Alan Blackwell

Joseph Cameron, Alan Blackwell

Understanding and manipulating timbre is central to audio synthesis, yet this remains under-explored in machine learning due to a lack of annotated datasets linking perceptual timbre dimensions to semantic descriptors. We present the Semantic Timbre Dataset, a curated collection of monophonic electric guitar sounds, each labeled with one of 19 semantic timbre descriptors and corresponding magnitudes. These descriptors were derived from a qualitative analysis of physical and virtual guitar effect units and applied systematically to clean guitar tones. The dataset bridges perceptual timbre and machine learning representations, supporting learning for timbre control and semantic audio generation. We validate the dataset by training a variational autoencoder (VAE) on its latent space and evaluating it using human perceptual judgments and descriptor classifiers. Results show that the VAE captures timbral structure and enables smooth interpolation across descriptors. We release the dataset, code, and evaluation protocols to support timbre-aware generative AI research.

Joseph Cameron, Alan Blackwell

We present a comparative evaluation of latent space organization in three Variational Autoencoders (VAEs) for musical timbre generation: an unsupervised VAE, a descriptor-conditioned VAE, and a VAE conditioned on continuous perceptual features from the AudioCommons timbral models. Using a curated dataset of electric guitar sounds labeled with 19 semantic descriptors across four intensity levels, we assess each model's latent structure with a suite of clustering and interpretability metrics. These include silhouette scores, timbre descriptor compactness, pitch-conditional separation, trajectory linearity, and cross-pitch consistency. Our findings show that conditioning on perceptual features yields a more compact, discriminative, and pitch-invariant latent space, outperforming both the unsupervised and discrete descriptor-conditioned models. This work highlights the limitations of one-hot semantic conditioning and provides methodological tools for evaluating timbre latent spaces, contributing to the development of more controllable and interpretable generative audio models.

Charith Perera, Saeed Aghaee, Ramsey Faragher et al.

Location sensing is a key enabling technology for Ubicomp to support contextual interaction. However, the laboratories where calibrated testing of location technologies is done are very different to the domestic situations where `context' is a problematic social construct. This study reports measurements of Bluetooth beacons, informed by laboratory studies, but done in diverse domestic settings. The design of these surveys has been motivated by the natural environment implied in the Bluetooth beacon standards - relating the technical environment of the beacon to the function of spaces within the home. This research method can be considered as a situated, `ethnographic' technical response to the study of physical infrastructure that arises through social processes. The results offer insights for the future design of `seamful' approaches to indoor location sensing, and to the ways that context might be constructed and interpreted in a seamful manner.

Tommy Nilsson, Alan Blackwell, Carl Hogsden et al.

BLE (Bluetooth Low Energy) is a new wireless communication technology that, thanks to reduced power consumption, promises to facilitate communication between computing devices and help us harness their power in environments and contexts previously untouched by information technology. Museums and other facilities housing various cultural content are a particularly interesting area of application. The University of Cambridge Museums consortium has put considerable effort into researching the potential uses of emerging technologies such as BLE to unlock new experiences enriching the way we engage with cultural information. As a part of this research initiative, our ambition has been to examine the challenges and opportunities introduced by the introduction of a BLE-centred system into the museum context. We present an assessment of the potential offered by this technology and of the design approaches that might yield the best results when developing BLE-centred experiences for museum environments. A pivotal part of our project consisted of designing, developing and evaluating a prototype mobile location-based BLE-centred game. A number of technical problems, such as unstable and fluctuating signal strength, were encountered throughout the project lifecycle. Instead of attempting to eliminate such problems, we argued in favour of embracing them and turning them into a cornerstone of the gameplay. Our study suggested that this alternative seamful design approach yields particularly good results when deploying the technology in public environments. The project outcome also demonstrated the potential of BLE-centred solutions to reach out and engage new demographics, especially children, extending their interest in museum visits.

Charith Perera, Saeed Aghaee, Alan Blackwell

This study explores the use of natural language to give instructions that might be interpreted by Internet of Things (IoT) devices in a domestic `smart home' environment. We start from the proposition that reminders can be considered as a type of end-user programming, in which the executed actions might be performed either by an automated agent or by the author of the reminder. We conducted an experiment in which people wrote sticky notes specifying future actions in their home. In different conditions, these notes were addressed to themselves, to others, or to a computer agent.We analyse the linguistic features and strategies that are used to achieve these tasks, including the use of graphical resources as an informal visual language. The findings provide a basis for design guidance related to end-user development for the Internet of Things.