2026-08-24 added shadows, the 3rd speaker, and ability to select (and resize) the scene.
Added more animations. Submitted a pull request to TalkingHead b/c mixamo changed their interface for no reason. Animations now work well, but they also lock me into 30fps. Which is okay.
Renamed cameras to reasonable names rather than numbers.
Added male voices, and submitted a pull request to HeadTTS since it was a bit obsolete.
Added the choice of the voice for each newspartial.
You see? Because of a correct decision I've made previously, I don't inflict pain in me at this step. Previously, I split up the whole newsvideo into multiple newspartials, because of certain limitations of free text-to-speech providers. I took the pain to make sure combining newspartials into a complete newsvideo works. Now, I can seamlessly use this. Voice is generated separately from the video, because voice also generates milisecond-time phonemes and visemes. It is only natural to have one newspartial per speaker, and then add them all up. They can share scenes and cameras. So, the decision to have piece-wise videos and mix them together from the beginning, was a good one.
And added multiple scenes. A good new one looks like this:
~ * ~ * ~ * ~
2026-08-23 updates wrote: I can so far create something like the following. Added multiple avatars, and camera transitions.
~ * ~ * ~ * ~
The UI for the studio is improving step by step, and right now it looks like this. The separate javascript studio:
And what you'd call a mixer that puts all the videos, audio and illustrations together:
Currently I spend a lot of time on matching speech and movements by the millisecond, so the timeline plays an important role here. The timeline runs from the top down, rather than from left to right.
Stay tuned for updates.