Where does my soundstage go?

[ErLan]

Hello,

I am deep in my journey to sound perfection in my living room (I guess it is a never-ending one...)

Background
My setup & configuration are as follow:

• Room: typical living room (75 Meter^3) non treated, but equipped with Sofas, curtains, bookshelves, etc.
• Rt & Lt speakers: Elac Carina 243.4,
• Subwoofer: Wharfedale WH-D10 (active)
• Amplifier: Behringer A800
• Music & movies are streamed to the amplifiers via. a NUC PC \ Asus U5 sound card.
• Crossing over (80Hz)\ bass-management is done via. EQAPO (controlling Lt, Rt, Sub separately). to the (active) Sub and Behringer amp through Asus U5 sound card.
• Typical dB in my MLP is 65-75dB, taking into consideration ~3m distance between MLP and Speakers we are talking ~75-85dB@1m
  * Bringing it up as per spinorama measurements this is still in the 'non-distorted' range of this speakers.

Now...
In the past months I am working to fine tune and integrate my Sub to the Lt & Rt.
Following the good advises of many ASR members I was able to test (both listening and REW measurements) several of the proposed approaches for alignment and EQ schemes (IIR, FIR, mixed). I believe that now I got to a good sound which I am (almost) feel comfortable with (which is not a trivial statement).

My current search is looking at sound perception of Lt & Rt vs. Lt & Rt & Sub configurations...

• When using Lt & Rt only, the sounds seem clear, and the holographic perception is very good (3D, very detailed);
• When engaging the Sub, the bass is more present (naturally), however it looks like the holographic view collapses from 3D to 2.5D, mostly impacting the soundstage depth'.

I confirmed (via measurements) that the sub integration is only extending the natural rolloff of the speakers (45Hz >> 30Hz), without boosting the low-end SPL.
I am also conservative regarding EQ, Where I am focused on room modes only (1/6th window, no boost, upto 300Hz, crossing low) in order to retain most of the natural behavior of Lt & Rt speakers.
I further checked this perception via High passing Lt & Rt & Sub at 45Hz so the output range of both configurations will be close; but I can still note the change between configurations.

Now, I am a puzzled
What may be the source of this observation?
Are there any additional quantitative\qualitative parameters that I should look at and may shed light on that?
Is it an expected behavior where you compensate one feature (bass extension) with the another (soundstage)?
Am I hallucinating?

Looking forward for your good insights!

 

[Wolf11Man]

Usually with a sub the soundstage is improved. Even if the Elac are only about 600€ a pair, that should work. Highpass them not lower than 80Hz, at 45Hz they are already stressed. Then match the sub to it. Get the phase right, this is the usual problem with a sub.

Make sure your source material has CD quality, with data reduction the soundstage is the first thing to disapear.

PS plug the port of the Elac's

 

[Godataloss]

As noted try reversing the phase of the sub, also I think the increased energy the sub provides affects the psycho-acoustic effect in ways that can make problems with your room more apparent and distracting.

 

[krabapple]

I don't quite understand.

You're crossing over the system at 80 Hz, and say "sub integration is only extending the natural rolloff of the speakers (45Hz >> 30Hz),"

According to these Klippel measurements, the 'natural roll off ' of the Carina appears to be well above 30- 45 Hz. More like 60Hz.

(Amir's had a 'BBC dip' at 2.8kH but the sub shouldn't impact that.)

Anyway, I'm guessing you need to do more comparative measurements of with vs without sub, from the MLP. Perhaps post them here for diagnosis. It would seem to be something happening in the room mode range. You are correcting up to 300Hz, have you tried restricting that range further?

 

[DVDdoug]

Am I hallucinating?

Well... soundstage is obviously an illusion since the sound is coming from the speakers...

Floyd Toole says:

The important localization and soundstage information is the responsibility of the recording engineer, not the loudspeaker.

But acoustics, speaker location and listening position, and your brain are also involved. And I assume that omni-directional or dipole speakers also make a difference.

with data reduction the soundstage is the first thing to disapear.

