ALMOST HEARD

Behind Almost Heard

unfriendly

This page steps outside the fictional world of Almost Heard to tell the story of how the project came to exist.

A Writer Before Almost Heard

Long before Almost Heard, Geoff Buffett wrote poetry.

Much of the earliest material now finding its way into Almost Heard began during his university years in the 1990s. The words existed, but for the most part they remained words on a page.

Then writing largely disappeared from his life for close to thirty years. Career, work and adult life occupied the space that writing once had.

The writing was not gone forever. It was waiting.

The Second Act

In his 50s, Geoff discovered AI music.

He was not a singer. He did not have the instrumental or production skills needed to build the music he could imagine around his words, nor the financial resources required to hire musicians, singers and a studio to do it for him.

AI changed what was possible.

For the first time, Geoff could take words he had written and hear them brought to life as songs. Poems that had existed quietly for decades could suddenly have voices, arrangements and performances.

The most important result was not simply that old writing became music.

The excitement of hearing his words brought to life awakened something that had been dormant for almost thirty years.

Geoff started writing again.

Old poems began becoming songs, but new songs began appearing too. What started as an opportunity to revisit earlier writing became a new period of creativity: a second act.

Becoming Almost Heard

The experiments grew.

Songs began to develop recurring voices. Those voices became artists. The artists developed appearances, personalities, histories and musical identities. Songs began finding the artist who suited them rather than simply existing as isolated experiments.

A collection of songs gradually became a fictional music world.

Almost Heard emerged from that process.

The name carries its own meaning. For decades, some of these words existed but had never truly been heard. Technology finally gave them a way to become something more than words on a page and gave their writer a way to seek the audience he had never been able to reach through music before.

At its core, Almost Heard is not a story about technology.

It is the story of a writer with something to share, trying to find an audience.

The Role of AI

AI is fundamental to Almost Heard, and the project does not attempt to hide or minimize that contribution.

Without AI, Geoff would never have been able to hear his words realized this way. He does not have the singing voice or musical skills to perform the songs himself, and traditional production involving musicians, singers and studio time would have placed much of this work beyond his practical reach.

AI removed those barriers.

But the substance at the centre of the work remains the writing: the words, stories, ideas, humour, questions and emotions Geoff wants to communicate.

AI provides an opportunity to present that writing in a form that can be heard by people who might never encounter it as poetry or lyrics on a page.

In that sense, the motivation behind Almost Heard is not especially new. It is the same motivation writers have always had: to make something, to share it, and to hope that somewhere it finds an audience.

AI didn’t just give old words a voice. It gave their writer his voice back.

The Creative Process

Almost Heard is not a one-button process in which a request is entered and a finished song simply appears.

A song can begin with a newly written lyric or with an older poem being revisited decades after it was written. The words may need to be adapted into song structure. Decisions are made about musical direction, tone, arrangement and which Almost Heard artist should perform it.

AI music tools make it possible to explore performances and arrangements that Geoff could not personally perform. Multiple interpretations can be generated, compared, rejected, revised and developed before a direction is chosen.

The selected material can then move through audio editing and production, artwork and artist development, video production, sequencing and release decisions.

Generative AI can make substantial creative contributions during that process. Almost Heard does not pretend otherwise. It is one part of a larger collaboration between writing, direction, selection, experimentation, editing and production.

The purpose of documenting the process is not to diminish what AI contributes or to inflate the human contribution. It is to be transparent about how the work is actually made.

Tools & Process

The Almost Heard creative workflow currently includes:

  • Suno for AI-assisted musical realization, including performances, voices and arrangements.
  • Audacity for audio editing and cleanup.
  • BandLab for additional audio production, editing and mastering.
  • OpenArt for artist visualization, artwork and visual development.
  • DaVinci Resolve for video editing and production.

The tools may change as the project develops. The reason for using them remains the same: they give a writer without the traditional means to produce music an opportunity to let his words be heard.