Question regarding WordBoundary events for subtitle generation

[mew1325]

Hello,

First, thank you so much for creating and maintaining the edge-tts library!

My goal is to create audio with word-by-word synchronized subtitles. I have successfully installed the latest version of the library directly from the GitHub master branch.

However, I've found that while I can successfully stream audio and receive SentenceBoundary events, the stream does not seem to contain any WordBoundary events. This makes it impossible to generate the word-level subtitles I'm hoping for.

I'm writing to ask if this is the current expected behavior, or perhaps if I'm missing a step to properly enable the word-level data. Is this a known issue, or is there a specific method to request WordBoundary events in the current version?

To provide clear context, I will attach the simple diagnostic script I used for testing, as well as its complete output log, which confirms the absence of WordBoundary chunks.

test_wordboundary.py

Here's the log: --- Starting WordBoundary Diagnostic Test ---
Using Voice: en-US-JennyNeural
Synthesizing Text: "Hello world. This is a simple test to see the word boundary events."

[Log] Connecting to server and receiving stream...

[INFO] Received SentenceBoundary chunk: {'type': 'SentenceBoundary', 'offset': 500000, 'duration': 16375000, 'text': 'Hello world.'}
[INFO] Received SentenceBoundary chunk: {'type': 'SentenceBoundary', 'offset': 16875000, 'duration': 38250000, 'text': 'This is a simple test to see the word boundary events.'}

[Log] Stream finished.
Diagnostic audio file saved to: diagnostic_audio.mp3

--- Diagnostic Result ---
❌ FAILURE: No 'WordBoundary' events were received.
This would indicate a persistent issue with the library or the service connection.