Audio Engineering

The relationship between crowd noise and crowd behavioral dynamics is a relatively unexplored field of research. Signal processing and machine learning (ML) may be useful in classifying and predicting crowd emotional state. This paper describes using both supervised and unsupervised ML methods to automatically differentiate between different types of crowd noise. Features used include A-weighted spectral levels, low-level audio signal parameters, and Mel-frequency cepstral coefficients. K-means clustering is used for the unsupervised approach with spectral levels, and six distinct clusters are found; four of these clusters correspond to different amounts of crowd involvement, while two correspond to different amounts of band or public announcement system noise. Random forests are used for the supervised approach, wherein validation and testing accuracies are found to be similar. These investigations are useful for differentiating between types of crowd noise, which is necessary for future work in automatically determining and classifying crowd emotional state.

Let S be a d-dimensional separoid of (k -1)(d + 1) + 1 convex sets in some "large-dimensional" Euclidean space IE N . We prove a theorem that can be interpreted as follows: if the separoid S can be mapped with a monomorphism to a d-dimensional separoid of points P in general position, then there exists a k-colouring ς: S → K k such that, for each pair of colours i, j ∈ K k , the convex hulls of their preimages do intersect-they are not separated. Here, by a monomorphism we mean an injective function such that the preimage of separated sets are separated. In a sense, this result is "dual" to the Hadwiger-type theorems proved by and . We also introduce ϑ(k, d), the minimum number n such that all d-dimensional separoids of order at least n can be k-coloured as before. By means of examples and explicit colourings, we show that for all k > 2 and d > 0, Furthermore, by means of a probabilistic argument, we show that for each d there exists a constant C = C(d) such that for all k, ϑ(k, d) ≤ Ck log k.

The interdisciplinary nature of foreign language dubbing involves the collaboration of actors, directors, producers, translators, sound and video departments, and often develops into an elaborate production project in its own right. Increasingly, particularly with motion pictures, the dubbing industry faces new challenges as studios are required to work at the pace of the original production and follow any and all modifications, leading ultimately to the global release of a film everywhere simultaneously, an economic model that has been gathering momentum in recent years. The technology for foreign dialogue dubbing has undergone a great optimization of the production work-flow to accommodate such new requirements, and this presentation will show various practices currently in use throughout Europe and North America. I will discuss the idealized work-flow versus technology-specific solutions, cost optimization, media security, and the advantages of certain types of technologies on foreign-language synchronization. Areas of development in foreign-language dubbing and internationalization will be discussed.

Two-channel stereo systems, despite the rise of multi-channel audio, continue to dominate the music industry. Beyond simple left-right sound, high-performance stereo can create a 3D illusion of a live event. This work focuses on the controversial element of 'elevation' in soundstages. Using a specialized high-fidelity setup, we provide psychoacoustic evidence that listeners can perceive verticality in audio images and discuss the biological mechanisms that make this 3D perception possible.

In this paper, we present a novel set of tempo-related audio features for applications in audio retrieval. As opposed to existing feature sets commonly used in the retrieval domain which mainly focus on local spectral characteristics of the audio signal, our features capture its local temporal behaviour w.r.t. tempo, rhythm, and meter. As a key component to obtaining a high level of feature robustness we introduce the cyclic beat spectrum (CBS) consisting of residual tempo classes which are constructed similarly to the wellknown pitch chroma classes. We illustrate the use of the newly constructed features by applying them to robust timescale invariant audio identification.

Audio thumbnailing, which aims at finding the most representative audio segment of a music recording, is an important task in music information retrieval. In general, the notion of a thumbnail is not well-defined and several musical parts may be good thumbnail candidates. For example, for popular music, both a verse and a refrain section may serve as suitable thumbnail candidates. Instead of considering only one thumbnail, we consider in this paper the problem of finding the two most representative segments that correspond to different musical parts. We denote these two segments as double thumbnails. As our main technical contributions, we propose two approaches for computing double thumbnails, both extending a previously introduced repetition-based thumbnailing procedure. In the first approach, which is straightforward, we simply apply the original thumbnailing procedure two times in an iterative fashion. In the second approach, we introduce a novel method for jointly estimating the two thumbnails within one optimization procedure. Finally, we report on experimental results demonstrating the performances of the two double thumbnailing procedures and indicate directions towards full music structure analysis.

This paper investigates the hook-driven songwriting and production methodologies associated with Max Martin through a creative practice research approach combining production analysis, reflective evaluation, and practical studio experimentation. Particular focus is placed on the concepts of "hook-first" songwriting and "melodic maths," examining how melody, rhythmic phrasing, repetition, vocal production, and arrangement structure contribute to listener engagement and memorability within contemporary Pop production. Through analysis of productions including Blinding Lights (The Weeknd), Teenage Dream (Katy Perry), Into You (Ariana Grande), and Style (Taylor Swift), the paper explores how Max Martin's production style has evolved whilst maintaining a consistent emphasis on melodic clarity and hook reinforcement. The study also reflects upon the application of these techniques within original compositions, evaluating how melody-first production workflows, layered vocal production, dynamic contrast, and structured arrangement influenced the development of commercially oriented Pop songs. The paper argues that contemporary Pop hooks are constructed through highly deliberate compositional and production strategies that balance repetition, variation, emotional pacing, and sonic clarity.

