Voice Acting Bonanza!

Yes, mp3 is the correct file format.
 
I think voiced quotes will be an awful lot of work, although it really could add to the mod mod. I liked Leoreth's mention of the Alpha Centauri, I always loved how the tech advances and secret projects advanced the story. You could maybe have more gruff voices for old and maybe more violent tech advances and then more refined and smoother voices for social advances.
 
Is there, like, some software that machine-learns a voice using samples and then you can feed it lines to read aloud? Or am I thinking for the year 2050 already?
 
...wow. We have reached the Digital Era! *beelines Artificial Intelligence*

I just checked out their samples. They still sound a bit robotic, although less than Google Translate. :crazyeye:
 
If people are prepared to put in the work, voices for techs would be nice. However having played C2C, the voice work they have ended up with I would describe as 'decent.' For what must have been a lot of work it added surprisingly little to the experience.
 
If people are prepared to put in the work, voices for techs would be nice. However having played C2C, the voice work they have ended up with I would describe as 'decent.' For what must have been a lot of work it added surprisingly little to the experience.

I always found that good voice acting added a lot to the feeling of satisfaction of earning a new tech. The bad voice acting was very annoying (and C2C has a lot of that).
 
Last edited:
With the wonders of modern AI, I wouldn't be surprised if within 5 years we'd be able to synthesize the quotes.

If I could choose a voice hybrid for the quotes, it'd be a mix between Arnold Schwarzenegger and Leonard Nimoy.
Now, we can. I found a website that could be used to synthesize voices. Leonard Nimoy is included, but the results sounded too differently from the recorded tech quotes (although it indeed sounds like him), so I made my own.

Here's Nimoy reading all of Ozymandias, for example.

 
They have cr1tikal. Now this is giving me ideas
 
Now, we can. I found a website that could be used to synthesize voices. Leonard Nimoy is included, but the results sounded too differently from the recorded tech quotes (although it indeed sounds like him), so I made my own.

Here's Nimoy reading all of Ozymandias, for example.

By you made your own, do you mean you trained your own AI with his Civ quotes? If so, please share it, that's awesome!
 
By you made your own, do you mean you trained your own AI with his Civ quotes? If so, please share it, that's awesome!

I used Tacotron2 for training and around an hour's worth of records of Leonard Nimoy's voice (mainly his autobiographical audiobook where he speaks the same way he speaks in Civ quotes) as the dataset. There are so many things you have to install first before you could use it in your local machine (or use Google colab to skip past much of the trouble), but apart from training your own model, you could load a pre-trained model file into the inference script and use it for voice synthesis. Here's a link to the model file I made with his voice.
 
Top Bottom