MDOH.

The Multimodal Digital Oral History project is developing a methodology and technical workflow for active engagement with the oral, aural and sonic affordances of both retro-digitised and born-digital oral history collections — across modalities including transcript, sound, waveform, spectrogram and metadata.

We are currently working on the detection of laughter in oral history interviews using AI tools, in combination with non-digital approaches. Laughter offers a compelling example of elements such as hesitations, vocal quirks and pacing typically omitted from manual transcripts.

Our approach encompasses multiple modalities of engagement with oral history materials, moving beyond traditional text-based analysis to incorporate the full spectrum of sonic and multimodal information contained within digital oral history collections.

We work across transcript, sound, waveform and metadata to develop comprehensive methodologies that preserve and analyse the rich, multidimensional nature of oral historical records — honouring both the content and the form of oral testimony.

AI routines enable the study of vocal features at the level of individual interviews and across larger collections. By combining computational approaches with traditional oral history methodologies, we identify and analyse patterns that would be impossible to detect through manual transcription alone.

This work represents a significant advancement in digital humanities methodology, opening new possibilities for understanding the full communicative richness of oral historical materials.