promotional bannermobile promotional banner

Forever Voiceover

Voices for quests and chatter
Back to Files

v0.1.5-forever

File nameforever-vo-v0.1.5-forever.zip
Uploader
FixpointFixpoint
Uploaded
Sep 25, 2026
Downloads
1.4K
Size
77.0 KB
Flavors
Forever
File ID
8976309
Type
R
Release
Supported game versions
  • 1.60.1

What's new

# Changelog

## 0.1.5

- Skyborne characters no longer leave "skyborne" in the lines they capture.
  The client reports the race as "Windshaper Skyborne" but reads `$r` as
  "skyborne" alone, so the word was never put back and Zamja's greeting was
  voiced "Can Zamja help you, skyborne?" for everyone. The last word of a
  multi-word race now counts. Lines captured by a Skyborne on an earlier
  release are fixed from the original text where the pipeline has it, and
  asked for again where it does not, so playing through them once on this
  release settles them.
- The first-login welcome no longer freezes the client when the Gamepad
  (Alpha) UI is on (#44). It was a standard Blizzard popup, and in gamepad mode
  the client hands every such popup to a path that addons are not allowed to
  trigger, then blames the addon and shows the same kind of popup again. The
  welcome is now a plain window of the addon's own, which that path never sees.
  Thanks to BrandtChristian for tracing it through Blizzard's code.
- No more "secret string value" error at NPCs whose identity the client hides
  from addons (the Disciple of Naralex was the first). The client hands over a
  sealed value in place of such a unit's name and GUID, and the addon now
  treats it as unknown: the speaker comes from the voice pack, or the line is
  skipped, instead of the gossip handler failing.

## 0.1.4

- The talking head shows the speaker you are actually talking to. It used to
  ask the client for the model by creature ID, which on this client comes back
  empty for some NPCs (Varimathras, for one) and left the previous speaker's
  face in the portrait. Now it takes the model from the NPC on screen, and a
  speaker the client cannot draw gets the narrator's book instead of a
  stranger.
- The export window now points to a "Contribute captured lines" issue form,
  one issue per export, instead of the shared inbox thread, which had grown
  past fifty comments. The bot decodes it, reacts with 🚀 and closes the
  issue. Comments on the old inbox still work.
- "Show the talking head" now hides just that: the speaker's portrait goes
  and the parchment, name and text stay, moved over to the left edge. Hiding
  the whole frame is the new "Show the panel" option (`/fvo panel`), for
  audio only.
- Quest lines that address you by gender ("lad" or "lass", "sir" or "madam")
  are no longer voiced with whichever word the first contributor happened to
  hear. The client resolves that choice before the addon sees the text, and
  unlike your name, class and race the other word is simply gone, so the
  pipeline now puts the branch back by comparing a male and a female reading
  of the same line. To make that possible the capture and `/fvo export` record
  your character's sex (one letter, nothing else new leaves the client), and a
  voice pack can ask for a line to be heard again: when the pack says it still
  lacks your sex's reading of a quest, that quest goes into your export even
  though it played. The same hook lets a future pack ask for any line whose
  capture has gone stale. Where the raw text is known (Classic, the beta quest
  cache) a single reading that matches it is fixed at once; thanks to
  jhaubrich for that part and for raising the problem (#28, #29).
- The export and logout messages count lines worth contributing, not only
  lines without audio.
- Captures now say which addon version heard the line and when. Every release
  so far has fixed something an earlier one recorded wrongly, and the pipeline
  can now prefer a line heard by a fixed release over one heard by an older
  one, whatever order they arrive in, instead of whichever was posted last.
  Until a quest line has been heard by this release or later, the pack asks
  everyone for it, so it goes into your export even though it played; the
  flawed recordings get replaced as people simply keep playing.