That's news to me... Have you got links to any studies on that? Good high-bitrate compression is often transparent in blind listening tests, which means there is often no audible difference at all. Usually the first (most subtle) difference is something called pre-echo. (The first measurable difference is the loss of the highest frequencies, but that's not usually what's heard.)

 

[krabapple]

That's news to me... Have you got links to any studies on that? Good high-bitrate compression is often transparent in blind listening tests, which means there is often no audible difference at all. Usually the first (most subtle) difference is something called pre-echo. (The first measurable difference is the loss of the highest frequencies, but that's not usually what's heard.)

Yeah, lossy vs lossless is not really a factor here.

 

[kemmler3D]

Hello,

I am deep in my journey to sound perfection in my living room (I guess it is a never-ending one...)

Background
My setup & configuration are as follow:

• Room: typical living room (75 Meter^3) non treated, but equipped with Sofas, curtains, bookshelves, etc.
• Rt & Lt speakers: Elac Carina 243.4,
• Subwoofer: Wharfedale WH-D10 (active)
• Amplifier: Behringer A800
• Music & movies are streamed to the amplifiers via. a NUC PC \ Asus U5 sound card.
• Crossing over (80Hz)\ bass-management is done via. EQAPO (controlling Lt, Rt, Sub separately). to the (active) Sub and Behringer amp through Asus U5 sound card.
• Typical dB in my MLP is 65-75dB, taking into consideration ~3m distance between MLP and Speakers we are talking ~75-85dB@1m
  * Bringing it up as per spinorama measurements this is still in the 'non-distorted' range of this speakers.

Now...
In the past months I am working to fine tune and integrate my Sub to the Lt & Rt.
Following the good advises of many ASR members I was able to test (both listening and REW measurements) several of the proposed approaches for alignment and EQ schemes (IIR, FIR, mixed). I believe that now I got to a good sound which I am (almost) feel comfortable with (which is not a trivial statement).

My current search is looking at sound perception of Lt & Rt vs. Lt & Rt & Sub configurations...

• When using Lt & Rt only, the sounds seem clear, and the holographic perception is very good (3D, very detailed);
• When engaging the Sub, the bass is more present (naturally), however it looks like the holographic view collapses from 3D to 2.5D, mostly impacting the soundstage depth'.

I confirmed (via measurements) that the sub integration is only extending the natural rolloff of the speakers (45Hz >> 30Hz), without boosting the low-end SPL.
I am also conservative regarding EQ, Where I am focused on room modes only (1/6th window, no boost, upto 300Hz, crossing low) in order to retain most of the natural behavior of Lt & Rt speakers.
I further checked this perception via High passing Lt & Rt & Sub at 45Hz so the output range of both configurations will be close; but I can still note the change between configurations.

Now, I am a puzzled
What may be the source of this observation?
Are there any additional quantitative\qualitative parameters that I should look at and may shed light on that?
Is it an expected behavior where you compensate one feature (bass extension) with the another (soundstage)?
Am I hallucinating?

Looking forward for your good insights!

Two thoughts:

IME bass direction is perceptible down to 60hz or so. So to the extent your perception of space is based on bass, the sub might be pulling the image around.

Even though it is EQ'd, ringing modes from the sub could be masking some spatial cues, maybe.

I am not really confident in either of these explanations, just what occurred to me reading your post.

 

[ErLan]

Many thanks for stepping into the discussion, few clarifications:

Usually with a sub the soundstage is improved. Even if the Elac are only about 600€ a pair, that should work. Highpass them not lower than 80Hz, at 45Hz they are already stressed. Then match the sub to it. Get the phase right, this is the usual problem with a sub.

Fully agree with this statement, this is why I was surprised to find out that this was not the case in my experience

I don't quite understand.

You're crossing over the system at 80 Hz, and say "sub integration is only extending the natural rolloff of the speakers (45Hz >> 30Hz),"

According to these Klippel measurements, the 'natural roll off ' of the Carina appears to be well above 30- 45 Hz. More like 60Hz.

The spinorama measurement (spinorama.org) points out cutoff of 45Hz; this value is aligned with my in-room measurements.
By 45Hz >> 30Hz I meant that, in the Rt & Lt configuration roll-off take place at ~45Hz, when while using the sub roll-off frequency is ~30Hz.

