24: The Audio Studio - transcript
Jessica: The QUILL Cast, episode twenty-four, closing the speech arc. I'm Jessica. Today, the outbound direction grew a whole workshop: the Audio Studio, where documents become audiobooks, recordings become chaptered books, and finished books become published shows.
Liam: I'm Liam. One address for all of it: Tools, Speech, Audio Studio. It's a wizard, and the first page asks one question, what's the journey? Narrate documents into speech audio or an audiobook. Combine a folder of recordings into one chaptered book. Or edit an audiobook you already have. Pick one and the wizard reshapes itself around your answer.
Jessica: The narrate journey asks its questions one page at a time, every step announced. What should I read, that's the folder, file types, and filters. Who should read it, engine, voice with a preview button, pace, and if you like, a round-robin of voices. How should chapters work, spoken headings, the page-turn sounder, the pauses. Then output, then the book page, then a summary in plain sentences. Back and Next the whole way, and Skip to summary jumps a repeat run straight to the end. Three keystrokes and an overnight run starts.
Liam: Two features turn that from an exporter into a workshop. First, incremental rebuilds. QUILL fingerprints every document it narrates, the text plus every setting that shapes the audio. Re-run the batch after editing chapter seven, and chapters one through six say, reused, unchanged since last run. Only chapter seven synthesizes. Editing one chapter of a forty-document book now costs minutes, not the night.
Jessica: Second, voice casting. The rotation is fine for variety, but casting is deliberate: a rule says every chapter titled like, interview, gets the guest voice, and chapter one gets the narrator. First matching rule wins, everything else falls back to the rotation. Dialogue-heavy fiction, interview shows, real casting instead of a blind cycle.
Liam: Journey two: you already have the audio. A folder of recordings, one file per chapter, becomes a single chaptered M4B or MP3, and you always review the chapter list before the merge. Point it at a folder of folders in library mode and every subfolder builds as its own book, unattended, titled after its folder. There's even a watch action, drop new recordings in a folder and the book rebuilds itself.
Jessica: Journey three is my favorite, the Chapter Workbench. Open any chaptered MP3 or M4B, or one with no chapters at all, and it's on the operating table. A chapter-aware player anchors the work: play, previous and next chapter, a position slider that speaks human time, and Where am I, which answers with the full glance, chapter four of twenty-four, three minutes in, nine to go.
Liam: Park the playhead where a boundary belongs and press Split at playhead. That's the fix-a-bad-chapter-by-ear move. Set start to playhead retimes a boundary, merge folds a chapter into the previous one, and Restore original undoes the whole adventure. Chapter lists import and export in five formats, Audacity labels, CUE sheets, timestamps, Podcasting two point oh JSON, CSV.
Jessica: And when the chapters are called track zero one through track thirteen? Propose AI titles. QUILL slices the opening minute of each chapter, transcribes it on your machine with the local speech model, the audio never leaves the computer, and asks your configured AI for a short title. The proposals land in the list for review. Nothing is applied blind, and Restore original is right there.
Liam: Then, publish. The Publish button offers explicit paths, each one a separate consent. A podcast feed written next to the book, entirely offline, RSS with iTunes and Podcasting two point oh tags. SFTP upload through QUILL's own SSH machinery, host keys checked, password in the Windows Credential Manager, with live progress you can actually cancel, uploading book dot m4b, forty-two percent.
Jessica: Or Auphonic, the mastering service, if you have an account: check it first and QUILL announces your credits and loads your presets, then uploads, waits, and downloads the mastered results next to the book. And for the self-hosting podcaster, the folder feed: every master in the folder becomes an episode with its own description, and one button regenerates the complete feed after every build, plus an accessible show-notes page, headings per episode.
Liam: Small kindnesses everywhere. Look up book details fills title, author, genre, year, and even downloads the cover from Open Library, all from a half-remembered title. Audition converts just the first document so you judge the voice before committing the night. Job files pin an entire run to a portable file you can edit in Notepad. ACX check tells you, in plain words, whether Audible would accept the loudness.
Jessica: And the old faithful is still there: File, Export, DAISY Talking Book produces the DAISY format that accessible libraries and dedicated players have run on for decades, your headings becoming the navigation levels.
Liam: The full circle of part five, then. Episode twenty-two, your voice becomes text. Twenty-three, anyone's recording becomes documents. Today, your documents and recordings become anyone's listening, chaptered, titled, mastered, and published, from one keyboard-first studio that speaks every step.
Jessica: Next episode we change arcs entirely: setting up AI in QUILL, including the paths that cost nothing. The transcript and every keystroke are in the show notes. Until then, this has been The QUILL Cast.
Liam: Made with QUILL, narrated by QUILL. See you in episode twenty-five.