Why does imaging quality change so much with the recording?

[garyrc]

When I sit dead center, in the audiophile (egocentric -- "Hell with the rest of the family" as George Carlin says) sweet spot, all is pretty good and the soundstage is often amazing. But with a listening area of 3 to 5 non-bulky chairs, those in the side chairs are often shortchanged.

So, I'd sure like to know:

1) Why -- on some recordings -- do my Khorns seem to fall silent, with the music coming from all around them and in a nearly ideal curtain across the front of the soundstage (with pretty good depth, as well), while -- on other recordings -- the music seems to hug the Khorn boxes, or, worse, favor one side, with very little coming out of the other, unless the listener is dead center? This happens both with and without the center channel Belle hooked up; a great and realistic wall of space differentiated sound some of the time, and clumps around the Khorns at times, but when there are clumps with the Belle connected, the Khorn clumps move slightly toward the center.

2) Why does the correct (best sounding, nearest the imagined original) center channel volume for the Belle (via NAD's simple center mix in "Enhanced Stereo I") vary from turned completely off to a level equal to the Khorns (when on, this range is 12 dB) depending on the recording? Oddly, the best sounding Belle setting does not vary nearly as much with the position of the listening chair (across 5 skinny chairs, maybe a 3 dB variation), as it does with the recording.

My CDs now sport little stickers with the correct center channel setting.

If it helps, here is what we listen to, in order of frequency, starting with the most frequently heard: classical, jazz, movies, folk, rock, pop, and other. Movies suffer the least from the above problems, probably because of the way they are channeled in 5.1 and the like.

The center chair is about 12 feet from the center Belle, and about 15 feet from the Khorns. I usually use NADs "distance" adjustment to compensate for the different distances.

[DrWho]

1) This can be one of two things. Either your current configuration is flawed in some way presenting a false image (thus hurting the imaging of other recordings). Or the recordings that don't image well just plain aren't recorded well. The natural directivity of horns make them relentless when it comes to imaging (fewer side-wall reflections to emulate it).

2) I wouldn't blaim it all on this, but in the studio there are various ways to treat the summing of the stereo mix. You can do equal power, -3dB, -6dB, 0dB, and a few other "stereo pan laws". Ideally there is a 60 degree spread between the mains and equal power is implemented. As studios deviate from the "ideal" they can change their stereo pan law to compensate - this process is never perfect. And on top of it you have the issue of your mains having a 90 degree spread, which is the maximum before you start getting huge holes in the center image. And then one other factor is the personal prefrences of the recording artist. So all that to say, there is good reason to see +-6dB on the center channel volume - merely as the result of how studios are being configured. The reason multichannel sources tend to sound better is because they have standardized the playback system (so the recording engineer only needs to mix for one system instead of averaging out a bunch of systems).

What are the dimensions of the room that you have the khorns in? And what wall are they sitting on?

[pauln]

I think it might have a lot to do with the way the peice was recored/engineered/mastered etc. If you are listening to an orchestra, it is possible that the recording was done in one take with mikes set up to capture the whole thing - imaging might be pretty good.

Other recordings may be built up from multiple tracks recorded at different times, even in different studios. When YES made the Close To The Edge album, they would record a few measures, then stop to figure out what to do next, then do some more, then stop and think again. Somehow it all came together, but the thing is that the tracks of a "recording session" may be distributed in time as well as space, and the resulting compilation may contain elements that don't sound natural or image right.

Live recordings often have incredible imaging because the performance was done in one place at one time - a good example is the 1972 recording of the live performance of "Miles of Isles" by Joni Mitchell, San Francisco. A very clean and well recorded album - if this does not image well for you then DrWho's suggestion that you might have a problem with the room would be the thing to consider...

This raises the question - what well known albums are known to be well imaged? If we had a group of well imaged albums to work with, this might be a good place to begin sorting out system problems...

[PrestonTom]

I certainly agree with PAULN's comments, he has rasied a number of important issues.

