Tech - AI Avatar Generation - Part 3

Working Zef Avatar with the following functionality so far

Using Sadtalker

Manual Audio WAV supplied
zef.png Graphic used

Attempting to add the following functionality

Google Deepvoice 3 TTS
Subtitling

Trying Google Deepvoice 3

Using the following URL

https://colab.research.google.com/github/r9y9/Colaboratory/blob/master/DeepVoice3_single_speaker_TTS_en_demo.ipynb#scrollTo=kemMMs6pg9Rv

The Google Deepvoice 3 Colab Notebook fails with multiple bugs

Have fixed a couple within my Googe Deepvoice 3 Colab example

Trying OpenAI

Created an OpenAI account and added $5 credit to allow for API calls

Created new Colab Notebook OpenAI and with code that works

Generated voice is very American, and I need a more British sound to it.

Looking into seeing whether can set British voice with OpenAI

Unfortunately the only options are American

Testing OpenAI Transcription Service

Have added OpenAI transcription code section to the colab notebok


Back to Google Cloud


Subtitling

Subtitling seems to be a 2 part process
Part 1 - Create the SRT subtitle file
Part 2 - Use mmpeg to merge the SRT to the video file




Comments

Post a Comment

Popular posts from this blog

Tech - AI Avatar Generation - Part 2

Image
AI Avatar - Part 2 Have created a Public GIT Repo with a GNU Public Type 3 Liecnce so others can use and adapt my code and create ZeF variants at will. https://github.com/rafez/zefavatar Image used - zef.png Audio Used - helloworld.wav And here is the final output If you would like to access the Google Colab I used to create this please visit my Colab Notesbook on colab.research.google.com/drive/1-GTjRuDnFAJfzZPO-JBllWABh9Ax0N8i?authuser=2 https://colab.research.google.com/drive/1-GTjRuDnFAJfzZPO-JBllWABh9Ax0N8i?usp=sharing
Read more

Intro - The Noorullah Foundation has slowly started

OK, How do we go about creating any organisation? Starting off by looking to create the systems necessary to support my foundation. Initially just a brain dumping ground as I explore the various different technologies needed to get my ideas realised. Systems currently in mind that need setting up Drupal CMS OpenProject  Image/Resource Library AI Avatar generator Text-To-Speech Engines Financial systems & controls Group Collaboration Tools
Read more