Open any WAV. One already annotated in OcenAudio arrives with its region comments as tiers; a fresh recording opens empty and you cut its first clip here — sweep a span on the wave view and press Enter. Click a line to hear it. Starting from a video? Open the video in OcenAudio and Save As → WAV; keep the video file beside it unchanged. Attach the video from the transcript toolbar to watch it follow along.
Libraries home: iOne folder holds every library. A directory grant is recursive, so you pick this once and it reaches all of them.
Or open a transcripts folder only — one root tree you work out of:
Or work on pages instead of audio — ᏗᏥᎶᏍᏙᏗ, the Digital Art Workshop:
This folder holds more than one recording — which one?
The filename is the title, so this renames the recording on disk — and the whole family follows it: the sidecar, the Data and Products folders, a same-name video, and (opened from a library) the recording’s own folder. Recents, an unsaved draft, and cited clips on the shelf are re-pointed too.
What was made under the old name keeps it: exported products and past activity-log entries are history, not links.
Word gaps in the syllabary are two spaces, so the marked syllabary box types the pair for you — one press of the space bar lays down the whole gap. The wave view's New clip text bar does the same.
For a single space instead:
- Shift + Space — one space, deliberately;
- space, then Backspace — the pair collapses back to one.
The doubling only happens between two real characters, so a space at the start of a line, or beside a gap already there, stays single however often you press it. Pasted text is never re-spaced — it arrives exactly as it was written.
English, romanization and the glossing lines are not syllabary and keep the ordinary single space, and so does every box when language tools is set to something other than ᏣᎳᎩ.
In ᏗᏥᎶᏍᏙᏗ the same gap is on by default (double space, in the text inspector), but a page's text box can hold English and Cherokee together — so there a gap doubles only when it touches syllabary, and two English words keep their single space.
Writes this whole card into the recording in one pass:
- the About record, as its own block inside the WAV;
- title, speakers and date, into the file's standard tags — what phones and music players show;
- four About fields into the standard tag slots archives read: Archive → Archival Location, Courtesy of → Source, Recorded by → Engineer, Permissions → Copyright. (In the WAV format those slots are named IARL, ISRC, IENG and ICOP.)
Saving into a folder also refreshes the sidecar copy. If the filename is outside the naming convention the About record still saves — only the standard tags are skipped, since there is no title, speaker or date to read out of the name.
This is a shortcut, not the only way: the ordinary Save changes to WAV — the one that writes your lines and regions — re-stamps all of the above with every write. Use this when the details are the only thing you changed. It wakes only when they have actually changed.
how the pages read — PDF & Word
Starts from what’s on screen. Glossing and word-for-word follow the screen too.
Picked companions ride along as pages at the end — choose them in the companions band.
every product
where they land
Beside the recording.
Speaker name database. It lives with the program as a CSV spreadsheet (opens directly in Excel); edits here save automatically in this browser — and, once you give the book a file below, to that CSV on every edit (it lives beside the program's other data). Columns are per-language; bare initials in region labels display using these names.
A font book for your languages. Keep your font files in the program's
fonts/fontbook folder and load that folder here — each language page
shows every font drawing its sample text. “Use in app” makes a font the display
face for the whole program and its products (transcripts, PDFs, Anki decks).
Fonts installed on this computer. Adding one copies your own font into the book in this browser only — it stays your personal copy, never becomes part of the program, and goes into your decks and print files like any book font (when its license allows embedding).
Reference grammars that come with your language package. The Feeling & Pulte 1975 grammar is a private reference copy — it reads inside the app but is never part of any shared release.
The original CSV-filename tools. The Transcript tab replaces them for WAVs tagged in the new embedded-annotation format; these remain for older material.
Convert an Ocenaudio regions CSV to use underscores as delimiters.
Split an audio file into WAV clips based on an Ocenaudio regions CSV (raw or processed — both work). Clips download as a ZIP.
