Blend
EN | FR
Blend
by
Timbre Lingo | Timbre and Orchestration Writings
Published: November 27, 2023
Take a listen to this haunting melody from the first movement of Claude Debussy’s La Mer (1905). It might not be so obvious to you what instrument is playing. Is it a woodwind? Or brass? If so, what is that very thin layer of sound that seems to be floating above it, moving to all the same pitches? You are actually hearing the blended sounds of an English horn and a muted trumpet, where the trumpet is playing an octave above the English horn. Debussy was a master of blending the sounds of different instruments, so much so that in this example it almost becomes difficult to identify either one concretely.
In orchestration theory and practice, blend is both a basic technique and a complex phenomenon that can be used to achieve a variety of orchestral effects [1], [2]. Blend can be understood across two main categories: timbral augmentation, in which a subordinate timbre embellishes a dominant timbre, and timbral emergence, where the combination of the two timbres creates a new timbre, unlike either of its constituents [3]. Blend is thus achieved when two or more timbres appear to “fuse” together. A particular blend of multiple timbres occupies a continuum between total fusion (completely blended) and total heterogeneity (not blended at all). As the degree of fusion increases, individual timbres become more difficult to identify [4]. Various factors affect our perception of blend between instruments; these can be categorized by whether they are observed in isolated (often lab-based) contexts or in musical contexts.
Here’s an example of timbral augmentation. Listen (and follow along with the score if you’d like) to how the bassoon and clarinet augment the dominant sound of the cello in this melody from Verdi’s La Traviata (1853).
Isolated context
Centroid composite: Defined by the sum of both constituent sounds’ spectral centroids, the centroid composite is inversely proportional to the degree of blend between these constituent sounds [1]. In other words, a greater amount of energy in the constituents’ upper harmonics (thus producing a higher centroid composite) is correlated with a decrease in blend. This finding is coherent with the suggestion that sounds that are described as very “nasal” (e.g., oboe) tend to be harder to blend with other instruments, because “nasality” is linked with a high spectral centroid [4]. This same nasality might go some way to explaining why the oboe sounds more prominent than the clarinet in the unison they share here, in Franz Schubert’s ninth Symphony (1824–1826).
Absolute centroid difference: The absolute difference between spectral centroids in a combination of two constituent sounds is inversely correlated with the level of blend [3]. In other words, the more “far apart” the centroids are, the less blend will occur.
Formant structure: Blend also increases with increased similarity of spectral envelopes in constituent sounds, more specifically with the proximity of the main formant frequency values. This means that two sounds that yield divergent formant regions would blend less (i.e. the individual sounds would be more identifiable), which could be due to a decrease in the masking phenomenon between both timbres resulting from the absence of “conflict” in the distribution of the spectral energy. In other words, the salient spectral traits of each instrument would be less concealed. However, we must keep in mind that due to the limits of our auditory system in the lower frequencies, formants that are situated below 500 Hz will be much less salient [1].
Distance in timbre space: A greater distance between the two timbres in a timbre space are correlated with a reduction in blend [4].
Pitch interval: Consonant intervals (e.g., octave, perfect 5th) result in a higher blend level compared to dissonant intervals (e.g., tritone, minor 2nd). Additionally, in general, a larger pitch distance between two sounds correlates with a decrease in blend of those sounds [5].
Onset synchrony: Blend is correlated with the onset synchrony. This correlation is especially important for impulsive instruments (such as percussion instruments) [6].
Attack time: Similarity in the attack slopes (i.e., the degree of “impulsiveness”) of two sounds correlates with increased blend [1]. When deliberately combining an impulsive sound with a sustained sound (e.g., marimba with clarinet), longer attack time, especially for the impulsive sound, is linked to a higher blend. In this case, it has also been shown that blend depends mostly on the characteristics of the impulsive sound, while overall timbre is influenced primarily by the quality of the sustained sound [6].
Amplitude envelope: Instruments with a similar amplitude envelope blend more readily [3].
Register: Higher pitches imply wider separations between partials. This makes the timbral identities more distinct, which has the effect of decreasing blend (by increasing individual identification) in higher registers [1].
Room acoustics: Recent research suggests that the amount of reverberation and longer early decay time in a room is positively linked to the degree of blend between orchestral instruments playing in that room [7].
