Social top

Perceptual Benchmarks

English Abstract

[Article en français]

The validity of the ideas presented in the previous sections, as well as below, is ultimately reliant on the faithful exportation of the mastered product to the public. Before continuing it will be useful to offer some precisions on exportability which describe the limitations of such an endeavour more explicitly than has been articulated in other publications (1).

A range of irregularities in frequency response are found in even the most expensive consumer grade systems, meaning that it is unrealistic to expect a direct transposition of the work from mastering studio X to person Y’s listening environment: colouration in varying degrees will be encountered throughout the spectrum. Only a very small percentage of systems — professional monitoring environments with controled acoustics — can boast a response within ±2 dB across the entire usable spectrum. While a number of problematic factors inherent to home systems, or in their use, limit the effectiveness by which the mastered product can be transposed to them (and to varying degrees), it is nonetheless found to be acceptable in the large majority of cases.

The other limit to the universal exportability of mastered works has to do with content. It would seem that with mastered works our perceptual mechanisms are capable of making adjustments to compensate for variances in response typical to less than optimal systems. However, from an æsthetic standpoint, works which are specificly concerned with inducing physical sensations through acoustic anomalies — intermodular distortion, phase “problems”, etc. — will not benefit from mastering to nearly the same degree as works concerned with euphony, because of their dependance on a much higher degree of precision in the transferral, a precision which is not available in most home systems.

Mastering is therefore limited to works concerned with the principle of euphony, to an æsthetic approach that aims to remove all acoustic ambiguities which are “foreign” to the composers intentions. The goal of this article is to lay out a number of objective benchmarks for mastering which are concerned with perception, and not with semantics, analysis or structure. Nearly forty years of commercial mastering have shown that once the aggressiveness of the sound and the major ambiguities have been resolved, the listener is not only more receptive to the work, but also more perceptive to its details.

Skepticism typically expressed at this point concerns whether the engineer can effectively correct problems in the work of another without introducing his own spectral preferences, or whether there is a risk that a “template” will be applied in the mastering process, both of which could compromise the work’s integrity. On the one hand, the bias of one engineer or studio will not be applicable in all cases, and on the other, the engineer has much to lose professionally by engaging in such practices. Further, the general disparity of material common to electroacoustics means that the subject of one person’s bias may not even be present in many works.

The question then arises: How does the mastering engineer detect incongruities, and what is done to correct them efficiently? In short, he concentrates on phenomena related to pre-æsthetic perception which are signaled and interpreted according to common perceptual foundations, and which provoke particular associative reactions. In order to properly perceive these phenomena the engineer must already possess a range of experiences and knowledge concerning the æsthetic possibilities typical to the genre which is to be mastered.

A selection of benchmarks gives an overview of the types of signals of potential problems the mastering engineer encounters: highly-resonant spaces and excessively-coloured reverberation; imprecise or ambiguous sound localisation; confusion between layers; distortion; erratic or unpredictable dynamic curves which are potentially dangerous for the auditory mechanism. The role of the engineer is to assess whether these phenomena are related to the composer’s workspace and equipment or are intentional compositional statements

A second list concerns signals of latent aspects of the work which suggest musical propositions but do not fully or convincingly realize them: size and quality of the stereo image; precision of sound localisation; issues of proximity; certain types of controled resonances. In such cases, the engineer can reinforce the identity of the materials, which in turn helps focus and clarify the composer’s compositional intentions to the listener.

In both cases, the engineer finds and implements appropriate solutions which benefit the æsthetic experience of the work: because of his advantageous position in terms of equipment and resources, he can offer the work a level of quality, efficiency and resources the composer may have wished for but simply did not have access to.

The receptiveness towards the work is at its greatest when the conditions of perception are optimal. This transparency is possible once the audio problems have been corrected and the latent aspects of the work have been brought into a more advantageous relationship with the æsthetic propositions of the work. The mechanisms of identification of audio quality are such that the greater the transparency, the more the production yields to the content. In order to judge the audio quality of a transparent product, it is necessary to discuss the audio content itself.

Despite the absence of a credible descriptive discourse singling out a definition to which everyone agrees, the confusion due to listening systems nor the ensuing apprehensions have prevented the unanimity of professionals concerning what constitutes audio quality and the relation it has to transparency. These aspects of the work are clearly perceptible to this group even without knowledge about the nature of the sources or the conditions of production. A precise mechanism is at clearly work here, and this precision has nothing to do with any sort of structural, cultural or semantic analysis.

One might ask whether the perceptual benchmarks discussed here can function in the context of music employing “unnatural” means and exploring spaces and spectral content which are heavily-transformed or artificial. It would seem that the universe of abstract sound is also a three-dimensional universe, and that its transposition to a fixed support would need to respond to the same criteria of spatialisation and euphony, as it invokes just as compellingly the entire gamut of perceptual benchmarks.


  1. See Dominique Bassal, “The Practice of Mastering in Electroacoustics,” eContact! 6.3 — Questions en électroacoustique / Issues in Electroacoustics (2003), pp. 49–50, section 2.3.4.
— English abstract by jef chippewa

Social bottom