I will add that the concept of imaging has a numer of aspects to it. The left -right distribution and the sweet spot (and it's fragility) are certainly the most important. But other aspects are a bit more subtle and also seem to vary with different recordings and setups & rooms. These would include the "compactness" of the individual instruments (this is frequenctly messed up with a poor recording and a poor room), the "precision" and "stability" of the instruments/vocals location (sometimes in a poor room the image will wander around). There are other aspects also that get into what some folks call "the soundstage". I find this a poor term (I never know what is meant precisely and different people use the term differently). I think the term "depth" may be better. I think this is also reflects the reverberation of the original recording (either from the original recording environment or what was added electronically). The last aspect, and this is very difficult to define, but you know it when you hear it, is whether the sound seems to come from the cabinets or have the cabinets in a sense "disappeared". This, I think, is very difficult to acheive and probably is determined by number of things. However the effect can be astonishing. I normally would not even bring it up since when you try and describe the sensation it ends up sounding like audiophile nonsense.

But back to the point, I agree, you should try and concentrate on the problem using well-recorded live music (not stuff "put together" in the studio).

Good luck,

-Tom

[Kudret]

Tom,

you should try and concentrate on the problem using well-recorded live music (not stuff "put together" in the studio

Do you have a recommendation for a well recorded music that can be used to check the soundstage. I am afraid, at times I might be using poorly recorded pieces in my tests and blaming my setup while the problem is with the recording.

In my understanding soundstage refers to the actual location of the instruments on the stage in 3-D. I have improved my setup to a point where I now have a 2-D soundstage on some recordings (which I consider good recordings). Having said that, I am still troubled by piano recordings which I find too difficult to localize. My sweet spot is also fragile for the most part. Vocals start to sound not right after a slight move to the left or right.

Considering what I had before, I am happier with my setup now, but I also know that there is much room for improvement. Next on my agenda are to build a pair of Type A networks with high tolerance v-caps and also install v-caps in my amps to make sure that both channels are equal. Tweeters and mid drivers also need a closer check to make sure that they are equally loud. Finally I am building false half-corners for my Khorns.

Kudret

[mas]

The folks above have made some very valid points.<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />

I think one has to examine the recording process itself to discover the fundamental flaw..

In assuming that imaging is to be preserved, we seem to have a preconceived notion that there was indeed some original event that is simply being transcribed to the medium which possesses 'an image'. And except in what has become a rare instance, this is a false assumption.

With multi-tracking, overdubs, remote recording and a myriad other techniques, not only are the individual parts not done at the same time, many are not even done in the same location! The individual pieces and parts are assembled into a whole. And any semblance of an image is an artificial creation.

There is one microphone/recording technique (ITE) that accurately retains all of the phase information and hence all of he acoustic cues for extremely accurate spatial recreation of the original integral event.

But unfortunately, in a composition comprised of many spliced pieces and parts, any resulting 'image' is a new creation of that which never was, as opposed to the recreation of an actual event. Thus, what most consider to be an 'image' is simply the skilled manipulation of the psychoacoustic elements sufficient to create an effective illusion.

P.S. If you are intent on verifying your system's ability to reproduce the spatial cues in a manner consistent with the recording process, the Prosonus SRD CD L.E.D.R test (track 51) will do this.

[PrestonTom]

Not all recordings are originally created in a small studio. We sometimes forget this.

Some are actually recorded in a performance venue (and not a rock band in a stadium).

Soundstage is still tricky, and I agree it is the added dimension of depth (although the term is frequently mis-used IMO). The perceived depth may or may not be accurate to the original performance and much of it is a function of reverberation that was present (or added).

For small groups of musicians I like the recordings made by Denon, Harmonia Mundi, Naxos and many other small labels from Europe. Many times these are actually recorded in small theaters, churches. When they do this, there are a minimum of mics and the when the sound is mixed there is little compression done on the levels or litttle artificial reverberation added. Instead, you will hear the acoustics of the instruments convolved with the acoustics of the hall.