Anyway, I'm guessing you need to do more comparative measurements of with vs without sub, from the MLP. Perhaps post them here for diagnosis. It would seem to be something happening in the room mode range. You are correcting up to 300Hz, have you tried restricting that range further?

I will consolidate all measurements in one compact file and upload it. Room mode is likely different due to frequency range and sub position; so this may be a possible source.

IME bass direction is perceptible down to 60hz or so

I was under the impression that the perception range is <100Hz (so 80Hz is 'inside'); is it this low? If so, it may strengthen the 'stereo bass' approach (that seems to be quite debatable) - where is the value of 60Hz coming from?

Consolidating some of the hints, a reasonable hypothesis is that the setup with sub generate room modes that are different than the setup with no sub, which may affect the stereo ques that support the sound stage.

Are there any works that correlates modes and soundstage that goes beyond general conclusions.

(*) As expected, my measurements do look different in the low end (30Hz-120Hz) due to the contribution of the sub in this range, but is there a way to better pinpoint which of the differences impair sound stage? or is it only 'in one's brain'?

 

[kemmler3D]

I was under the impression that the perception range is <100Hz (so 80Hz is 'inside'); is it this low? If so, it may strengthen the 'stereo bass' approach (that seems to be quite debatable) - where is the value of 60Hz coming from?

I am not making any specific claims about *how much* stereo is perceptible, but I think vague direction is perceptible lower than people normally say. I have stereo subs in my office. I tried listening to a tone ping-ponging between the subs, and the lowest frequency where I could convince myself there was any directionality was around 60-65hz.

While this is not a super-scientific test on my part, it's been studied and there seems to be evidence that you can still detect direction in that range and lower, although resolution suffers as frequency decreases. https://aaltodoc.aalto.fi/server/api/core/bitstreams/e43fa993-e3c9-4784-a6c2-05a292aab188/content#:~:text=Results indicate that humans can,for reaching the MAA threshold.

I think it's plausible that adding mono, dislocated bass around 80-100hz could throw off stereo image, a little.

FWIW I wouldn't say stereo bass is worth it for the stereo effects. Most recordings don't have much stereo separation (if any) in sub frequencies so even if you can hear it, it doesn't come up very often / ever in music. I am never sitting here thinking "wow, such stereo width in the bass!" It just happens that putting my desk in the middle of two subs makes sense for the room.

 

[Old_School_Brad]

To troubleshoot stereo imaging issues, start by checking channel separation. Play test tones through each speaker one at a time to confirm that only the intended speaker produces sound -left channel tones should not be audible from the right speaker, and vice versa. This helps rule out any crosstalk that could blur the stereo image.

This should be done with and without subwoofer to identify any issues that the use of subwoofer and crossover could introduce.
You might want to take a sweep with REW and look at phase and impulse response as well to see if anything odd is introduced with the subwoofer engaged.

 

[Chris A]

(I just stumbled onto this subforum. The breadth of my reading here is limited.)

My current search is looking at sound perception of Left & Right vs. Left & Right & Sub configurations...

• When using Lt & Rt only, the sounds seem clear, and the holographic perception is very good (3D, very detailed);
• When engaging the Sub, the bass is more present (naturally), however it looks like the holographic view collapses from 3D to 2.5D, mostly impacting the soundstage depth.

I've done many REW measurements over the years in-room (12 x 4.7 x 2.7 metres--152 cubic metres total volume, RT30 is generally ~0.4s from 100 Hz to 6 kHz). The room's Schroeder frequency is ~100 Hz. In other words, the low frequency performance of the room likely allows me the hear sound directivity at a lower frequency than many home listening rooms.

What I've found with subwoofers is that it's really their higher frequency harmonics (i.e., harmonic distortion) that largely gives away their location in-room subjectively. This is an important observation since the goal is to not hear their location(s) but only their added low frequency extension, as if your front loudspeakers have wider frequency response.

