Thursday, March 9, 2023

What Does The Fox Say?

And the whole thing stitched together via YouTube...

Post literally speaks for itself but a few notes:

Astute viewers or, more accurately, listeners will have noticed the voices are different on each of the videos. This is because there are almost two dozen voice actors to choose from and I am fickle. I did explain in one of the sections how it worked but it was fronted by a cartoon fox and failed to generate a valid result. I also forgot to copy the text before I deleted it and then, when I re-wrote it, I couldn't remember if I'd already used the part about the voices in a previous section, so I left it out. Turns out I hadn't. 

I think the last one is the one I'd keep. They all have names but naturally I can't remember which one it was. Sorry, Jane or Amber or Ashley or whoever. Not, I imagine, that those are the actors' real names.

I was also going to add some AI-created music in the background but that, too, came to nothing. I did find a couple of places that made a decent fist of creating something bearable, using the same prompt I used for the image. Unfortunately, if understandably, they all required that I set up an account before I downloaded the results and I just could not be bothered. My tolerance for filling in forms is limited to once every sixty minutes at most.

My original inspiration for doing any of this came from a video called Using AI Text-to-VIDEO Generators (animation and realistic) on a channel called Sharp Startup. It's only seven minutes long and it's an interesting watch. It has links to three websites that have some kind of free trial or service but the only one I've tried so far is the one I used to make the videos above: D-ID. I most likely will be trying the others, along with some more I found elsewhere.

If there was an unlimited, free version of D-ID or something similar, I would be quite keen on using it to create an animated, talking head version of Inventory Full, which I guess I could post on my YouTube channel. Or, indeed, here. Not all the time, of course, but some posts would really suit the format. At the moment, a free service doesn't seem to be a possibility and I'm not likely to start paying to do it but I'm pencilling it in as a possibility. I believe Google are working on something.

You have been warned!


  1. You may not have mentioned gender, but the AI apparently knows that "fox" is also (somewhat dated) slang for an attractive woman. That's likely where the female part came from.

    Still... Fascinating stuff. I'd almost be afraid of what happens when someone inevitably tries to make an x-rated version of these.

    1. I thought the 1970s meaning of "fox" might have been a factor but I got a lot more actual foxes than foxy ladeez. The problem was almost none of them were useable. I could have specified the prompt more accurately, I guess, but I liked a couple of the results so I just left it at that.

  2. Damn impressive! I didn't know you were working on this, and am really impressed by the result you managed to achieve.

    I prefer the fox to either of the humanoids. It's a bit less uncanny valley, and a bit less distracting. You could make it male if you liked by just switching narrators, of course.

    Big fun! I will follow your YT channel now.

    1. I don't make nearly as much use of the YT channel as I should. Mostly I just upload stuff there if I want to use it for the blog. The advances in AI do make me feel like I couild do more, though. Of course, the irony is that I personally prefer to read text than watch videos so moving content over there would be a bit of a "Do as I say, not as I do" moment.


Wider Two Column Modification courtesy of The Blogger Guide