Musical context (additional factors)
Auditory Scene Analysis principles: For blends of two different voices in orchestral works, principles of the theory of Auditory Scene Analysis become very important. For example, if you are familiar with the rules of counterpoint, you probably know that parallel fifth and octave voicings are generally forbidden in this idiom. A perception-based rationale for this rule is that we tend to fuse sets of voices that follow identical pitch movements (frequency comodulation principle), especially in the case of perfect-consonance intervals (such as fifths and octaves), which have numerous common harmonics (harmonicity principle). Considering these two principles, voices that move in parallel fifths or octaves tend to fuse with one another, giving the impression of a singular voice, and thus reducing the expressivity of the counterpoint.
Dynamics: The intensity with which a musician rubs, plucks, blows, or strikes an instrument influences the energy distribution throughout the partials of the produced sound, with stronger gestures generally producing sounds with higher spectral centroids. Dynamics can therefore alter the spectral properties of timbres by raising their spectral centroids, thus invoking several principles described above (i.e., centroid composite and absolute centroid difference). It has also been shown that pitch and dynamics covary for wind instruments (which refers to the notion of pitch-driven dynamics in orchestration), underscoring the interdependence of the many factors that are involved in our perception of blend [1].
Assigned roles in performance: Musicians that are designated as followers, when playing musical phrases with a leader, tend to adjust their timbre towards a “darker” sound (lower spectral centroid/main formants). They do this consciously, aiming to achieve a better blend while relinquishing the emphasis to the leader. It has been shown that leaders also tend to reduce their main formant frequencies in order to increase blend, but in a much less dramatic way than followers do [1]. Considering these factors and many others, our perceptual system synthesizes all information it receives from the auditory scene and determines the level of blend we experience. Research that unpacks this process thus makes it possible to generate robust theories that can predict the perception of blend in orchestration, which can be helpful for composers and orchestrators [2].
These principles have been used to create an algorithm that generates a prediction on the perceived blend of various instrument groupings in digital orchestral scores [8]. The algorithm works on parameters of onset synchrony, pitch harmonicity, and parallelism in pitch and dynamics. The study’s authors suggested that an updated version of this model, combined with machine learning techniques, could be very promising in automatic identification of blend effects. In a different vein, another study developed a morphing algorithm to help blend the voices of backing vocalists with a lead vocalist when the voices differ in timbral quality [9]. The process is based on spectral envelope matching, and can be used in the mixing process of a music piece.
REFERENCES
[1] Lembke, S-A. (2014). When timbre blends musically : perception and acoustics underlying orchestration and performance [PhD thesis, McGill University]. EScholarship. https://escholarship.mcgill.ca/concern/theses/6h440w776
[2] McAdams, S., Goodchild, M., and Soden, K. (2022). A Taxonomy of Orchestral Grouping Effects Derived from Principles of Auditory Perception. Music Theory Online, 28(3). https://mtosmt.org/issues/mto.22.28.3/mto.22.28.3.mcadams.html
[3] Sandell, G. J. (1995). Roles for Spectral Centroid and Other Factors in Determining "Blended" Instrument Pairings in Orchestration. Music Perception, 13(2). 209-246.
[4] Kendall, R. A., Carterette, E. C. (1993). Identification and blend of timbres as a basis for orchestration. Contemporary Music Review, 9(1-2), 51-67. https://doi.org/10.1080/07494469300640341
[5] Sandell, G. J. (1991). Concurrent timbres in orchestration: a perceptual study of factors determining blend [PhD thesis, Northwestern University].
[6] Tardieu, D., McAdams, S. (2012). Perception of Dyads of Impulsive and Sustained Instrument Sounds. Music Perception, 30(2). 117-128.
[7] Thilakan, J., Gomes, O. C. and Kob, M. (2021). The influence of room acoustic parameters on the impression of orchestral blending [Conference paper]. Euronoise 2021.
[8] Antoine, A., Depalle, P., Macnab-Séguin, P. and McAdams, S. (2021). Modeling human experts’ identification of orchestral blend using symbolic information. International Symposium on Computer Music Multidisciplinary Research. https://doi.org/10.1007/978-3-030-70210-6_24
[9] Roddy, M., Walker, J. (2014). A Method of Morphing Spectral Envelopes of the Singing Voice For Use With Backing Vocals. 17th conference on Digital Audio Effects (DAFx-14).