I would define "image quality" as the ability of a system to clearly and unambiguously present a phantom "image" in the sound scene and to do this without any blurring or shifting with signal level or the sources pitch.
It is a hard thing to quantify, so hard, in fact, that decades ago Ford Motor commissioned a research study at Waterloo University in Canada to come up with a metric of this. The report was interesting and comprehensive, but the metric was not felt to be effective. This work was all published in JAES.
If the phase were presented as well as the magnitude then I could exactly produce the decay.
I could also approximately produce the decay by first finding the minimum phase response from the magnitude by using a Hilbert Transform, and then use that to construct the impulse response and thus the CSD, but this procedure would not necessarily be exact. It would be only as accurate as the speaker actually is minimum phase, which most speakers are.
May have to do with delay time and amplitude. Delays less than about 20ms cause, or are perceived as, comb filtering. They can also affect perceived virtual source position. More than that and they start to be perceived more as delays. Wait long enough and there are enough reflections bouncing around to make it sound like reverb.
I thought that "first few" was self explanatory, but maybe not. In most situations where the speakers are setup to be away from the front wall, the first reflections are going to be off the nearby walls, ceiling and floor. Discounting the floor and ceiling for now as they should be dealt with, but are not major conflicts with imaging, which leaves the sidewall reflections. These can arrive with in the summing localization time and will strongly interfere with a stable image. So to me the first sidewall reflections are the ones to be concerned about, but front wall reflections can also be a problem if they are not dealt with. As you can see, there are just a "few" reflections that I am talking about, and they arrive "first" before all others, so "first few" made sense to me.
Years ago when I was experimenting with OB design I found that simply producing a slight curve to the front of the baffle (so that it was slightly concave) it had the effect of focusing the acoustic image of the sound somewhat without it affecting the rear radiation into the room. This gave the OB that more 'in your face' presentation that many box designs have with which we grew up with. The baffle can have a subtle effect on the presentation of sound even if we tweak the life out of the frequency response, dispersion most definitely has its effects and can be manipulated to good effect if not overdone.
C.M
C.M
OTOH see the "Archimedes" EU-funded scientific research project: EUREKA Project E!105 ARCHIMEDES - EUREKA
It was a joint project of Danmarks Tekniske Universitet, Bang & Olufsen and KEF led by Prof. Soren Bech.
From one of the member of the test panels:
see: A round table about room acoustics, and bass traps.
Last edited:
I agree that floor reflections must be handled, they are disruptive, but they do not interfere with imaging in the same way that lateral ones do - image is predominately a lateral effect. And once the floor is handled, the side walls present the next problem. I do the ceiling with an absorptive grating and the floor with a thick area rug between speakers and listener, and then a coffee table to obstruct the remaining reflection as well. Toole says to leave the side walls reflective, OK I agree, but then the very first reflections must be handled through directivity if a strong image is to be maintained.
As far as imaging, the stereo mix itself has the biggest influence. But if you want to design a speaker with good imaging think of a bat's sonar. It uses sonic reflections to "see" in the dark.
The speaker drivers themselves illuminate the baffle and nearby objects with sound. Thus eliminate these illuminations and get great imaging.
There are many approaches. One is to have more directional speaker by any number of means such as beamier drivers or waveguides. Another is to cover the baffle is sound absorbing material. I did this for the ones I have now and they image better than anything I've ever heard. I believe this is the best for near field listening, however the material will absorb some frequencies more than others which may require equalization.
The other approach (perhaps a bit more fancy) is to build a sleek cornerless shaped speaker that eiminates diffraction and early reflections. However, these speakers will have problems without plenty of space around them.
Arrays for example need lots of lateral space to sound right because of the high lateral dispersion, but the floor and ceiling don't matter as much because of the lobing effect. So it all depends on the room and what you want, more than one way to do it.
Early reflections will make any speaker producing high frequencies sound bad.
Last edited:
Would a coffee table not be an added early reflection? I would think anything between the speaker and listener would be just as bad (or worse) as having a bunch of stuff around the speaker.....tv, furniture etc. This is something I go around and around with. Sometimes when I listen I slide the coffee table out of the room (easy to do). That said in room measurements from the lp show a bit smoother response from 100-1khz.
Described in here is a rather extreme way to deal with floor bounce as promoted by Tom Hidley.
The BOP Studios Story | (scroll to bottom)
The BOP Studios Story | (scroll to bottom)
Member
Joined 2009
Paid Member
That's a lot of work, most folk can't have such fancy rooms. But not to worry. In the future, we'll have our iphones implanted directly into the skill, providing audio and picking up audio. Then we'll be able to listen to music directly beamed to our heads. Speakers will be gone and no more compromises.
When talking about measurements, two things are getting confused.
1. the effectiveness in communicating meaningful information to humans, and
2. the abstract data contained in the display.
BTW, The mathematically inclined can tout the wisdom and purity of this or that display and how it is exactly equivalent to some other display and to reality. But there are various "simplifying" assumptions always made about the math and about how the real world works that are hidden.
Geez, nobody but motional feedback fans care about Group Delay? I was fooling around last night with MF and ran REW with my mic hanging a few inches over a gross car woofer with a convenient dual voice coil*.
Here are GD's, one with a few dB of feedback from the second voice coil and one without. Which do you like better?
B.
*parts count for device needed to create feedback circuit (if you already own a mic mixer): zero
1. the effectiveness in communicating meaningful information to humans, and
2. the abstract data contained in the display.
BTW, The mathematically inclined can tout the wisdom and purity of this or that display and how it is exactly equivalent to some other display and to reality. But there are various "simplifying" assumptions always made about the math and about how the real world works that are hidden.
Geez, nobody but motional feedback fans care about Group Delay? I was fooling around last night with MF and ran REW with my mic hanging a few inches over a gross car woofer with a convenient dual voice coil*.
Here are GD's, one with a few dB of feedback from the second voice coil and one without. Which do you like better?
B.
*parts count for device needed to create feedback circuit (if you already own a mic mixer): zero
Attachments
Last edited:
fat chance there will still be audiophiles that will contend that the differences in size,shape and thickness of the human skull will result in perceptual differences...
Last edited:
This thread goes all over the place. But point source vs array would be an interesting discussion.
I still think mtm or wmtmw type speakers can sound the best, because they cut down on floor and ceiling reflections and also the driver pairs cut down on distortion.
But they really need a larger space and distance.
I still think mtm or wmtmw type speakers can sound the best, because they cut down on floor and ceiling reflections and also the driver pairs cut down on distortion.
But they really need a larger space and distance.
In many instances that I have seen the coffee table is a major reflection, so I share your concern. But in my instance the table top surface is deliberately small, so it acts something like a diffusor. But I don't think that something far out into the room would be as bad a diffraction problem as something close by - could be wrong about that I suppose! I can move mine as well and sometimes do, but there is not a huge effect. If it were optional then I would tend to say not to have a coffee table, but since my room is a home theater with usually more than one person, a place for drinks, feet and controls is kind of a must.
Your not thinking of those being an "array" are you? Because they more closely simulate a point source.