I briefly considered using them as placeholders for characters' voices(on a Scratch Mod), but I worried that might be gen AI

And there would be hell to pay if I used AI regardless of reason, at least from what I've seen.

Is the TTS extension AI? Or at least the kind people don't like?

text to speech is ai, but its a much older than the ai everyone uses now

I'm doing this to allow myself to tackle lip syncing, while I figure out how I can get voice actors. On a nothing budget.

The project in question relates to a blanket banned franchise, so I will not ask for VAs here.

i don't think tts is trained on anyone's voice but i'm not particularly skilled in this topic so i wouldn't know

i do believe its some sort of ai though, maybe not the harmful kind

Text to Speech extensions were made before generative AI was made, text to speech models are usually trained on a single person for each voice that say basic syllables and words, then those words and syllables get put into an extension that tries to interpret the words/syllables being said in the text so that the audio can put together a string of audio clips into one.

For example, “The apple is red” would get read by a TTS extension as the words “the”, “apple”, “is”, “red”, then it would take the audio clips where those words were said and splice them together.

This is a lot less intensive than generative AI, since all the data is pre-trained with less processing in general, and it doesn't have any unethical training pitfalls since it was specifically trained on a set list of people that worked to specifically train the TTS extension.

the TTS extension interfaces with amazon poly, which amazon lists as an “artificial intelligence” product offering, and it ‘generates’ audio. BUT scratch doesn't use amazon's new “Generative voices” engine, just the “Standard voices” engine which works by putting together pieces of recorded speech without any super fancy tech.

might be a kind without a neural network (like what most game NPCs are) but otherwise no.

think like google translate

I don’t think so, at least not generative ai.

No, it is an AI (early AI to be specific, which uses lots of if-statements), but it isnt generative. Although its output is a mix of its data, there is no actual new data formed from it that it didnt have before

Jeffalo wrote:

(#6)
the TTS extension interfaces with amazon poly, which amazon lists as an “artificial intelligence” product offering, and it ‘generates’ audio. BUT scratch doesn't use amazon's new “Generative voices” engine, just the “Standard voices” engine which works by putting together pieces of recorded speech without any super fancy tech.

So, to give a direct answer: according to Amazon, Scratch's text-to-speech extension literally uses “generative” “AI”, but not in the same way as modern AI products like Sora. That is to say, this is not "the kind [of AI] people don't like".

I believe this topic has been answered sufficiently, so I'll ask for it to be closed.