested by ear, and my corrections are usually not as brute force as OP’s but I don’t see how a physical solution can solve this problem (aside from getting an AVR and a center channel)

There are two problems presented, the one belonging to the thread title and a second related to a different system with an issue between the two channels integrating.

My comments refer to the latter which is why I quoted those sections. In that case the problem is one of the distance between the speakers, the boundaries and the listening position. Adjusting those for the best compromise before attempting electronic correction. No anechoic chamber required.

 

[ernestcarl]

A very important point, manipulating phase while only looking at graphs is an all too easy way to destroy the sound.

For summation problems like that it is important to look at the cause, why do the two speakers cancel each other out? Looking at the SPL and Phase is one way, the spectrogram can also give a clue as often nulls develop in time. Sometimes it is better to attenuate the problem frequency range in one speaker and boost the other side a little to compensate so they don't fight each other as much. Placement can often solve these sort of things without any processing being needed.

Processing like Dirac needs to be the icing on the cake not a band aid for taming acoustic issues that are easily solvable with physical solutions.

Yeah, steady state graphs showing only the magnitude will not show the full effect of the phase EQ filter(s).

I set the AP filter at 55Hz Q=3.5

Whether or not one can clearly hear that filter as sounding better/worse/neutral would depend on the final system as a whole. I don't know how you created your filter, but if you used rePhase it should "look" something like the ff:

Wavelet 1/6 smoothing

*Now, the last one seems rather extreme -- so I am interested in knowing what the actual in-room listening results are...

 

[ppataki]

My friend did not have the time to test (listen to) the AP filter I sent him
However I did test it on my own system

This is how my Left and Right channels look like with the L+R summation highlighted

Pretty nasty cancellations between 50-60Hz and at about 150-180Hz

I tried to fix the 50-60Hz dip with an AP filter (54.5Hz Q=7) in Jriver's PEQ applied on the right channel

As you can see the simulation (and also the actual measurement) looks much better

But when I listened to it I could hear that some bass notes were actually 'crawling' out of the stage from the right channel - as if those notes were coming from a third loudspeaker outside the stage

This is how the spectrogram looks like for the normal summation:

and with the AP filter

I am wondering if it is the elevated peak of the energy curve that I can hear.... any comments are welcome

For me the conclusion is that these tricks with the AP filters do look great on the FR curve but when actually listened to - not really
I guess @ernestcarl that was your conclusion too

 

[ernestcarl]

My friend did not have the time to test (listen to) the AP filter I sent him
However I did test it on my own system

This is how my Left and Right channels look like with the L+R summation highlighted

View attachment 278570

Pretty nasty cancellations between 50-60Hz and at about 150-180Hz

I tried to fix the 50-60Hz dip with an AP filter (54.5Hz Q=7) in Jriver's PEQ applied on the right channel

View attachment 278571

As you can see the simulation (and also the actual measurement) looks much better

But when I listened to it I could hear that some bass notes were actually 'crawling' out of the stage from the right channel - as if those notes were coming from a third loudspeaker outside the stage

This is how the spectrogram looks like for the normal summation:
View attachment 278573

and with the AP filter

View attachment 278575

I am wondering if it is the elevated peak of the energy curve that I can hear.... any comments are welcome

For me the conclusion is that these tricks with the AP filters do look great on the FR curve but when actually listened to - not really
I guess @ernestcarl that was your conclusion too

Shallower q and basically first order all-pass filters are largely “benign”. In a system that is very linear in both time and frequency, second order and higher all pass filters is something that I can hear in the bass with critical listening. It’s not necessarily worse, but it can mask the midrange or make it less clear.

The actual sound-field in your listening room, as you’ve explained before, is rather complicated so I wouldn’t really be able to able to draw conclusions of the exact cause of the shift perceived just by looking at those graphs alone.

 

[ozzy9832001]

You have a lot of standing waves and slower decay times. That can create the effect that you are describing. If you are near to a corner the effect can be amplified, especially if you are listening in the nearfield. You're probably also getting some reflections off that angled wall back at you which are early compared to the ones from the left corner into the alcove. This would explain why the left speaker is fine but right is not.

You also have dips at 160 and 180. Probably model as well.

A subwoofer would probably fix the sub bass issue and possibly the others as well if positioned properly and crossed over a bit higher.

 

[neRok]

I would love to understand if there is a better way than just trial and error

Use filtered IR's. REW's Overlay screen has a few predefined settings, but since I discovered the Alignment Tool > Impulse alignment, it has become much easier.

I'm guessing that one speaker has more reflections than the other, and that is causing a phase shift during its IR, and thus they don't sum together good. Same thing happened to this user who has a rectangular room;

So if you allpass one speaker in the problem area, which phase shifts it 180, the bulk of the power will sum. But there will still be a bit that won't sum properly (at the start, because the direct waves will be out of phase, but the reflections will be in phase).

Now I can see why you would have this in your room = because it is not symmetrical. My room is similar, in that I have a section of angled wall behind me that acts as a "2nd back wall". You didn't describe your friends room, so it could be anything.

If your room is to scale, it might be particularly bad because you are sort of equal distance from your angled wall and your front wall, putting you right in the middle of a sort of null.

 

[ppataki]

Thanks for the advice, much appreciated!
To be honest the way I fixed it (using mid-side processing) works and sounds perfect so I think I am settled with this issue

Next year we might rearrange the whole living room so I might get back to this thread to share my new experience