Generate an Anki deck from a folder of audio clips named
<Cherokee>_<English>.wav.
Renames clips from Modified Cherokee syllabary to simple phonetics.
Renamed copies download as a ZIP (inside a phonetic/ folder).
Activity log
About ᏗᎪᏣᎳᎩᏍᎩ
ᏗᎪᏣᎳᎩᏍᎩ (digotsalagisgi) turns annotated Cherokee recordings into clips, Anki decks, and transcripts.
All processing happens in your browser — audio never leaves your computer. Nothing is uploaded and nothing is sent anywhere; the recordings you open stay on your own machine. Once the app has loaded it keeps working with no connection at all.
Keys
Every key below was read off the running code. Most belong to the shared wave view, so they work the same in ᎦᏂᏱ,Ꮩ'Ꮧ and ᎠᏫᏒᏗ.
| what you want to do | press | when it works | comes from |
|---|---|---|---|
| moving around the recording | |||
| Zoom in / out | + − or the wheel, or pinch | always | wave |
| Fit the whole recording | f | always | wave |
| Zoom to the span you swept | z | a span is swept | wave |
| Back to the zoom before | ShiftZ | always | wave |
| Pan sideways | shift-wheel, a trackpad glide, or drag the time ruler | always | wave |
| Jump to the previous / next clip edge | [ ] | always | wave |
| Nudge the playhead | ← → · hold Shift for a second at a time | always | wave |
| Go to the start / the end | Home End | always | wave |
| making and changing clips | |||
| Sweep a span | drag across open wave | always | wave |
| Turn the swept span into a clip | Enter | a span is swept | wave |
| Move a clip's edge | drag the edge · or drag its top strip to slide the whole clip | always | wave |
| Nudge the selected edge, one sample | , . · Shift for coarser | a clip is selected | wave |
| Delete the selected clip | Delete | a clip is selected | wave |
| Drop a mark at the playhead — it opens in the ⚑ panel | m | the wave has focus | wave |
| Cancel — close the form, clear the sweep | Esc | always | wave |
| playing | |||
| Play or pause — the swept span if there is one | Space | always | this program |
| elsewhere in the hub | |||
| Mark this moment, mid-take | m | only while recording | ᎦᏂᏱ,Ꮩ'Ꮧ |
| Save the marks | ⌘S | always | ᎠᏫᏒᏗ |
| Previous / next page | ← → | always | ᏧᏪᏘ ᏗᎪᎵᏰᏗ |
| key | ᎦᏂᏱ,Ꮩ'Ꮧ | ᎠᏫᏒᏗ | ᏗᎪᏣᎳᎩᏍᎩ | ᏧᏪᏘ | ᎶᏍᏛᎢ |
|---|---|---|---|---|---|
| from the shared wave view — the same three by construction | |||||
| + = | zoom in | zoom in | zoom in | — | — |
| − _ | zoom out | zoom out | zoom out | — | — |
| f | fit | fit | fit | — | — |
| z | zoom to span | zoom to span | zoom to span | — | — |
| ShiftZ | zoom back | zoom back | zoom back | — | — |
| [ ] | clip edge | clip edge | clip edge | — | — |
| , . | nudge edge | nudge edge | nudge edge | — | — |
| Home End | start / end | start / end | start / end | — | — |
| Enter | add the clip | add the clip | add the clip | — | — |
| Delete | delete the clip | delete the clip | delete the clip | — | — |
| each program's own | |||||
| Space | play | play | play | — | — |
| m | mark the moment recording only | — | drop a mark wave focused | — | — |
| ⌘S | — | save the marks | — | — | — |
| Esc | cancel | cancel | cancel | close | close the popup |
| ← → | nudge the playhead | nudge the playhead | nudge the playhead | previous / next page | pan the map |
The last row is why this view exists: one key, three different meanings across the hub. Nothing is broken — no two of them are ever on screen together — but it is the kind of thing worth seeing on purpose rather than discovering.
