Since the 3rd crunch on downloads is coming.... I decided to do a bunch of experimental albums. This is one of those songs from a new album.

An AI singing photo generator turns one still image into a video of that face performing a song. You upload a photo, you upload audio, and the model maps phonemes to mouth shapes, adds blinks and head movement, and returns a finished clip. The reason this format keeps spreading is simple. It needs no camera, no filming, no editing skill and no face of your own. One image and one audio file is the entire input. The tools are not equal though, and the gap is wider than the marketing suggests. Most of them are talking head generators with a music upload bolted on, which is a different problem. Speech has pauses. Singing has sustained vowels, held notes and fast lyric runs, and that is exactly where the cheap ones fall apart. I put the same five images through all of them. HOW I TESTED Five source images, because the failure modes only show up when you leave the easy case: A clean studio portrait. A grainy family photo from the 1960s. An illustrated cartoon character. A dog. A classical painting. Same 30 second vocal for all of them. I scored four things: sync on hard consonants, whether the face survives long held notes, whether non human subjects work at all, and export quality. Here is where each landed. 1. ZOICE The only tool that cleared all five images without a failure. The cartoon, the dog and the painting all produced usable performances, which is where most of this category simply gives up and returns a warped face. Sync held through the sustained notes rather than drifting the jaw, and it stayed locked across the full clip instead of degrading after the first few seconds. Output runs to 4K, and the photo to video path is very easy. Best for: anyone working outside clean portraits, and anyone producing at volume who needs the same quality every time rather than one good result in five. Watch for: test at 720p before committing a batch, then push the keeper to full resolution. 2. HEYGEN The most polished product here and very safe on a well lit front facing portrait. It also handles pets and illustrated characters. The 1960s photo is where it slipped, with the grain and soft focus turning the mouth region mushy.

Curious as to how far out everyone sets their project release date in Distrokid, etc. - Do you release yours right away? Do you set yours a week or two out? Looking forward to your replies. Many Thanks

I just got an email a few minutes ago about Artist Roster, I checked my channel that I had been waiting on my OAC approval for and it was now active! I'm still missing 2 releases, but my first one is showing in the Releases section now!

It’s not just about Claude skills and two-cent comments around here lol It took me a little while to post about this because I was waiting for approval on Instagram. And yes, the album has an explicit track (and you can listen to it on Instagram without any issues). I’m not sure if I can or should share album URLs here, but the real reason for this post is: If I did it, you can do it too. Think about what you wanna do, but don't get paralyzed by overthinking, get started. Just start. You need to take action even to make mistakes. And the sooner you break through that barrier to action, the closer you get to your goals.