This paper investigates the production and songwriting methodologies associated with Max Martin through the application of contemporary commercial Pop production techniques within a portfolio of original compositions. Adopting a creative practice research methodology that combines practical studio experimentation, reflective analysis, critical listening, and production evaluation, the study explores how hook-driven songwriting, melodic reinforcement, vocal production, and layered arrangement techniques were integrated into the tracks Autonomy, Breathless, and Slow It Down. Particular focus is placed on the role of "melodic maths," rhythmic precision, vocal clarity, and sidechain processing within contemporary Pop and R&B production workflows. Influences from Phil Spector's "Wall of Sound" production methodology are also examined in relation to density, spatial depth, and layered instrumentation. Through practical experimentation and critical evaluation, the paper considers how contemporary commercial songwriting balances repetition, memorability, and emotional engagement whilst maintaining clarity within dense arrangements. The project concludes that applied creative experimentation provided the most effective method for understanding contemporary Pop production practices, demonstrating how structured compositional and production techniques can support both melodic cohesion and professional production outcomes within songwriting practice. Within this paper, 'melodic maths' refers to the strategic repetition and reinforcement of melodic motifs across vocal and instrumental elements to strengthen memorability and listener retention

Over a hundred years have passed since the advent of sound recording and reproduction, and throughout this time, a key challenge has been the preservation and transmission of the spatial sensations experienced by the listener in a concert hall or theatre. Attempts to solve this problem have led to a transition from monophonic sound recording and reproduction to stereophonic, as well as to more complex multi-channel spatial sound reproduction systems (for example, 7.1 or 22.2 systems). However, their invention is not a definitive solution to the problem of spatial sound, which is why we are seeing alternative solutions, such as binaural stereo, ambisonics and others. Among these solutions is a special system of spatial sound reproduction known as Wave Field Synthesis (WFS). Wave Field Synthesis (WFS) is a technique for spatial sound reproduction that places virtual  sound sources in real space. The generally accepted theory explaining the creation of a phantom sound source is the ‘localisation summation’ theory. According to this theory, stereophony works through the physical summation of signals from two loudspeakers, which create a virtual source. The theoretical foundations of wave field synthesis were laid as early as the late 17th century by Christiaan Huygens in his article on the nature of light. He set out the following principle: any wavefront (a surface connecting points that are in the same phase of oscillation) can be regarded as a continuous distribution of point sources, each of which emits

Many sound engineers test the acoustics of an auditorium by clapping their hands,
in an attempt to assess the room’s acoustic qualities. It is well known that it is impossible
to achieve an ideal response, similar to a mathematical delta function in free space,
because air possesses viscoelastic properties.
In our experiment, we recorded the actual impulse response of a sharp clap. What are the acoustic properties of such a response?
The sound pressure recording of the clap  was made directly next to a person’s hands using Pro Tools, with a sampling rate of 96 kHz.
The impulse represents a rapidly decaying oscillatory process; its duration is less than 5 ms, and the decay time does not exceed 15 ms. The short leading edge lasts less than 2 ms, allowing the sound engineer to hear the first reflections from the hall walls.

Advances in science and technology made it possible, in the late 19th and early
20th centuries, to convert sound waves into an electrical signal. This opened up
opportunities for amplifying, recording and transmitting sound. By the 1930s,
sound recording and reproduction technology had reached a high level of development.
It became possible to reproduce sound across the entire range of audible frequencies and
even beyond, for example from 16 to 30,000 Hz. Non-linear distortions
were minimised to a level imperceptible even to highly experienced
specialists.
But you don’t need to be an expert to distinguish the natural sound
of an orchestra or a human voice from the sound produced by audio equipment.
As research has shown, one of the main drawbacks is the ‘monophonic nature’
of conventional sound reproduction. A listener perceiving the sound in a concert hall is
able to localise musical instruments from different directions. The concert
hall also plays its part by adding reverberation. In this way, we
obtain a sound image rich in spatial information.

Cinema, as a form of the screen arts in general and
in particular in the form of a feature film, is an audiovisual phenomenon in which the visual-plastic sequence is synthetically combined with the auditory (musical and sound) sequence
and thus forms an ‘audio-visual’ image. Of all the varieties
of audiovisual arts, cinema represents the most extensive
platform for the application of means of creating a screen image and the most
comprehensive set of tools regarding the possibilities for depicting reality.
The auditory component in feature films is also more extensive than in other
forms of screen art, not only in terms of the number of its components, but
also in the diversity of the means by which they are applied. However, the complexity of creating
the auditory aspect of the cinematic image lies not so much in the number of its components,
but in its specific nature. The specificity of the auditory component of the cinematic image lies in
the particularities of the viewer’s perception of on-screen sound.
Cinematographic practice shows that the on-screen image is perceived by the viewer
primarily through visual perception of the montage sequence, whilst
the sound sequence is perceived by the viewer subconsciously (with the exception of the conscious
306
perception of individual sounds or musical fragments, which
is achieved by the director through directorial techniques). Thus,
the art of creating an audio sequence is geared towards the predominantly
subconscious perception of on-screen sound by the viewer; in other words, towards
the ‘invisibility’ of sound.
Consequently, in order for the director to create an audiovisual cinematic image,