[METHOD] Recording audiobook with AI (Black Hat)

rijiwi9723

Registered Member
Joined
Jan 3, 2023
Messages
84
Reaction score
67
Hello, everyone!

I'm excited to share a fantastic method that allows you to earn a few bucks (or even more!) by utilizing AI to create predub audiobooks.

1. Download any royalty free audiobook from torrent website. Many audiobook websites and forums offer a "classic" category that features a selection of older books. A significant portion of these titles meet our criteria. It's essential to ensure that any downloaded audiobook is divided into individual chapters and available in popular formats like mp3 or wav, as this organization will prove convenient for future use.

Screenshot 2023-08-01 at 14.11.28.png

2. Download and install any AI voice converter. Most of the website just using free open source. It will convert original audio to desire voice.For the best results, utilize the F0 harvest method. It has proven to deliver excellent outcomes in my experience. You need to download AI voice model. Use voice which is not mainstream and everybody knows him but also is good for audiobook narrating. If origital voice is male one use male model as well so you dont need to do pitch correction.

3. Split your large audio files into smaller chunks using software like Audacity. This will make it easier to manage and process individual segments. Convert each smaller audio chunk into the desired voice. Utilize the F0 harvest method for optimal results. Once you've converted all the files, merge them back together into a cohesive audiobook using Audacity or any other audio editing software. To ensure compatibility and standardize the audio, convert the merged audiobook to a 44 kHz sample rate.

Screenshot 2023-08-01 at 14.15.04.png

4. Create cover with Canva or Photoshop , use made up narrator name (use same one if you upload multiple books with same voice model).

5. Create your account on any audiobook distributors. Don't forget cut intro and outro and replace them with your own converted voice.

I uploaded 3 audiobooks and making ~$3 per audiobook month.

Screenshot 2023-08-01 at 14.21.29.png
 
Not a good process for only 3$ a month? I don't get that sentence with your capture.
 
5. Create your account on any audiobook distributors. Don't forget cut intro and outro and replace them with your own converted voice.

I uploaded 3 audiobooks and making ~$3 per audiobook month.

There are many. What are your top 3, and what one specifically are you using to get those earnings in? Cheers. Well done.
 
The idea is not bad, I think there are more than 3$ possible and the whole thing is very scalable
 
Thank you for the share! This really sounds promising! Scalable & possible to add your own twists.
 
There are many. What are your top 3, and what one specifically are you using to get those earnings in? Cheers. Well done.
I uploaded only 3. Good thing is that they are selling without any marketing. I think that lots of audiobook subscription website have small audiobook catalog so if you have good cover you will pretty easily rank with anything.

what's the F0 method?
There are different method how you can convert original sound to target one. You basically choosing between high quality and slow conversion or low quality and fast conversion. Also AI model must be train same way.

Screenshot 2023-08-03 at 08.39.11.png

Only downside I find so far is that converting one 10 hour book took me 5 days to make. Tried using colab to use faster GPU cannot make it running.
 
Nice method :D

You are using RVC for this? Seem like you can make converting 10x faster if install latest update and exporting with RMVPE rather than Harvest.

I just tested on my own PC and 10 min audio took 200+ seconds with Harvest, only 20 seconds with RMVPE. But quality is just as good if the used voice model is optimized for talking
 
Nice method :D

You are using RVC for this? Seem like you can make converting 10x faster if install latest update and exporting with RMVPE rather than Harvest.

I just tested on my own PC and 10 min audio took 200+ seconds with Harvest, only 20 seconds with RMVPE. But quality is just as good if the used voice model is optimized for talking
Yes RVC. I was trying to run FastVC but most of this stuff not working at all or I kill 5 days to installing different version of libraries and rewriting code.

RMVPE is pretty new. I'll check it, looks good. Thanks
 
Can you name a few audiobook distributors where we can upload our audiobooks?
 
Can you name a few audiobook distributors where we can upload our audiobooks?
Just google it. I tried only one of top 10. You can have much better results with different one.
 
Hello, everyone!

I'm excited to share a fantastic method that allows you to earn a few bucks (or even more!) by utilizing AI to create predub audiobooks.

1. Download any royalty free audiobook from torrent website. Many audiobook websites and forums offer a "classic" category that features a selection of older books. A significant portion of these titles meet our criteria. It's essential to ensure that any downloaded audiobook is divided into individual chapters and available in popular formats like mp3 or wav, as this organization will prove convenient for future use.

View attachment 274478

2. Download and install any AI voice converter. Most of the website just using free open source. It will convert original audio to desire voice.For the best results, utilize the F0 harvest method. It has proven to deliver excellent outcomes in my experience. You need to download AI voice model. Use voice which is not mainstream and everybody knows him but also is good for audiobook narrating. If origital voice is male one use male model as well so you dont need to do pitch correction.

3. Split your large audio files into smaller chunks using software like Audacity. This will make it easier to manage and process individual segments. Convert each smaller audio chunk into the desired voice. Utilize the F0 harvest method for optimal results. Once you've converted all the files, merge them back together into a cohesive audiobook using Audacity or any other audio editing software. To ensure compatibility and standardize the audio, convert the merged audiobook to a 44 kHz sample rate.

View attachment 274479

4. Create cover with Canva or Photoshop , use made up narrator name (use same one if you upload multiple books with same voice model).

5. Create your account on any audiobook distributors. Don't forget cut intro and outro and replace them with your own converted voice.

I uploaded 3 audiobooks and making ~$3 per audiobook month.

View attachment 274480
interested, can you check the message I contacted you
 
Thank you for sharing. BTW, I also heard about generate caricature with AI, that may be something worth try.
 
The method is scalable, and you shared useful information.
 
Thanks for the great idea

I am planning to try this method
 
Interesting, will check this greate idea, thank you for sharing.
 
Back
Top