Project Gutenberg has been providing free ebooks to the world since before the Internet existed. It was founded in 1971, when founder Michael S. Hart typed the U.S. Declaration of Indepdence into a mainframe and made it available over ARPANET.
Fast forward and in 2023, they have over 70,000 documents. These are books and texts that are out of copyright/public domain, so don’t expect to find Harry Potter there, but you will find a lot of classic literature – everything from Sherlock Holmes to Shakespeare to Dracula and Frankenstein.
The formats have evolved as well. Originally these were text documents, but today they’re very e-reader friendly, suporting EPUB, PDF, and other formats.
But What About Audio?
Producing text copies of books is relatively straightforward and does not take a great amount of effort. But what about audio?
That’s a lot more work. You need to have a narrator who takes the hours and hours to read a book, correcting mistakes, editing the audio, etc. Of course, just recording itself is critical – it can’t just be some guy talking into his AirPods, but rather someone using a professional setup (if you doubt this, listen to a few previews on Audible and see how much the microphone can change the experience).
There is a project that’s been putting Gutenberg-type books into the public domain since 2005: LibriVox. They have over 15,000 audio books available for download, and they’re all read by volunteers.
In my experience, the quality is uneven, both in recording and narrator talent. Personally, if I’m paying Audible for an audiobook, I expect a high quality recording by a narrator who is top notch. But for a free audiobook, my expectations are lowered…and there’s a certain charm to having books read by volunteers. It’s more like a friend reading a book to you than a “stage production”. Also, there’s often several different recordings for popular audio books so you can choose from different narrators. And of course, you can’t beat the price!
And Now, the Robots
Microsoft recently started a collaboration with Project Gutenberg to record its audiobooks via AI:
A team from Microsoft approached Project Gutenberg about a collaboration to produce thousands of high-quality audiobooks using an AI-driven solution and then give them back to the Project Gutenberg community. These new audio recordings have made Project Gutenberg’s books more accessible to a wider audience of people around the world, including those facing accessibility challenges.
The capability turns the text of each book into audio using advanced human-like voices that can even convey emotion. “This is an AI innovation that reads text in a lifelike voice,” Newby explains. “The voices are trained to mimic humans in order to sound natural, and the result is convincing—a big upgrade over older versions of text to speech.”
Some of the AI-generated audiobooks, such as recordings of Shakespeare’s plays, include different voices to portray various characters. They can also generate human-like voices in other languages and adjust the voices to read with more emotion.
And the time to generate these audiobooks? 30 seconds per book.
Gutenberg has made some stabs at automated audiobooks in the past, but they’ve been execrable (example).
The new effort sounds much more pleasant. Here’s an example:
Will You Listen to a Computer Reading to You?
Of course, this has been done for decades in an accessibility context. There aren’t professional narrators for each and every web page and email someone needing accessibility support might need. But here we’re applying the technology universally for all.
I’m not sold on AI readers. They speak perfectly, but I miss the warmth of knowing it’s an actual human reading – the occaisional flubs, the choices (good and bad) about voicing and emphasis, etc. We’re evolving into a three-tiered model:
- Best quality, professionally-produced works such as you find on Audible (though certainly not all Audible productions!)
- Volunteer-produced works (LibriVox). These sometimes are just as good as the pros, though the distribution skews downward a bit.
- AI-generated content.
What do you think about all of this? Let us know in the comments below!
- Glitch Servers’ 30% Off Winter Sale is Now Live! - December 8, 2024
- Have You Missed Any of these LowEndBoxTV Videos? - December 7, 2024
- Black Friday Never Ends!Check This Out: VeloxMedia has a 1GB VPS for £10/Year in California! - December 6, 2024
Leave a Reply