search

EDCD / EDDI

Companion application for Elite Dangerous
Other
440 stars 81 forks source link

EDDI Core: Speech prosody presets #130

Closed Tkael closed 6 years ago

Tkael commented 6 years ago

Proposal

Full emotion with TTS is very difficult (and outside of the range of this project), but with SSML prosody tags we could create "preset" functions to provide some variation and suggest different emotional states for voices that support them:

For example:

{FeelsHappy("I'm happy.")}
{FeelsAngry("Now I'm angry.")}
{FeelsSad("Now I'm sad.")}
Now I'm not.

would translate into:

<prosody pitch="+5%" rate="+10%" volume="+20%">I'm happy.</prosody>
<prosody pitch="+3%" rate="+10%" volume="+100%">Now I'm angry.</prosody>
<prosody pitch="-3%" rate="-20%" volume="-20%">Now I'm sad.</prosody>
Now I'm not.
Tkael commented 6 years ago

Note: depends on #127

Tkael commented 6 years ago

Decided this probably isn't worth doing at this time.

yasmeenh commented 6 years ago

I have tried this, but it doesn't work!!. There are no changes in the voice emotion. Can you help me to get happy, sad, fear, and angry emotions using SSMLprosody tag?