If you suppress the level of subwoofer harmonic distortion at higher frequencies (including and especially harmonic distortion) and achieve subwoofer time delay/phase coincidence with your front left-right-center loudspeakers in the crossover band, the subwoofers will disappear acoustically such that you can't hear their locations (generally below the room's Schroeder frequency). This also means that you cannot hear any soundstage/imaging effects that the subwoofers may bring to the total.

__________________________________________________________________________________

So now the clue is that your subwoofer is likely producing frequencies that are too high (including its harmonics) and perhaps not in phase/time delay with the front loudspeaker direct arrivals, thus affecting the overall soundstage/imaging depth.

1) The first experiment I would try is to reduce the crossover frequency of the subwoofer relative to the front loudspeakers.

2) The next would be to reduce the amplitude of motion of the subwoofer's cone--by adding more subwoofers (one or more), which disproportionately reduces nonlinear harmonic distortion amplitude due to halving the required woofer motions.

3) The third experiment would be to add absorption material and a physical barrier around the subwoofer (leaving the lower frequencies to still transmit into the room to attenuate your subwoofer's higher harmonic distortion frequencies above ~100 Hz. This could be as simple as placing the subwoofer(s) inside a large tube trap. You should not be able to see the subwoofer's diaphragm from your listening position. Make sure that the subwoofer's measured relative phase/time delay relative to the front loudspeakers is readjusted to provide in-phase/time alignment performance.

There are perhaps other experiments that could be tried.

Chris

 

[krabapple]

In a 'typical' system using a subwoofer, there will already be filtering applied either in the AVR (from two sources : its 'LPF for LFE' setting, which throttles the bandwidth of the LFE channel particularly, and 'crossover' setting, which bass manages 'small' speakers) and in the subwoofer itself (its variable 'crossover').

settings for these are typically user-adjustable but the recommended 'defaults' are usually:
LPF for LFE (bass from LFE track of the source) : 120 Hz
AVR crossover (bass from 'small' speakers): 80Hz
subwoofer: 'off' or set to 'maximum' (250Hz)

An AVR bass management crossover set to 80 Hz isn't going to generate much subwoofer location signal. The LPF for LFE is another story.

The LFE track was invented for very loud, often brief, bass content in movies, like gunshots, eruptions, etc. The Dolby standard for LFE bandwidth is 120 Hz.
Mix engineers shouldn't be putting broader bandwith content that than, in the LFE.
But they certainly do, especially since, for no very good reason, LFE tracks are included in multichannel music releases. There, I have even seen full-range content put into LFE.

Since even 120 Hz includes 'location signal' frequencies ( i.e., >80 Hz) there is always a chance the content itself will reveal the subwoofer. For movies this should be only an occasional risk. For multichannel music releases, maybe more often. If one is often concerned about it, enforce an ~80Hz LPF/crossover from AVR, or at the subwoofer itself.

 

[neRok]

What I've found with subwoofers is that it's really their higher frequency harmonics (i.e., harmonic distortion) that largely gives away their location in-room subjectively. This is an important observation since the goal is to not hear their location(s) but only their added low frequency extension, as if your front loudspeakers have wider frequency response.

If you suppress the level of subwoofer harmonic distortion at higher frequencies (including and especially harmonic distortion) and achieve subwoofer time delay/phase coincidence with your front left-right-center loudspeakers in the crossover band, the subwoofers will disappear acoustically such that you can't hear their locations (generally below the room's Schroeder frequency). This also means that you cannot hear any soundstage/imaging effects that the subwoofers may bring to the total.

__________________________________________________________________________________

So now the clue is that your subwoofer is likely producing frequencies that are too high (including its harmonics) and perhaps not in phase/time delay with the front loudspeaker direct arrivals, thus affecting the overall soundstage/imaging depth.

I agree. My first suggestion is that the sub and speakers might be in phase at the crossover frequency, but that doesn't mean they are in time. You need to measure everything individually whilst using an acoustic timing ref, and then check out the impulse alignment, like I describe here. If there's still issues after that, then I would be looking at the "phase slopes", because they probably aren't parallel, which means they are drifting out of sync. And if that's good, then you want to look at the filter order (freq slopes) and position, paying attention to things like Chris mentions.