Phase Issue Impacting Right Channel Only

[Bugal1998]

Hi all,

I have a puzzling phase issue that is deteriorating the spaciousness and envelopment of the audio system in a dedicated listening room, and I'm interested in hearing any thoughts on what may be happening. I made a series of changes to my room and I can't pin down exactly when the issue began occurring, but after putting new seating in the room that replaced giant bean bag chairs I have a phase issue with the right channel.

Previously I was able to get mono pink noise played out of phase to be completely diffuse across the entire spectrum and music was spacious and enveloping. Lately I've been noticing that music is less spacious, and when I checked mono-out-of-phase pink noise there's a notable amount of sound that isn't diffuse across the spectrum.

With all room correction disengaged I took phase measurements of the left and right speakers at the mouth of each JBL M2 horn. Single-channel phase measurements were then taken at the listening position with the mic acoustically time aligned between the speakers (the speakers should be very nearly symmetrical in the room as well). The close mic phase response between the speakers has very good agreement, but at the listening position the right channel has a lot more phase variation. This phase variation is shown by both OmniMic and Trinnov, whereas previously the phase responses matched at the listening position.

Left channel phase at horn mouth:

Right channel phase at horn mouth:

Left channel phase measured at MLP:

Right channel phase measured at MLP (consistent with left channel until ~900HZ, then greater variability):

Questions popping into my head as I type this include:

Is the right speaker level and in position, or did it get bumped during the furniture install?
Diffraction effects from inconsistent toe-in; again is it out of position?
Could this be caused by a visually imperceptible difference in asymmetry of the newly installed leather seating?
Any ideas?

Many thanks in advance!

 

[Curvature]

The change in envelopment doesn't make sense. There's nothing telling in the measurements you posted.

Speaker placement shouldn't be so sensitive, but you have a dedicated room with a lot of diffusers. Even then, I don't think small changes or errors in position would have that much effect.

I'll take a guess: leather chairs are thick, wide and have high backs with headrests? If so, and if you are now listening in one of those compared to what you had, below, then it's probably the chair headrest blocking sound form the sides and behind, catching extra sound from the front, and in general changing what reaches your ears.

 

[Bugal1998]

The change in envelopment doesn't make sense. There's nothing telling in the measurements you posted.

Speaker placement shouldn't be so sensitive, but you have a dedicated room with a lot of diffusers. Even then, I don't think small changes or errors in position would have that much effect.

I'll take a guess: leather chairs are thick, wide and have high backs with headrests? If so, and if you are now listening in one of those compared to what you had, below, then it's probably the chair headrest blocking sound form the sides and behind, catching extra sound from the front, and in general changing what reaches your ears.

View attachment 295313

Good thoughts! But why is only the right channel showing the extra phase shifts above 1k? Seat placement?

Edit: Or more correctly, why is the right channel showing significantly more phase shifts between 3k and 10k than the left channel at the listening position?

 

[Curvature]

Good thoughts! But why is only the right channel showing the extra phase shifts above 1k? Seat placement?

Edit: Or more correctly, why is the right channel showing significantly more phase shifts between 3k and 10k than the left channel at the listening position?

I think you should probably ignore the phase plot for now, which is extremely sensitive at HF, and try moving the leather chairs around, or try the lawn chair again and see what happens. If that doesn't work I'm out of immediate ideas.

 

[Bugal1998]

I think you should probably ignore the phase plot for now, which is extremely sensitive at HF, and try moving the leather chairs around, or try the lawn chair again and see what happens. If that doesn't work I'm out of immediate ideas.

It's entirely possible my implied hypothesis about phase differences being either a symptom of the culprit or the actual culprit is false. Despite the sensitivity, the measurement is repeatable and the differences are consistent; it's always the right channel.

If fine tuning speaker and furniture positions is able to resolve the issue it will be interesting to see if the L/R phase responses also come into better agreement.

 

[Curvature]

It's entirely possible my implied hypothesis about phase differences being either a symptom of the culprit or the actual culprit is false. Despite the sensitivity, the measurement is repeatable and the differences are consistent; it's always the right channel.

If fine tuning speaker and furniture positions is able to resolve the issue it will be interesting to see if the L/R phase responses also come into better agreement.

I would suggest using your ears as an expedient. The problems you identified with pink noise and music are fairly obvious, I would imagine.

It's interesting that the measurement is repeatable. If you have a laser measure I would calculate path length directly and for reflections to satisfy curiosity (it's likely asymmetrical between L/R). Although, like you said, I would not bet the measurable phase difference is the main cause. It should be something obvious. At HF reflections are so dense in a room that phase isn't an important component of audibility.

In general I would consider using REW. I find it much easier to make comparisons using its interface and you can also attach the mdat files to this thread for us to look at.

 

[Bugal1998]

If you have a laser measure I would calculate path length directly and for reflections to satisfy curiosity (it's likely asymmetrical between L/R).

I do and I have, but I need to revalidate.

In general I would consider using REW. I find it much easier to make comparisons using its interface and you can also attach the mdat files to this thread for us to look at.

That's a good push for REW.

 

[Bugal1998]

Quick update. Setting Trinnov Room Optimizer to account for early reflections almost entirely resolved the issue. This is opposite behavior compared to the settings that provided the most spacious results before changing seating.

I don’t love the slight tonality change in the mids and highs with this setting, so still more tweaking, but there's a path forward.

Busy time, so additional measurements will have to wait a week or two.

 

[Curvature]

Quick update. Setting Trinnov Room Optimizer to account for early reflections almost entirely resolved the issue. This is opposite behavior compared to the settings that provided the most spacious results before changing seating.

I don’t love the slight tonality change in the mids and highs with this setting, so still more tweaking, but there's a path forward.

Busy time, so additional measurements will have to wait a week or two.

It makes sense some if the seating is the cause, although I'm not fully familiar with how Trinnov works.

 

[Bugal1998]

It makes sense some if the seating is the cause, although I'm not fully familiar with how Trinnov works.

You were correct about the impact of phase. I resolved both the tonality issues and the envelopment issue with minor changes to seat positions, but the phase differences persist.

Thanks again for the nudge in the right direction.

Next step is to experiment with targeted FIR filters to manipate phase response to yield any further improvements...