Piano music is tricky and I agree about the "expansiveness" of the image. Part of this is due to the decision of whether to place the mic(s) close to the piano or further back (to get some of the hall's acoustics in the recording). Sometimes the image of the piano will be expansive due to room asymmetries in your living room. Although it is heretical, sometimes solo piano is best listened to through a single speaker or in mono. Yes, the imaging will be dead on and stable.

As far as "modern music" goes, the recordings by Steely Dan, Dire Straits, and Paul Simon, are frequently well-recorded. Jazz from the late 50s through mid 60s can be quite good on labels by Blue Note, Impulse and some others. The tape hiss will be annoying, but they are frequently nicely recorded - some of the artists became absolute legends.

Good Luck,

-Tom

[D-MAN]

There are probably hundreds or even thousands of variables in the way. Whether the transients and small signal information actually get to the recorded medium or not is a concern of course, cause if it ain't there, you ain't gonna hear it no matter how good your system is.

The recording of spacial reflections is not something that a lot of engineers are concerned with. The sense of depth and imaging is actually an artifact of how we perceive sound, and the chances are that the spacial references of the reproduced sound technically has little or nothing to do with the spacial dimensions of the actual space the recording was made in. Add to that some reverb is going to be artificial, and may have nothing to do with defining actual spacial information at all. We've all heard the reverb on some instrument or singer, but not on the drums, etc. How do our ears interpret that in the context of a soundstage? Not well, at least, not convincingly.

Then there is the room and the speakers that you listen in with its own absorption and reflection characteristics. Given that your (and our) rooms are subject to their own coefficients based on their respective dimensions, contents, and layout, it's a miracle that anything sounds good, isn't it? Add to that the various elements that effect the signal, electronics and wiring, networks and speakers, and that is virtually unaccountable.

I suppose that in the off chance that the recorded reverberation information (if there and ACCURATE) happens to fall within the limits of the particular room variable coefficients and speaker propagation characteristics within that area, certain recordings could reproduce the spacial information in some believable or convincing sense without the room variables altering it too drastically. Providing the math all lines up, which is not very likely. However, there would always be a window or threshold where the math lines up ENOUGH, even only in a certain portion or section of a recording...

However, you CAN artificially enhance the subjective effect of imaging by reducing the reflections from the walls by moving the speakers away from the walls, closer together and out in front of you, like we have all seen and experienced with the small footprint direct-radiating tower-style speakers so in vogue. Unfortunately, these speakers are going to compress the available dynamics in doing so. They tend to give you the "Orchestra on a Table Top" sort of soundstage. Small, compressed, but still a very stable soundstage. Compressed tends to say it all, doesn't it?

I think that we can easily agree that fixed-angle (45 deg) corner horns are not ever going to create a fabulous soundstage for a couple of important reasons, even if you are perfectly aligned with each speaker on a 45 degree axis from the listening position. There are still many, many variables in the way.

DM

[DrWho]

I slmost forgot to mention - try sitting on the floor...works wonders with khorns.

[ZAKO]

Imageing? I have one recording (RCA) that used 28 mics. If yoy can imagin mics on high 15 foot booms & stands pointing down at the bald heads in the orchestra, you can imagine what deapth your going to get. Some of the new recording schools have been giving bad advice.....The Tone Meisters in europe or even worse....Pick up a Deutsche Grammophon record a good example of multi mic recording . I have dozens of photos showing over miked situations.

[mas]

The fundamental problem is that the original complex phase relationships of the various sources (being recorded - not the mics!) relative to each other in a live recording are neither captured nor maintained throughout the mix process.

Without this all you have is a composite image of discrete images that are simply postitioned by a pan!

