Audio’s “Circle of Confusion” is a term coined by Floyd Toole  that describes the confusion that exists within the audio recording and reproduction chain due to the lack of a standardized, calibrated monitoring environment. Today, the circle of confusion remains the single largest obstacle in advancing the quality of audio recording and reproduction.
The circle of confusion is graphically illustrated in Figure 1. Music recordings are made with (1) microphones that are selected, processed, and mixed by (2) listening through professional loudspeakers, which are designed by (3) listening to recordings, which are (1) made with microphones that are selected, processed, and mixed by (2) listening through professional monitors...... you get the idea. Both the creation of the art (the recording) and its reproduction (the loudspeakers and room) are trapped in an interdependent circular relationship where the quality of one is dependent on the quality of the other. Since the playback chain and room through which recordings are monitored are not standardized, the quality of recordings remains highly variable.
Creating Music Recordings Through An Uncalibrated Instrument
A random sampling of ones own music library will quickly confirm the variation in sound quality that exists among different music recordings. Apart from audible differences in dynamic range, spatial imagery, and noise and distortion, the spectral balance of recordings can vary dramatically in terms of their brightness and particularly, the quality and quantity of bass. The magnitude of these differences suggests that something other than variations in artistic judgment and good taste is at the root cause of this problem.
The most likely culprits are the loudspeakers and rooms through which the recording were made. While there are many excellent professional near-field monitors in the marketplace today, there are no industry guidelines or standards to ensure that they are used. The lack of meaningful, perceptually relevant loudspeaker specifications makes the excellent loudspeakers difficult to identify and separate from the truly mediocre ones. To make matters worse, some misguided recording engineers monitor and tweak their recordings through low-fidelity loudspeakers thinking that this represents what the average consumer will hear. Since loudspeakers can be mediocre in an infinite number of ways, this practice only guarantees that quality of the recording will be compromised when heard through good loudspeakers . This is very counterproductive if we want to improve the quality and consistency of audio recording and reproduction.
Another significant source of variation in the recording process stems from acoustical interactions between the loudspeaker and the listening room - Below 300-500 Hz, the placement of the loudspeaker-listener can cause >18 dB variations in the in-room response due to room resonances and placing the loudspeaker in proximity to a room boundary.
Evidence of acoustical interactions has been well documented survey of 164 professional recording studios where the same high-quality, factory calibrated monitored was installed . Figure 2 shows the distribution of in-room responses measured at the primary listening location where the recordings are monitored and mixed. The 1/3-octave smoothed curves show a reasonably tight ± 2.5 dB variation above 1 kHz. However, below 1 kHz, variation in the in-room response gets progressively much worse at lower frequencies. Below 100 Hz, the in-room bass response can vary as much 25 dB among the different control rooms! You needn’t look any further than here to understand why the quality and quantity of bass is so variable among the recordings in your music library.
Evaluating Loudspeakers When the Recording is a Nuisance Variable
Loudspeaker manufacturers are also trapped in the circle of confusion since music recordings are used by listening panels, audio reviewers, and consumers to ultimately judge the sound quality of the loudspeaker. The problem is that distortions in the recording cannot be easily separated from those produced by the loudspeaker. For example, a recording that is too bright can make a dull loudspeaker sound good, and an accurate loudspeaker sound too bright . A review of the scientific literature on loudspeaker listening tests indicates that recordings are a serious nuisance variable that need to be carefully selected and controlled in the experimental design and analysis of test results.
At Harman International, we try to minimize loudspeaker-program interactions in our loudspeaker listening tests by using well-recorded programs that are equally sensitive to distortions found in loudspeakers. Listeners become intimately familiar with the sonic idiosyncrasies of the different programs through extensive listener training and participation in formal tests. In each trial of a loudspeaker test, the listener can switch between different loudspeakers using the same program, which allows them to better separate the distortions in the program (which are constant), from the distortions in the loudspeaker.
Through 25+ years of well-controlled loudspeaker listening tests, scientists have identified the important loudspeaker parameters related to good sound, which can be quantified in a set of acoustical measurements , By applying some statistics to these measurements, listeners’ loudspeaker preferences can be predicted . The bass performance of the loudspeaker alone accounts for 30% the listener’s overall preference rating. Good bass is essential to our enjoyment of music, which unfortunately is a frequency range where loudspeakers and rooms are most variable (see Figure 2). Controlling the behavior of loudspeakers and rooms at low frequencies is essential to achieving a more consistent quality of audio recording and reproduction. Fortunately, there are technology solutions today that provide effective control of acoustical interactions between the loudspeaker and rooms.
Breaking the Circle of Circle of Confusion
As Toole points out in , the key in breaking the circle of confusion lies in the hands of the professional audio industry where the art is created. A meaningful standard that defined the quality and calibration of the loudspeaker and room would improve the quality and consistency of recordings. The same standard could then be applied to the playback of the recording in the consumer’s home or automobile. Finally, consumers would be able to hear the music as the artist intended.
 Floyd E. Toole, Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms, Focal press (July 2008).
 Floyd Toole, “Loudspeakers and Rooms: A Scientific Review,” J. Audio Eng. Soc., Vol. 54, No. 6, (2006 June). A free copy of this paper can be downloaded here
 Sean E. Olive and William Martens “Interaction Between Loudspeakers and Room Acoustics Influences Loudspeaker Preferences in Multichannel Audio Reproduction,” presented at the 123rd Convention of the AES, preprint 7196 (October 2007).
 Aki V. Mäkivirta and Christophe Anet, “The Quality of Professional Surround Audio Reproduction, A Survey Study,”19th International AES Conference: Surround Sound - Techniques, Technology, and Perception (June 2001).
 Todd Welti and Allan Devantier, “Low-frequency Optimization Using Multiple Subwoofers,” Audio Eng. Soc., Vol. 54, No. 5, (May 2006). A free copy of this paper can be downloaded here
 Sean E. Olive, John Jackson, Allan Devantier, David Hunt, and Sean Hess, “The Subjective and Objective Evaluation of Room Correction Products,” presented at the 127th AES Convention, New York, preprint 7960 (October 2009).
 Sean E. Olive,”The Preservation of Timbre: Microphones, Loudspeakers, Sound Sources and Acoustical Spaces,”8th International AES Conference: The Sound of Audio (May 1990)
 Floyd E. Toole, “Loudspeaker Measurements and Their Relationship to Listener Preferences: Part 1,” J. Audio Eng. Soc., Vol. 34,No.4, pp.227-235, (April 1986). A free copy of this paper can be downloaded here
 Floyd E. Toole, “Loudspeaker Measurements and Their Relationship to Listener Preferences: Part 2,” J. Audio Eng. Soc., Vol. 34, No.5, pp. 323-348, (May 1986). A free copy of this paper can be downloaded here
 Sean E. Olive, “A Multiple Regression Model for Predicting Loudspeaker Preference Using Objective Measurements: Part II - Development of the Model,” presented at the 117th Convention of the AES, preprint 6190 (October 2004).