Revisiting "Imaging"

[marco_gea]

On a recent thread: http://www.audioheritage.org/vbullet...ded-crossovers one of several divergent topics from the original topic dove into the discussion of imaging.

Like everything else in this hobby, there are many opinions on how to get great imaging. That said, I'm not even sure we are all taking about the same thing when we discuss imaging.

Based on some of the responses in this very thread, I doubt we all are (talking about the same thing)...

I define imaging as the the speaker's soundstage. In my opinion for there to be imaging you typically do need to have a setup where you can sit at an equidistant vertex of a triangle. I realize for many this is simply not possible. While it is still possible to have a very satisfying musical experience I don't think outstanding imaging will be possible.

100% agreed (bold emphasis mine).

I would add that a true test for the ability to reproduce such an illusion of "soundstage" must start with a recording where natural, low-level acoustical cues of the original recording space are present in the first place. Multi-mike, multi-tracked, artificially pan-potted recordings - such as e.g., The Dark Side of The Moon mentioned in another post in this thread - can certainly sound immersive and impressive in their own right, but they aren't the right tool to ascertain a system's ability to truly "image", in my opinion.

Marco

[marco_gea]

Absolute symmetry of the location of the L & R loudspeakers is important for imaging.

Yes, this is always true.

This includes toe in, front and side wall distance.

The importance of equal side wall distance depends on the horizontal directivity of the speakers, especially in the mid range. The higher the directivity, the less sensitive the speakers' imaging potential is on the proximity of the side walls.

Balance of the individual drivers levels and source of L & R levels is also important.

A little known fact is that the L & R level of a vinyl cartridge can vary +- 1.5 dB. Careful adjustment of a tone arm and test equipment can minimise this.

Tracking of the volume control L & R can also impact significantly on the imaging.

Yep!

[marco_gea]

In my listening history, one pair of small single driver speakers imaged so well they blew away all that came before. They are Omega TS2Rs, a model with 6.5 inch Fostex drivers that came out around 2002. At 94dB/watt/meter efficiency they are not difficult to drive. I first heard them powered by a small tube amp and later by a First Watt F2J. (The imaging was similar but the bass response was extended much lower with the latter, being a transconductance design. Everyone who has heard the TS2R/F2J combination asked about what subwoofer was being used. None was, of course.)

...

For the well-heeled I can recommend an equally rewarding but different experience, the MBL Radialstrahler 101 E MK II speakers set up in a good room. Their omnidirectional presentation is perfectly executed and the sound field is so thrillingly immersive you forget all about the usual imaging impressions and considerations. ...

I agree on the single driver experience. Good imaging can be one of the strong points of well-implemented so-called "full range" drivers, especially those without "whizzer" cones, which create all sorts of phase issues of their own.

I strongly disagree about the MBL's, though. I have heard them many times, and they invariably superimpose the same overly spacious and "reverberant" (for lack of a better way to descibe it) rendition on pretty much any music that is fed to them. To me, that is not good imaging, but fake imaging. From a technical standpoint, I think it results from excessive early reflections from the listening room's walls swamping the low-level spatial cues that are present in the recording.

[marco_gea]

My low tech guess about why my 4345s suck pond water at imaging is that their baffles are too damn big to pull it off. ...but from what you say about the M2s, perhaps not. I have noticed that when sitting on the floor in front of them, the stereo image kicks in incredibly close to the baffles. Not much over a foot.

It has much more to do with their relatively crude crossover, which does not pay much if any attention to inter-driver phase tracking.
The width of the baffle per se is only of secondary relevance, especially with high-directivity mid-hi frequency drivers such as those.

[marco_gea]

There is no soundstage behind soffit mounted speakers though.

With all due respect, I think this is an unsubstantiated blanket statement.

What soffit mounting precludes is additional artificial "spaciousness" created by in-room reflections from the front wall. But this has little to do with properly reproducing the soundstage cues that are present in the recording (if they are there to begin with, of course!)

[marco_gea]

In my experience this spacing works best with most speakers. Most of my systems ended up in this range even if the room is huge and the listener can be back further and there is ample room to push the speakers wider apart. I have found I prefer 9-10' between the speakers and that same distance to the listener.

I'm not sure if this is a personal preference or a more universal thing.


Widget

FWIW, I found the same. Sitting at the apex of an approximately 3-metre side equilateral triangle is the "sweet spot" in speaker-listener positioning.

