search

genea-workshop / Speech_driven_gesture_generation_with_autoencoder

This is the official implementation for IVA '19 paper "Analyzing Input and Output Representations for Speech-Driven Gesture Generation".
https://svito-zar.github.io/audio2gestures/
Apache License 2.0
9 stars 8 forks source link

Wrong frame rate in audio encoding in the GENEA'22 codebase #1

Open Svito-zar opened 2 years ago

Svito-zar commented 2 years ago

I have discovered a pretty bad issue in this baseline code: the audio features were encoded at 20 fps, while the motion is 30 fps.

Svito-zar commented 2 years ago

This issue was fixed in the commit https://github.com/genea-workshop/Speech_driven_gesture_generation_with_autoencoder/commit/9b5a895beb3a3083aa9b1055471ac96ba2d305b0