Welcome to Our Publishing House!
Audio and Ai
Just my thoughts on the subject
I know there are a lot of strong feelings around using ai. And I’d like to start by saying that I whole heartedly agree that using ai voices to read audiobooks is just a solid
NO!
But having said that, there are times when I feel ai can be useful.
I remember Adobe announcing Project VoCo back in 2016. It’s hard to find the video now so here it is. If you don’t want to watch it, it was basically a tool for analyzing audio dialogue and then manipulating that audio as if it were text. Quite a cool concept for the time.
As someone who has spent many, many hours proofing/editing audiobooks, I thought:
“This is great! We can fix little mistakes on the fly without having to get the voice talent back in the studio,” which involves scheduling, so more delay. And I figured that since narrators are paid by the finished hour, why not save them that extra 1 to 3 hours (occasionally more) doing a pickup session.
But Adobe later announced that the tool was never intended to be released. It was just a research prototype. The internet started to suggest it was just some kind of joke.
Years later we now have:
You can click on the links if you’re interested in looking at those tools. They are not affiliate links and I’m not even bothering to track them.
I have played with ElevenLabs. I was trying to re create the voice of someone long gone and all there was to go one was an old cassette tape recording from the 80’s. But the tape had too many sound issues to be usable.
I still think the idea of fixing little mistakes–a stumble on a word or inserting an ‘and,’ or, ‘if’ that was missing–would be of great value to audiobook production. The books still need an actor/narrator to bring the text to life and I can’t see ai ever being able to replace a human narrator. Mimic emotion? Sure, a little. But read, understand and emote all that is in a full work of text. Nah. I could be wrong but still….NAH!
But this brings me to the more recent use…music
I’ve been experimenting with Suno. An ai tool that generates music (and lyrics if you want it too).
I prefer to write my own lyrics. Although, for two songs I’ve created, I used poetry from the early to mid 1800. Whether you love or hate the idea of ai creating music. It is an interesting way to scratchpad song ideas. Or rather to hear what your lyrics could sound like in different styles, without spending hours recording those styles. I think, once you’ve heard something that sounds right, you could then start creating that music yourself.
I’ve created an account on SoundCloud to post some of the things I’ve created. I’m not looking to create heart stopping masterpieces, I’m just having fun and the lyrics reflect that. At least I think they are funny. You may disagree. I’ve also had fun trying to push the envelope, as it were, by making the voice sing with a lisp or like Elmer Fudd in the B Bunny song (songs 1 through 5 are my lyrics – all rights reserved).
I quite like my B Bunny song. If someone ever decides to write a Bugs Bunny musical theatre show, maybe I can sell it :-)
But Can I Create Music?
Why yes, I can. I’ve been a musician for over 40 years. I’ve written music for short and long films as well as live performances. Theatre, dance and such.
I’ve only posted a few things on Spotify (so time consuming).
So why would I be messing around with ai?
For no other reason than it’s fun.
I realize that there will be a lot of people creating and posting a lot of crap. I’m sure Spotify will be inundated with garbage but then Spotify isn’t all that great for independent artists anyway.
Carl Craig, Composer, Lyricist, and the guy who makes Post Hypnotic Press sound good. :)