[marco_gea]

Based on my experience with a) 4345's; b) numerous 2- and 3-way direct driver speakers in indifferent boxes; c) two-ways with a fifteen or fourteen crossed in the hundreds to a "windswept" horn on top of the woofer cabinet; d) small full ranges; e) Altec 604's; f) Tannoy/Manley time aligned ML10's vs the same 2558 D.C. driver in the stock, non-time aligned SGM 10B, I would say good imaging depends on the speakers tracking frequency response and also tracking coherent phase response, absence of diffracting edges and good placement with reflection management. There are multiple ways to get there.

Yes!

[rusty jefferson]

With all due respect, I think this is an unsubstantiated blanket statement.

What soffit mounting precludes is additional artificial "spaciousness" created by in-room reflections from the front wall. But this has little to do with properly reproducing the soundstage cues that are present in the recording (if they are there to begin with, of course!)

I feel the point is, the "spaciousness" and localization of performers is not artificial. It's time delay of the instruments or voices to the microphone(s) and reflections in the theater environment also picked up by the microphones. Played back on soffit mounted speakers, there will be no reproduction of the "actual" instruments/voices in space in the proper location, behind the loudspeakers, as the performers were behind the microphones.

Please see post #16 about the Stereophile Mapping the Soundstage track. Playing that track back on soffit mounted monitors will simply be a left to right pan on plane with the loudspeakers.

[Robh3606]

Multi-mike, multi-tracked, artificially pan-potted recordings - such as e.g., The Dark Side of The Moon mentioned in another post in this thread - can certainly sound immersive and impressive in their own right, but they aren't the right tool to ascertain a system's ability to truly "image", in my opinion.

Hello Macro_gea

Maybe not but they can make excellent set-up tools from system to system and typically if you can get the best out of them a more natural imaged piece is quite easy to hear. I use a couple of specific recordings that I know well to set-up my speakers in my systems. I also find that with that type of recording the imaging changes from song to song from the differences in mix down from song to song. I agree about speakers that superimpose there characteristics by design over recordings. In my mind they should all be different especially in multitrack recordings and it should be easy to hear the differences from song to song even on the same CD/Album.

Rob

[1audiohack]

I feel the point is, the "spaciousness" and localization of performers is not artificial. It's time delay of the instruments or voices to the microphone(s) and reflections in the theater environment also picked up by the microphones. Played back on soffit mounted speakers, there will be no reproduction of the "actual" instruments/voices in space in the proper location, behind the loudspeakers, as the performers were behind the microphones...

I don’t mean to be contrary but if the cues are in the recording, anything that wraps around the speakers and comes back to the listener must be considered an alteration and an addition to the recorded original.

It seems to me that if a speaker pair can create an image or perceived placements of sounds wider than themselves and deeper than themselves that information should be in the time and space cues of the material and as long as the acoustic space is inert like outdoors or soffit mounted in an infinite baffle, or in a symetrical room properly treated, the effect should be detectable.

If you hold that the sound must interact with the space, then it’s the space.

I understand how an asymmetrical space will damage the effect.

If this is not so, please tell me how and why?

Again, I am not attempting to be contrary but to understand.

Thank you all.
Barry.

[1audiohack]

I did not mean to kill the conversation!

Should I delete my questions and observations and just spectate?

Barry.

[rusty jefferson]

Hey Barry,

I was thinking how I could explain the physics to your questions/points, but honestly, I'm not smart enough. So here's a link to a page at the Linkwitz website that may help explain it properly. http://www.linkwitzlab.com/The_Magic/The_Magic.htm I'm not promoting Linkwitz products, but thought him to be a reliable source.

The room is a very important factor for imaging. Properly time delayed reflections are critical as are many of the things previously discussed, great off axis response (including dipole), low diffraction cabinets (or no cabinet), wide dispersion as from the small point source, etc. To hear the instrument 4-5 ft. beyond the width of the speakers as Clark mentioned, or JA banging the cowbell on the left/right side of the stage on the test track showing as being behind the loudspeakers, you need delayed reflections. It won't happen in an anechoic chamber or with soffit mounted monitors (especially in an acoustically dead studio setting). The sound will be "stuck on the speakers" if you will. They will not disappear.