## 0.1.3

- Stage directions are read by the narrator. Lines like "Hmm... <Jorgen looks
  up at you through squinted eyes.> All right, I'll help ya" used to skip the
  part in angle brackets; now the narrator says it between the NPC's words, in
  whichever narrator voice you picked. Lines that were nothing but a stage
  direction, silent until now, are voiced too. Needs a voice pack built after
  this change; older packs play as before.
- The "no voice pack found" messages name the CurseForge packs to install.
- A quest read from an item, or turned in at a game object, is no longer
  credited to the last NPC you spoke to. The client's "npc" unit outlives its
  dialog, so Admiral Proudmoore's orders were captured and voiced as Gar'Thok
  and the Corpse Laden Boat's turn-in text as High Executor Hadrec, face and
  all. Quest events now trust the quest giver unit the way Blizzard's own frame
  does, an item-started quest is named after its item, and the book shows for
  it instead of the turn-in NPC. Packs built after this record the giver and
  the turn-in speaker separately, so a turn-in at an object shows the object
  even when the client does not say who is speaking.
- The Classic voice pack now comes as two downloads, Base (quests to level 40
  and all gossip) and Base Endgame (quests from 41), because CurseForge caps a
  file at 1 GB. The "no voice pack found" message names both.

## 0.1.2

- A character whose name is an ordinary word ("It") no longer has that word
  eaten out of every line it hears. The client always writes a character's
  name capitalised, so the name is now matched only in its own case; class and
  race still match in either case, because the server writes both.
- Names beginning with a non-ASCII letter ("Ösel", "Élodie") are redacted
  again. The whole-word check in 0.1.1 could not fire next to such a letter,
  so those names were left in captured text and in `/fvo export`. Thanks to
  jhaubrich for the report and the fix.
- Quests whose greeting branches on your gender but whose turn-in does not now
  play the turn-in. The pack recorded one flag for the whole quest, so the
  addon asked for a gendered file that was never made and played silence. New
  packs record which events branch; packs built before this still work. Thanks
  to joergensentroels for the fix.
- The pipeline now reads the addon version an export carries, so it can tell
  which repairs a submission needs.

## 0.1.1

- Lines no longer call you by the wrong class. The client fills in $n, $c and $r
  before an addon can read the text, so a quest first heard on a rogue was
  recorded saying "rogue" and then said that to everyone. Captures now store the
  placeholders, and the narrator says "adventurer" instead.
- `/fvo export` no longer carries your character's name, class or race: the
  placeholders go back in as the line is captured, so a submission says what the
  NPC said and nothing about who heard it.
- Gossip matches whoever is reading it. A greeting recorded on one class used to
  hash differently for every other class, so it often fell back to fuzzy matching
  or went silent.
- The narrator's voice is now yours to pick. Quests and gossip from objects and
  items have no speaker, so a narrator reads them; voice packs can carry those
  lines in several voices, and Options > Audio > Narrator voice chooses one
  (`/fvo narrator` cycles). Lines the chosen voice has no recording for keep
  the default narrator.
- The talking head now uses your faction's parchment by default; clear
  Options > Talking head > Faction parchment style for the dark panel.
- Fixed the queue panel's "nothing else is waiting to play" line hanging off
  the left edge of the panel.
- The play buttons on the quest list rows are gone: they covered the status
  icon the quest log draws there ("..." for in progress, "?" for ready to turn
  in). Open a quest to read it and the Play button beside Back does the same
  job, next to the text it reads.

## 0.1.0

First release: the player addon, without a voice pack. This version collects
the lines players see so the pack can be generated from them.

- Talking head styled after the client's own, with the speaker's model, name,
  quest title and the text paged in time with the audio.
- Queue with pause, skip, clear and reorder; a queue panel; play buttons in the
  quest log list and next to Back in the quest details.
- Options under Escape > Options > AddOns, an addon compartment entry with a
  playback menu, `/fvo` commands.
- Capture of every quest and gossip line seen, `/fvo export` to contribute
  them, a welcome note and chat reminders while no pack is installed.
- Voice packs register through `ForeverVO.RegisterPack`; quests are keyed by
  ID and gossip by speaker and text hash, with a fuzzy fallback.

This mod has no additional files