So, after tons of test rendering I stop where I am happy with the result. Perfectionism is compromised.
What have I learned?
In the beginning I used one phonetic pose for each letter in the script. This looked very hard and exaggerated after rendering compared to lipsync in modern games. So I thought I must tone down each pose to a smaller expression. Not working. It still looked fast and strange. At last I simply removed poses that was not clearly heard in the sound file:
In the words "Are you" the old was:
A R E Y O U
to
A R Y O U
to finally
A Y U
What have I learned?
In the beginning I used one phonetic pose for each letter in the script. This looked very hard and exaggerated after rendering compared to lipsync in modern games. So I thought I must tone down each pose to a smaller expression. Not working. It still looked fast and strange. At last I simply removed poses that was not clearly heard in the sound file:
In the words "Are you" the old was:
A R E Y O U
to
A R Y O U
to finally
A Y U
This is a very rough render with minimal samples.
The animation has only 15 fps and is not super smooth, but I am still unsure if I really want to go higher. But I have found rendering cheats good enough to increase the game resolution to full screen!
And, yeah, I have a voice actor now! Her name may be revealed later :)
