What creates good imaging?

[jaydillyo]

I am on the hunt for new speakers. I am currently running some old MB Quart QLC-404s. I picked these speakers up a long time ago for pretty cheap when I was much poorer. These speakers have served me well for about 8 years, but a substantial update is long overdue.

I have been auditioning speakers locally. Today I listened to some GoldneEar Aon 2s and 3s. They both sounded way better than my old MB Quarts, but the number one thing that jumped out at me was how well they imaged.

When I was listening to the Aons a trumpet sounded like it was dead center of the soundstage. On my MB Quarts its just a giant wall of sound in which its hard to pinpoint where things are in the soundstage. Which leads me to my question.

What speaker qualities create good imaging?

-- jaydillyo

 

[Steve81]

One thread on the subject:
http://forums.audioholics.com/forums/general-av-discussions/82268-imaging.html

 

[jaydillyo]

Thanks for that link. So it sounds like imaging is more function of the listening room and speaker setup than the speakers themselves. That's unfortunate if that's the case. It's a lot easier to change speakers than the room.

-- jaydillyo

 

[jaydillyo]

So, an interesting note. I just picked up a Sony bdp-s185 blu-ray player and was using this for Audio CD playback. It's awful and is the source of a lot my imaging problems. When I streamed the same song over my Apple TV it sounded so much better than when played on the sony. I haven't even watched a blu-ray on it yet, but I'll be returning it to Best Buy tomorrow. The CD playback is that bad. I didn't think it would be that hard to convert CD audio to lpcm and stream it over HDMI without ruining the sound. Apparently I was wrong.

-- jaydillyo

 

[jaydillyo]

Another follow up on this. My Apple TV and my MacBook Pro (iTunes) sounds much more open than the Sony bdp-s185 and my PS3 do. I'm left wondering which is "right". The sound from the bdp-s185 and PS3 just seem a little bit muted in the highs in comparison.

I apologize that this thread is getting away from the original topic. I'm just surprised at the differences that I'm finding.

-- jaydillyo

 

[j_garcia]

A good speaker designer creates good imaging.

 

[jonnythan]

A good recording and, I would guess, speakers that are low-resonance with good waterfall plots. I would think that the thing that ruins imaging is sloppiness - that is, sounds don't roll off fast enough.

 

[AcuDefTechGuy]

I think the actual recording & an accurate speaker with smooth off-axis response.

 

[GranteedEV]

When I was listening to the Aons a trumpet sounded like it was dead center of the soundstage. On my MB Quarts its just a giant wall of sound in which its hard to pinpoint where things are in the soundstage.

My guess would be inter-driver phase differences.

 

[JerryLove]

My guess would be inter-driver phase differences.

Let's assume that relative phase affected imaging.
Let's also assume that the drivers were out of phase.

That would affect imaging across the crossover point.
It would not affect imaging entirely on one side or the other of the crossover.

So falsifiable test for our hypothetical model: is the imaging issue only occurring on sounds which cross the crossover point?

I think room interaction is a far more likely cause.

 

[jaydillyo]

Well, I picked up some Aon 3s today. Local store is being shuttered and I got them for 40% off msrp. Couldn't pass it up.

These Aons definitely image better than my MB Quarts. No doubt about it. They are just better speakers. I'm sure there are some things I can do to my room to make things even better.

-- jaydillyo

 

[GranteedEV]

Let's assume that relative phase affected imaging.
Let's also assume that the drivers were out of phase.

That would affect imaging across the crossover point.
It would not affect imaging entirely on one side or the other of the crossover.

Alright, then let's assume a common 1st order transitioning to 3rd order crossover centered at 3khz. At 1.5 khz your tweeter is only 6db down, and at 750 hz your tweeter might still be 18-20db down which may still be audible in the form of phase cancellation. The same applies to the upper limit - at 6khz your woofer is only 6db down, and it's also probably going through damped diaphragm break up. At 12khz your woofer is still contributing at -12db of the response.

So from about 1khz to 12khz you've got two drivers with enough output to affect imaging, and they're varying between being about 90 degrees out of phase to as much as 140 or more degrees out of phase. For phase quadrature summation like this, there isn't even a real great way of verifying "desired" phase tracking (although this is possible during a Linkwitz-Riley type slope which is exactly 180 or 360 deg out of phase - thus reversing polarity would give a strong reverse null.

So my question is, is the crossover for that two way a narrow band, or a very wide one?

Now consider a 2.5-way, 3-way, or even 4-way.

 

[cpp]

Good imaging as it pertains to non concert music a.k.a. in person (live) starts with a good recording then good room acoustics then a good set of speakers that can replicate the intended sound correctly as recorded.