Not listed yet: the shortcuts belonging to this program's own panels and to ᏧᏪᏘ ᏗᎪᎵᏰᏗ's reader. Those still need a sweep before this page can call itself complete.
Guide
The program front to back, in the order you'll meet it. It prints as one document (⌘P) if you want it on paper. Keyboard shortcuts have their own page — Keys, in the ☰ menu.
What this program is
ᏗᎪᏣᎳᎩᏍᎩ (digotsalagisgi) turns annotated recordings into language materials: interactive transcripts, audio clips, Anki flashcard decks, printable and editable documents, subtitles, and archive-ready files. It was built first for Cherokee, but the core works for any language — the Cherokee-specific tools simply appear when a file needs them.
Your recordings never leave your computer. Everything happens in your browser, on your own files: nothing is uploaded anywhere, there is no account, and once the page is open it needs no internet connection.
Getting to a transcript
- Connect library… (the big button on the first screen) — pick your Libraries folder, the one folder that holds every library inside it. You pick it once; the browser will ask you to allow access — allow it, and that one grant covers everything inside.
- The dashboard appears: every recording in every connected library, the recently-opened ones first. A red working draft badge means the transcript is still being worked. Click a card and the recording opens in its own tab, browser-style, at the top of the window.
- No library is required, though: Open recording folder…, Open a WAV…, and Import ELAN file (.eaf)… open loose material, and Recent recordings ▾ brings back anything you've had open before.
Listening
Click any line to hear exactly that clip — it stops at the end. Space plays and pauses. ▶ Play all (Listen ribbon tab) plays the whole recording, highlighting the active line and scrolling to follow; when two people talk at once, every simultaneous line lights up.
Also on the Listen tab: a volume bar (100% is the file's own level; drag up for quiet recordings), a speed bar with the 🐢 button — set a slow speed once, then flip between it and normal with one click, pitch untouched — Consultant view, which opens a clean second window for a consultant or a class (no controls, no working notes, mirrors you live), and Attach video… for recordings that came from one. The thin scrub line under the ribbon seeks anywhere with a click or drag.
The wave view ▾ toggle in the transport bar opens the waveform itself: zoom with the wheel or +/−, sweep a span to play just it, drag clip edges to move boundaries, and read the level meter while it plays. Every key it answers to is on the Keys page.
What the lines are
Each clip is a small stack of lines, and only some of them are typed by hand. For Cherokee: the marked syllabary is the authored tier — the spelling you type, carrying the pronunciation marks. From it the program derives the clean syllabary and the phonetics automatically, live. The English translation (and a romanization, where one exists) is authored too. Other languages skip the derivations and just show what you typed.
The Display ribbon tab decides what shows: the marked tier, the h-mark in phonetics, italics, glossing (morpheme-by-morpheme lines for linguistic work), word for word (each word over its plain-English meaning, in columns), footnotes, and a text-size stepper. Hide-able means the data is still there — showing less never deletes anything.
Editing, and making clips from nothing
Tick edit (Products ribbon tab). Every authored line becomes a text box in place; the derived lines update as you type. In the syllabary box the space bar lays down the two-space word gap for you — Shift+Space when you really want a single space.
A recording with no annotations yet is transcribed right here: open the wave view, sweep a span across a sentence, press Enter, and type into the New-clip bar. Sweep, Enter, type — that is the whole loop.
With edit on, the ✎ beside the title renames the recording — the WAV, its sidecar, its folders, and every reference move together. The gain… button in the wave toolbar is the one edit that changes the audio itself (it warns you about clipping first); ⌘Z takes wave edits back.