Here's a paragraph from the link above, and this quote from another section. I like how he refers to imaging as the magic, as you did in an earlier post.
"The magic is difficult to describe in pictures or words but is recognized within 30 seconds when heard. It usually elicits a big smile or even laughter from the listener. Naive listeners, audiophiles and professionals alike recognize the naturalness of presentation. On many recordings it is 3D in front of the listener and resembles a concert experience..."
3.2.1 Loudspeaker & Setup Requirements
For the magic to occur there are five requirements, which I have found to be essential:

1. The off-axis frequency response of the loudspeaker in the horizontal plane must mimic the on-axis response. The vertical polar response is not as critical but the formation of lobes should be avoided, i.e. the loudspeaker should be acoustically small. Such a loudspeaker has a neutral signature and its reflections and the reverberated sound field will have the spectral signature of the listening room. The polar pattern could be omni-directional, cardioid or dipolar with frequency independent power response. The omni pattern would produce the strongest interaction with the room and be the least desirable of the three.
2. The loudspeakers must be free of resonant radiation such as coming from vents or panels. Non-linear distortion must be low enough not to identify the speaker location during loud music passages. This demands adequate volume displacement capability of woofers and tweeters.
3. The loudspeakers must be set up at least 1 m distance from left and right side walls and 1 m from the wall behind them. The resulting time delay of about 6 ms is necessary for the ear-brain perceptual apparatus to separate direct from reflected sound streams. With neutral excitation of the room response, the listener can then withdraw attention from the room and process primarily the direct sound streams from the loudspeakers. This is similar to not hearing the ticking clock or to the cocktail party effect where attention and hearing is drawn to information streams of interest. Everything else is moved beyond the acoustic horizon as in survival mode.
4. The wall behind the loudspeakers is preferably diffusive and the wall behind the listener absorptive. The sound waves from the loudspeaker should be allowed to travel freely past the listener and not be reflected from a wall behind.
5. The listening room should have a reverberation time around 450 ms, i.e. be neither acoustically dead nor overly live but comfortable for conversation and entertainment. If anything it should err in the direction of liveliness.

[marco_gea]

To hear the instrument 4-5 ft. beyond the width of the speakers as Clark mentioned, or JA banging the cowbell on the left/right side of the stage on the test track showing as being behind the loudspeakers, you need delayed reflections. It won't happen in an anechoic chamber or with soffit mounted monitors (especially in an acoustically dead studio setting). The sound will be "stuck on the speakers" if you will. They will not disappear.

Sorry, I don't have time for a complete response to your long post. But I can assure you that this particular assertion is demonstrably wrong.

For instance, the first track on Roger Waters' "Amused to Death" has a barking dog that's supposed to sound as coming from far away, and way outside of the triangle defined by the physical position of the speakers (I forget now whether further to the right w.r.t the right speaker, or to the left w.r.t the left one).

I have listened to that track on a large number of systems, including soffit mounted ones, as well as on my own, where the speakers are floor-standing but positioned right up against the front wall. And I can assure you that - with all speakers having properly designed crossovers - the dog always sounds exactly where it is supposed to sound, i.e. behind and outside of the triangle defined by the physical position of the speakers.

Marco

[Ian Mackenzie]

Hi Rusty and Marco,

This is a good discussion and l found a few useful links on imaging techniques as applicable to recordings and loudspeakers below.

In the last link reference is made to creation of stereo imaging in the “mix”.

I have no doubt that plugins used to enhance or create interesting stereo reproduction would be frowned upon by engineers devoted to traditional mike techniques of classical recordings. To that end l crave for well recorded classical music in the analogue domain.

I also post a quote regards my comments about the use of imaging in stereophonic recordings to creat interest.


Paul Wilson

https://audiophilereview.com/audioph...overrated.html

“Imaging makes recorded music sound decidedly more interesting. It calls the listener to become more involved with the music. It makes the listener almost think he or she is at a live concert. Spatial cues in terms of musical depth and width provide an added dimension of realism to recorded music.

Imaging has the ability to make music more interesting. Hearing a musical presentation made in a realistic fashion may transport the listener to a time remembered - maybe a favorite concert, or some memorable past event.”

https://en.m.wikipedia.org/wiki/Stereo_imaging

Quote

“The quality of the imaging arriving at the listener's ear depends on numerous factors, of which the most important is the original "miking", that is, the choice and arrangement of the recording microphones (where "choice" refers here not to the brands chosen, but to the size and shape of the microphone diaphragms, and "arrangement" refers to microphone placement and orientation relative to other microphones). This is partly because miking simply affects imaging more than any other factor, and because, if the miking spoils the imaging, nothing later in the chain can recover it.

If miking is done well, then quality of imaging can be used to evaluate components in the record/playback chain (remembering that once the imaging is destroyed, it cannot be recovered).“

https://www.waves.com/tips-for-wider-stereo-mix

I hope everyone finds this productive to the discussion.

Ian

[Ian Mackenzie]

This link provides basic information about the impact of acoustics on stereo image

https://www.google.com/amp/s/www.cru...acoustics.html

https://www.gikacoustics.com/audioph...oom-acoustics/