 

[JerryLove]

So from about 1khz to 12khz you've got two drivers with enough output to affect imaging, and they're varying between being about 90 degrees out of phase to as much as 140 or more degrees out of phase. For phase quadrature summation like this, there isn't even a real great way of verifying "desired" phase tracking (although this is possible during a Linkwitz-Riley type slope which is exactly 180 or 360 deg out of phase - thus reversing polarity would give a strong reverse null.

24db-octive is more common (except in speakers *trying* to get time coherent).

phase cancellation looks just like non-flat frequency response. So you can't cite that as a cause of imaging unless you want to include both causes.

So yes. Your flutes should image well and your human voices poorly in the above scenario if phase is the problem.

 

[Chu Gai]

Look to eliminate the earliest - not necessarily the first - reflections.

 

[KnarlyCharly]

I believe what creates good imaging is the speakers ability to produce a good off axis response, in other words, it's ability to produce soundwaves well beyond it's rated range. Not only off axis, but the design inherent to input tuned environment in a said anechoic chamber. Another factor is the physical aspect of the cabinet design being mirror imaged of one another. On the other hand, soundstage is another matter. Most of our hearing is in the middle range of most capable speakers anyways. Did it ever occur to you that due to the age of your older speakers, that the crossovers might be due for a rebuild? I have seen it one too many times, speakers sound poor due to age and deterioration, only to evoke a gaping response of amazement after I restore them back to original sound paramaters. Some cases I've even tweaked them a little with even better results. Don't write them off just because they don't sound quite right.

 

[GranteedEV]

24db-octive is more common (except in speakers *trying* to get time coherent).

Is it? It's actually very expensive and generally seen in more superior speakers that... image well.

there is no single dominant choice of crossover rate or frequency. That's why speakers sound so different.

phase cancellation looks just like non-flat frequency response. So you can't cite that as a cause of imaging unless you want to include both causes.

Phase cancellation does not necessarily appear as non-flat frequency response on a single axis. A 3rd order butterworth slope has drivers summing such that they are individually 3db down in level at the crossover frsequency. If they summed in phase, this would cause a 3db peak in the frequency response (given that two sources playing the same signal will create a +6db summation in response compared to a single one). They already sum 90 degrees out of phase to create a flat-looking frequency response where the drivers are individually 90 degrees apart. Another 30 degrees (or 120 deg out of phase) would not neessarily affect response notably - but it may be audible.

 

[JerryLove]

Is it? It's actually very expensive and generally seen in more superior speakers that... image well.

The Infinity Primus 163 uses a 24db crossover: Primus 163 - Infinity

That was the one I was using as an example, and the first I looked up. It's less that $100 / pr. How is that "very expensive"? I tried looking up the Pioneers that are popular, but I don't see the crossover slope listed.

there is no single dominant choice of crossover rate or frequency. That's why speakers sound so different.

Oh? I though cabinet resonance, off-axis performance, cross-driver dispersion, and frequency response might have more to do with it than crossover slope.

Phase cancellation does not necessarily appear as non-flat frequency response on a single axis. A 3rd order butterworth slope has drivers summing such that they are individually 3db down in level at the crossover frsequency. If they summed in phase, this would cause a 3db peak in the frequency response (given that two sources playing the same signal will create a +6db summation in response compared to a single one). They already sum 90 degrees out of phase to create a flat-looking frequency response where the drivers are individually 90 degrees apart. Another 30 degrees (or 120 deg out of phase) would not neessarily affect response notably - but it may be audible.

If it's flat where you are listening then it's not happening where you are listening. If it's happening where you are listening, then it's not flat (unless it's lucky enough to be exactly countering some other issue) where you are listening.

 

[monkish54]

The Infinity Primus 163 uses a 24db crossover: Primus 163 - Infinity

That was the one I was using as an example, and the first I looked up. It's less that $100 / pr. How is that "very expensive"? I tried looking up the Pioneers that are popular, but I don't see the crossover slope listed.

My guess is they are talking about 4th order acoustic slope.

Open PCD and play with two drivers you think will mate well together. Set the high pass and low pass as 4th order electrical. It probably won't be so cheap. Cost can quickly get out of hand when using 4th order electrical, especially when working with a budget.

The Pioneer XO only uses 6 components, so it can't be 4th order...least not high pass and low pass. My guess is (again with the guessing) Andrew used a 2nd order electrical for both the high and low pass, and an Lpad for the tweeter. 2+2=4 4+2=6 6 components.

 

[3db]

Let's assume that relative phase affected imaging.
Let's also assume that the drivers were out of phase.

That would affect imaging across the crossover point.
It would not affect imaging entirely on one side or the other of the crossover.

So falsifiable test for our hypothetical model: is the imaging issue only occurring on sounds which cross the crossover point?

I think room interaction is a far more likely cause.

Maybe I'm over simplifying this a little but wire one of the speakers out of phase and see if you can correct it with acoustic treatments.