Typing the marked syllabary
| you type | it means |
|---|---|
, after a character |
silences its vowel (ᏓᏥᎪᎯ, → datsigo); silences that character's own h too |
' |
the h-mark — an h that survives (Ꮤ'Ꮕ, → tahn) |
+ |
a real pause comma in the sentence (since , is taken) |
] |
a literal apostrophe (since ' is the h-mark) |
-- trailing |
misspeak / false start — kept, and rendered in italics on the derived lines |
- trailing |
the speaker was cut off — never styled |
" ? |
real punctuation is just itself |
Word gaps in syllabary are two spaces; products always render them that way. One space or two while typing are equivalent — and in edit mode the space bar handles it for you.
Saving
Save changes to WAV (or ⌘S / Ctrl+S
from anywhere) writes everything — text, speaker colors, the About record,
flags, footnotes — back into the WAV itself, copying the audio
byte-for-byte, never re-encoded. A .sidecar.json beside the file
keeps a spare copy of the annotations. The WAV is the master document: back
the pair up together and the whole transcript travels with the sound.
Working with OcenAudio on the same file is safe — the program watches the disk and merges rather than overwriting. The one manual step: after saving text here, do File → Revert in OcenAudio before editing audio there.
Flags — working notes
A flag marks what you're unsure of: "check this translation", "couldn't hear this word". Press m with the wave focused (or ⚑ ＋ mark in the flags panel) and a mark drops at the playhead, ready for a note. The ⚑ handle on the transcript's right edge slides the panel out: every flag, click one to jump to it, ✓ Resolve when it has served. Flagged words highlight in their type's color; hovering shows the full note. Flags are temporary and never enter decks, subtitles, or documents unless you tick "Include flags" for a review printout.
Footnotes
Footnotes are the opposite of flags: real content, meant to be kept. Each footnoted clip gets a number, in sequence; the numbers appear as superscripts on screen and in the PDF and Word transcripts, and the notes fold out at the right margin under the footnotes toggle. They are typed in the transcript and saved into the WAV like every other tier.
Products
Deck builder ▴ (Products ribbon tab) opens the drawer. Turn on select clips to choose material where a product wants a selection.
| product | what you get |
|---|---|
| Anki deck | a ready .apkg — audio embedded, your Fontbook font embedded, a speaker credit on every card |
| PDF transcript | print-ready, same script layout as the screen |
| Word transcript | an editable .docx in the same layout |
| Subtitles (SRT) | one cue per clip, named so video players load it automatically |
| ELAN (.eaf) | the archive-deposit file — every tier, paired to the WAV by name |
| Clips to a folder | one tagged WAV per selected clip, named in the filename orthography |
| Send to ᏗᏥᎶᏍᏙᏗ… | clips travel to the page composer — audio, every tier, and a citation — for picture books that play |
A Pseudonyms option (Speaker 1, 2…) swaps real names out of generated
products only — master data always keeps the real names. Dated products carry a
vMMDDYY stamp naming the day you made them, so exports never
collide.
The rest of the ☰ menu
| Speakers | the speaker book — initials → real names, per language. A CSV that lives beside your libraries and opens in Excel; the transcript header credits come from it |
| Fontbook | collect fonts for your languages, compare them on real text, and Use in app — the whole program and its products follow |
| Grammar | published grammars, readable beside your transcript |
| Legacy | the four original CSV-filename tools, kept for older material |
| ᏗᏥᎶᏍᏙᏗ | the page composer — picture books, fliers, documents, with transcript clips wired to play on the page |
| About · Keys | what this program is; every keyboard shortcut, printable |
Where recordings come from
Any WAV opens. The filename is the source of truth for title, speaker, and
date: Title-with-hyphens_initials_MMDDYY.wav — four-digit years
(05031972) for archival material, initials joined with hyphens
when several people speak. Annotate in OcenAudio (one region per
sentence: text in the comment, initials in the label), record fresh sessions
in ᎦᏂᏱ,Ꮩ'Ꮧ, the recorder — or skip both and cut clips right here in
the wave view. Files named outside the convention still open fine; they just
carry less header information.