Automating the Production of the Balance Mix in Music Production

Jillings, Nicholas (2023) Automating the Production of the Balance Mix in Music Production. Doctoral thesis, Birmingham City University.

Nicholas Jillings PhD Thesis published_Final version_Submitted Nov 2022_Final Award Oct 2023.pdf - Accepted Version

Download (15MB)


Historically, the junior engineer is an individual who would assist the sound engineer to produce a mix by performing a number of mixing and pre-processing tasks ahead of the main session. With improvements in technology, these tasks can be done more efficiently, so many aspects of this role are now assigned to the lead engineer. Similarly, these technological advances mean amateur producers now have access to similar mixing tools at home, without the need for any studio time or record label investments. As the junior engineer’s role is now embedded into the process it creates a steeper learning curve for these amateur engineers, and adding time onto the mixing process.

In order to build tools to help users overcome the hurdles associated with this increased workload, we first aim to quantify the role of a modern studio engineer. To do this, a production environment was built to collect session data, allowing subjects to construct a balance mix, which is the starting point of the mixing life-cycle. This balance-mix is generally designed to ensure that all the recordings in a mix are audible, as well as to build routing structures and apply pre-processing. Improvements in web technologies allow for this data-collection system to run in a browser, making remote data acquisition feasible in a short space of time. The data collected in this study was then used to develop a set of assistive tools, designed to be non-intrusive and to provide guidance, allowing the engineer to understand the process.

From the data, grouping of the audio tracks proved to be one of the most important, yet overlooked tasks in the production life-cycle. This step is often misunderstood by novice engineers, and can enhance the quality of the final product. The first assistive tool we present in this thesis takes multi-track audio sessions and uses semantic information to group and label them. The system can work with any collection of audio tracks, and can be embedded into a poroduction environment.

It was also apparent from the data that the minimisation of masking is a primary task of the mixing stage. We therefore present a tool which can automatically balance a mix by minimising the masking between separate audio tracks. Using evolutionary computing as a solver, the mix space can be searched effectively without the requirement for complex models to be trained on production data.

The evaluation of these systems show they are capable of producing a session structure similar to that of a real engineer. This provides a balance mix which is routed and pre-processed, before creative mixing can take place. This provides an engineer with several steps completed for them, similar to the work of a junior engineer.

Item Type: Thesis (Doctoral)
25 November 2022Submitted
16 October 2023Accepted
Uncontrolled Keywords: Audio Engineering; Software; Artificial Intelligence; Music; Mixing; Genetic Algorithms
Subjects: CAH10 - engineering and technology > CAH10-01 - engineering > CAH10-01-10 - others in engineering
CAH25 - design, and creative and performing arts > CAH25-02 - performing arts > CAH25-02-02 - music
Divisions: Doctoral Research College > Doctoral Theses Collection
Faculty of Computing, Engineering and the Built Environment > School of Computing and Digital Technology
Depositing User: Jaycie Carter
Date Deposited: 29 Nov 2023 16:27
Last Modified: 29 Nov 2023 16:27

Actions (login required)

View Item View Item


In this section...