The Only recording technique that completely preserves this relationship is the ITE recording method. And this provides a 360 degree complete 3space environment where ALL of the acoustic cues relative to localization are completely maintained and reproduced. They are uncanny to experience.

http://www.livesoundint.com/archives/2004/aug/tech.pdf#search=%22ITE%20recording%20technique%22

[DrWho]

Don't you have to listen to ITE recordings with headphones?

I remember reading articles written by people that talk about how crazy it is when the listener moves their head during the recording - like it induces crazy vertigo for the listener.

[PrestonTom]

Don't you have to listen to ITE recordings with headphones?

I remember reading articles written by people that talk about how crazy it is when the listener moves their head during the recording - like it induces crazy vertigo for the listener.

Mike you are correct, the ITE recordings were primarily designed for headphone use. The versions from Germany were referred to a binaural recordings. If you listen over speakers, you get a "lively" sound but you lose the intended effect.

The other issue is the role of head movements. When you move your head, the spatial cues should correspondingly change. In this case they do not (in auditory virtual reality displays they actually can - but that is another topic). The consequence is that when you move your head (and the image does not remain in the the same external position) you lose the sense of externalization of the sound and the image "collapses" to an internalized percept (like most headphone listening). There are some tricks to get around this, such as introducing some reverberation is the most typical, but the problem is still there. In order to induce vertigo, you typically need to go to a fuller virtual reality display where you have a visual channel also. In this case the vertigo is a result of misligned auditory and visual cues (and others also - it is a complicated issue that is not quickly summarized).

BTW Mike, in some of the other threads, you are getting easily suckered into some stupid and hostile arguments. You are being baited. There is no reason to let them get the better of you...

Good Luck,

-Tom

[mas]

Actually, no.

The ITE method involved suspending a silicone tranducer down the ear canal into the pressure zone of the eardrum. Each mic appears as a bobbin earring [] (very stylish!) The technique extended the PZM concept to the pressure zone of the ear, thus accurately capturing all of the phase information exactly as your ear drum receives it just prior to its being transformed from air based energy.

What makes this system so unique is that it captures all of the acoustic phase exactly as it exists after being processed by the ear - the entire pinnae structure.It is MUCH more advanced then the simple approach taken by the various less sophisticated ( but nevertheless still advanced techniques) such as the Crown SASS system.

The playback is via a front LF pair and a rear LF pair, and the gain is adjusted to the rear pair so that it is just apparent that there is an ambient source.

The recording is localized from the perspective of the person who 'wore' the microphones. Thus, if an orchestra condustor wears them you will experience the concert from the location and perspective of the conductor - with violins to your left, tympani to the far right rear, and the annoying audience member who talks during the perfomance over your left shoulder about 10 rows deep. This was the first recording I listened too - with my eyes wide and mouth open I might add!

The second recording was both a treat and a trick that verified beyond ANY doubt the voracity of the technique. It was made at a Grateful Dead show in 1990 where Don and Carolyn Davis had been invited to help with an issue with the sound system. During the course of the recording, someone not familiar with the fact that the person wearing the ITE microphones was recording, walks up from behind the person , taps them on the shoulder to ask them a question.

Each one of us listened to the recording one at a time sitting in the 'sweet spot' of the 4 speaker F-R pair. We were only allowed to witness others experience this after we each had our turn to listen first hand. So we had no prior idea of what we were going to experience. In other words, if you had not listened first hand, you had NO idea what was about to transpire. So, at the precise moment during the playback of the show when the 'interruption' occured, each individual who was listening in the sweet spot, IMMEDIATELY turned to look over their left shoulder as if someone had just tapped them on the shoulder and asked a question. You actually were prepared to respond to the accurate acoustic cues enabling such complete and detailed localiztion of the source. It was uncanny to say the least! It was (or it certainly felt) real!

Of course, as the reaction was an automatic response, you quickly realized that no one was there and that it was part of the recording, and you promptly felt very sheepish, while simultaneously being absolutely amazed, as the others who had already been caught themselves in exactly the same way were able to share in your surprise by laughing themselves silly at your expense! But then it was your turn to share in the delight as the next 'victim' took their seat to listen. I have never experienced anything that came even remotely close to this effect of "is it live or is it Memorex'! You are truely THERE!

