Description
QuestWordHunter — German Voiceover
Quest text you can read is already here. This is the part you can listen to: the German a quest giver hands you, spoken, and any single word out loud when you click it.
Status: spoken. All of it — every quest passage in every expansion, and every word in the German dictionary. 341,538 clips in twelve packs.
Why it has to be pre-generated
The WoW client cannot synthesise speech. There is no API for it and no way to
reach one from an addon. Everything the pack says has to be generated in
advance, encoded as Ogg Vorbis, and shipped as files the client plays with
PlaySoundFile.
That decides the shape of the project. It is not an addon with a tool beside it; it is a generation pipeline whose output happens to be an addon.
How big it is
Measured, not estimated, against the German corpus in
WordHunterWoW-Dictionary-DE:
| clips | audio | |
|---|---|---|
| quest passages — offer, progress, hand-in | 237,264 | 281.3 h |
| dictionary words | 104,274 | 28.8 h |
| total | 341,538 | 310.1 h |
Objectives and quest titles are not in it. Nobody says them out loud — they are read off the screen.
The clip count is sentences, not passages: there are 71,775 passages, and each is cut into one clip per sentence so a reading can be followed, paused and resumed a sentence at a time.
Both hour figures are measured rather than projected, which they did not use to be. The quest hours are the sum of every value in the eleven packs' own duration tables; the word hours are read off the clips themselves. The table used to say 368 h and 397 h, extrapolated from the 107,489 clips that existed at the time at 5.59 s each — the finished corpus averages 4.27 s, and carrying the early mean across the rest is where the extra 87 hours came from. It also said no word had been spoken yet, which stopped being true the day the word pack was committed.
At the pace an RTX 4090 runs the reader, that is days of generation, not weeks, and it only has to happen once. Afterwards each run does the new quests and whatever text Blizzard rewrote.
On its own, or alongside the word hunter
This addon does not need QuestWordHunter. Install it with a sound pack and nothing else, open a quest, and it reads: it watches the quest events itself rather than waiting to be told, and the window that shows who is speaking — with its stop and pause — is its own. So is the options page.
What QuestWordHunter adds is a surface rather than a capability. Its quest panel is where the play button beside each paragraph is drawn, and where the English sentence lights up in step with the reading. Neither can exist without a panel to draw into, so without it they are simply absent, and the reading is unchanged.
Clicking a single word to hear it is the one thing that does want the base addon, because clicking a word is something its panel offers.
Retail 12.1 and Classic Era. GPL v3 — see LICENSE.


