Maybe Bill recorded in stereo and panned each vocal take hard to one side, so all cancellations between vocal takes would be final on mono but spacial on stereo playback. The popularity of pitch-shifted flanged vocoder vocals starting a dozend years ago also was of the kind of "common stereo is broken anyway, let's break it some more".
I think so. Maybe, many of the vocal effects (or vocal recording technics) were invented and used to deal with phantom center issue, while the engineers probably didn't know why it sounds better than pure dry vocal. Mono recordings of Ella Fitzgerald are dry and has no additional effect (except unintentional tube distortion), but they sound great with one Altec speaker. With 2 modern speakers, not very good.
The notches in the phantom center are an artifact of the stereo pair playing identical signals. It's not something you would hear with a sing source right in from of you, be it a loudspeaker, or a human speaker. 
The paper linked to in post #2 claims that fixing those notches increases speech intelligibility, which is great for dialog. To me ear, it makes the center too bright for music, if your system had a good tonal balance before the fix. It did not have a huge effect on image in my system, just tonality.
The paper linked to in post #2 claims that fixing those notches increases speech intelligibility, which is great for dialog. To me ear, it makes the center too bright for music, if your system had a good tonal balance before the fix. It did not have a huge effect on image in my system, just tonality.
That would presume good recording/panning of dialog which is pretty rare. At least when I listen to dialog on radio it often is a total mess with two voices left or right and nothing at the other channel.
When tonality is good in stereo with music I would not bother too much with phantom centers.
When tonality is good in stereo with music I would not bother too much with phantom centers.
It's a matter of taste, I suppose. And it's not a huge problem, just one that audio nuts like us tend to be bothered by.
My desire to fix the problem arose out of noticing that the tonal balance of two speakers together was not the same as with just one speaker playing. It was easiest to hear on sounds panned across from one side to another. I wanted the same tonal balance in all three positions, Left, Center, Right. See post #648
My desire to fix the problem arose out of noticing that the tonal balance of two speakers together was not the same as with just one speaker playing. It was easiest to hear on sounds panned across from one side to another. I wanted the same tonal balance in all three positions, Left, Center, Right. See post #648
Or to get rid of some stink by appliing some perfume.
Electro-acoustical stereophonics is a difficult thing. See, how long it took, until electronic reproduction of stereo vision became feasible, in case it really is. In order to provide spatial cues, which I define as signals occuring two hours to the left and rite hand side and anywhere inbetween, spanning a third of a circle in front of and around the listener, one needs a horizontal, concave array of many loudspeakers.
In a dedicated room of a villa we would employ maybe 31 small loudspeakers and amplifiers each fed by its special grade of xL+-yR. As we want to listen with company, we must allow quite arbitrary listening positions. Runtime difference between two adjacent loudspeakers sin(alpha)*distance must stay below half an inch in order to not cause cancellations within hearing range. Therefore we need many loudspeakers.
I've done the 1 speaker per instrument thing, both live and recorded. It's fun and can sound pretty good. Have also done speaker grouping quite often in live. A cluster for vocals, guitars in a stack, keyboard in another, etc. There are diminishing returns.
Having some surround or ambience channels doesn't hurt, tho.
Actually, no, we don't. But it's easier to achieve a good stereo image with cheap speakers in a bad room by using multiple speakers. With good room acoustics and good speakers, stereo can be a very convincing illusion.
Having some surround or ambience channels doesn't hurt, tho.
Two important points.
Depending on your choice of metaphors. For sure even the swimsuit models on the cover of Sports Illustrated are "enhanced" with PhotoShop and so is the extensive cooking that audio gets on its way to our speakers. Yes, both easy and essential to enhance audio.*
The point many in this thread are missing here is that there is no abstract, math, or purist logic that totally defines outputting a signal from one, two, or three speakers in your room at home. Exclusively a matter of what folks things sounds right and it improves as the art and science progresses**.
Now, the point about stereoscopic vision is a very interesting point to consider. First, don't forget the stereoscopic cue is just one kind of cue from among a half-dozen that are cues to distance. And your mind does over-ride that cue even if that would be totally incomprehensible to an engineer.
The stereoscopic cue works exceedingly well and provides high-fidelity information in a Victorian stereoscope or virtual reality headset when present eye-by-eye.
Hey, isn't that like headphone listening!
Ben
*the second movement of Beethoven's 7th Symphony wouldn't make sense unless the producer cranked up just a few bars of the oboe solo in the middle... not to mention that every female pop singer in the history of recording has become more "breathy" on the recording than in real life
**and by the time everybody has electrostatic speakers at home
Last edited:
This is where I have to step in because I cannot accept this premise or almost any of the others presented here.
First "Stereo" is perfect, period. It is what it is as a medium of distribution of sonic arts. The producer is perfectly aware of its limitations and works with those. What he/she puts out as a "recording" is perfect. It is exactly what they want you to hear no matter what you personally think of it. It doesn't need "correction" or "improvement", it needs to be presented to the listener exactly the way the producer intended. Nothing more, nothing less.
Hard to tell if your comments are meant satirically.
The producer has no solid idea what it sounds like in your house. So different producers cook their sound a bit differently as the same artist(s) go from studio to studio.
Producers do crank up oboes and female pop singers. Some more and some less. Except for some esoteric headphone cults, they have to fiddle or it just wouldn't sound like they think you want it to.
They do this because they feel that you'll like the recording more - for whatever reasons of your own that, in their commercial and artistic judgment seems likely. Producers of classical music cook the sound for their customers in ways that are different from the way pop music producers cook for their customers.
I don't think anyone could dispute the simple description above. But for those of us who are pondering the theoretical questions of music reproduction at home, the point of view that needs to be laid to rest is the question whether there's a physically correct and comprehensive theory of home sound reproduction. (My answer remains, "no" at least regarding anything beyond basic outlook.)
Ben
Last edited:
I am perfectly serious.
The task is not for the producer to guess how it sounds in your listening room, but for you to set up your system so that it sounds like his. Admittedly, the lack of standards for recording studios is a problem, but this is not true for film. So standards do exist, just not universally. If they did then there wouldn't be any question about what is "perfect".
The task is not for the producer to guess how it sounds in your listening room, but for you to set up your system so that it sounds like his. Admittedly, the lack of standards for recording studios is a problem, but this is not true for film. So standards do exist, just not universally. If they did then there wouldn't be any question about what is "perfect".
If the producer has a terrible system, and I mimic it, then that's "perfect"?
For example, the producer's system suffers from numerous nulls at the listening position. Even though his recording contains information at those frequencies, the "perfect" thing to do is copy the nulls?