Likewise, if you were to walk through a Morrocan bazaar wearing the microphones, you would hear all of theacoustic cues in a 360 degree arc both on the horizontal plane as well as the vertical plane. It would capture the TOTAL ambience in all regards.

Let me make a couple of calls and see exactly what the status is of the technology. I know that the IMAX folks were interested in it. But I am not aware of any additional commercial use. Additionally, for 5-6 years everyone who visited Don and Carolyn became an appointed volunteer in the making of pinnae measurements in order to ascertain the best ear response topology - and just the variation in this is utterly amazing as well! I know they were interested in using this information to create an artificial 'head' for the purposes of using the technique - but again, allow me to make a few calls if you folks are interested.

[PrestonTom]

MAS,

I got confused. You are correct about the ITE system and I was stuck on the binaural recording version.

However, my confusion remains. If the ITE is recorded using an "in the ear canal" recording, then you wil get all the spatial cues arising from the the sound interacting with the "first" listener's head and pinnae or the head related transfer function (HRTF). If the sound is now reproduced using external speakers, then the sound will now be further convolved with the "second" listener's HRTF (from the listener in a given postion etc). Is this what you mean? If so, now you have got two sets of HRTFs that are convolving the signal and only one of the HRTFs is "correct". What am I missing?

-Tom

[mas]

Tom, Let me make a few calls and I will attempt to get the exact details of all aspects of the apparatus and microphones, as well as a detailed description of the functional principals - I have been meaning to do this for a while now and this is a good kick to get me movin...

[dhtman]

I ran a jazz label in Europe in the '90s and sat in the studio control room when a good few albums were being mixed. A couple of points about studio mult-track recordings:

First, a mixing studio has its walls covered with acoustic damping materials, so that the engineer and artists aren't subjected to relfections from the monitor speakers.

Second, whilst terms like forward, back, up and down are used in the mixing process - they are in relation to amplitude of any recorded track in relation to the others. I never heard any discussion about 3D imaging or soundstage height - these are audiophile artifacts induced by high frequency reflections in a listening room allied with the way the ear/brain interprets the reverb on individual instruments. In live recordings it's possible for this to be perceived as depth.

The whole "holographic 3D imaging" concept is a fabrication invented by certain audio reviewers/magazines, in support of the "higher price/better sound" scam that serves the magazine emplyees and their advertisers.

For the OP, the best way to test your speakers/room would be either with a recording using simple stereo crossed pair mikes (find some local amateur musicians who are happy to help out) or alternatively a test cd like the one from Linkwitz Labs.

[Duke Spinner]

I I never heard any discussion about 3D imaging or soundstage height - these are audiophile artifacts induced by high frequency reflections in a listening room allied with the way the ear/brain interprets the reverb on individual instruments

The whole "holographic imaging" concept is a fabrication invented ( like PRaT ) by certain audio reviewers/magazines, in support of the "higher price/better sound" scam that serves the magazine emplyees and their advertisers.

to test your speakers would be either with a live recording using simple stereo crossed pair mikes or alternatively a test cd. like the one from Linkwitz Labs.

now ...That ....

is a well reasoned Answer.................[]

and, I concurr wholeheartedly

[dhtman]

I think PRaT (pace, rhythm and timing) is a label that classifies a particular type of distortion - characterized by good leading edge transient response followed by overdone decay of the notes (in electronics) and massive deletion of low frequencies and complexity (in turntables)!

Point is, in a well constructed audio system, PRaT should be in there, along with tonal balance, low distortion and noise, good detail resolution, etc. It gets a bad name in systems where the other stuff is missing - e.g. Naim amps don't do detail very well (IME).

[ZAKO]

Good answere.....Especially when the mics are pointed down on the performers bald heads.