This morning I published my latest newsletter (read it here if you haven’t yet!) about AI music and how it may very well be able to influence our emotional state of mind.
After getting home from work, I started diving head-first into the latest hype coming from open source voice cloning and inference: F5-TTS.
On my local PC I was able to clone my voice and that of „GLaDOS“ (from the Portal Game franchise) with only 15 seconds of original audio each.
Asking Perplexity to come up with a fictional podcast between GLaDOS and me was the easiest part. I then pasted the results into the Gradio Web interface of F5-TTS and hit the „Create“ button. After some fine-tuning and repeated „renders“ I received a fictional podcast all from this one program. Of course, creating the video was pure manual work: Davinci Resolve.
Maybe I will choose to create an accompanying podcast for each newsletter now?
Tell me if that’s what you’d like to have each week or rather keep it simple with the written AI-Logbook!