Respeecher's AI Voice Cloning Revolutionizes YouTube Content Creation

Jolly is a popular YouTube channel run by two friends — Josh and Ollie. They came to fame in 2013 from another channel they started called Korean Englishman, with the goal of introducing Korean culture to the masses.

The two friends create different types of funny content and always try to come up with new ways to amaze their subscribers. As they say about themselves on their channel, “On this channel you'll find two best friends who have spent too much time on the internet, and in any given video pretty much anything can happen!”. “Anything can happen” directly relates to the YouTube content of these besties and the way they treat each other.

Last year, Ollie secretly ghost-wrote and published Josh’s autobiography which became a great success for the Jolly duo. However, Josh had no interested in reading his “own” masterpiece, even for a much requested audiobook.

That’s why Ollie stepped in and presented the audio version of Josh's autobiography…  told by Josh himself!

So you don’t get confused, let’s go through all the details of how Ollie created this extraordinary present.

The Challenge

So, Ollie secretly wrote Josh’s autobiography and voiced it himself. But the main idea was to give this story the voice of its hero. That’s why it had to have Josh’s voice. 

The book was around 2,5 hours long and involved lines with different characters. This meant that special attention had to be paid to maintaining certain timbres, voice tones, cadences, emotional accents, and so on. 

Another major hurdle was to deliver the project in the midst of an ongoing war. The execution of the project coincided with Russia’s full-scale invasion of Ukraine. Ollie contacted us to say that there is no urgent need to continue our work on the audiobook under the current circumstances. However, we felt like we had to keep working, produce new projects, and spread the Ukrainian product worldwide.

The Solution

To turn Ollie’s idea into a reality, Respeecher trained models that were capable of addressing the cadences and emotions of the people in the story to deliver the best experience possible. The technology delivers:

• Pitch perfect quality. Our voice swaps are virtually indistinguishable from the original — and always sound natural. They convey all the nuances and emotions of human speech and maintain the highest production value.

• Eliminate text-to-speech pitfalls. Our clients tell us that even the best text-to-speech solutions have a robotic, non-emotional delivery. They also struggle with unusual words, foreign languages, or nuances like humming and giggling. TakeBaker allows for speech-to-speech conversion technology that delivers far superior results.

• Creative control. The tool allows for making changes to the scripted words at any time during the creative process without re-recording the target voice.

In order to achieve natural-sounding results and create a pitch-perfect voice-to-voice swapping model, Respeecher trained a neural model of Josh’s voice. Once the model was ready, Ollie could use it to convert his voice into Josh’s.

The Result

As a result, Ollie’s audio autobiography of Josh manifested as if Josh was telling his own story himself. The audio track is completely natural-sounding, considering the different characters, tones of voice, and emotional accents.

The process of creating a synthetic voice is always engaging and interesting, especially as that voice develops into the final product. But we are always very eager to see (or should we say — hear) the result and get the reaction from the customer. Since the customer was such a creative person, we were happy to see the total range of emotions upon delivery of the completed project. 

Ollie then recorded a video, in which he presented the audiobook to Josh. 

“You are about to become the first person in history to unknowingly read, record, and
release your own audiobook!” said Josh right before presenting the surprise to Ollie.

Just take a look at how surprised and impressed Ollie is!

Voice Cloning for YouTubers

With the help of speech synthesis, YouTubers can free up more time as voiceovers can be done without them. This allows a content creator to spend this reclaimed time coming up with new creative ideas. Here are some of the critical benefits that voice cloning delivers to the YouTube community:

Coping with Burnout

If you follow any popular YouTuber, you know they never take a break. While not always noticeable to viewers, burnout has a considerable impact on the health and motivation of video bloggers.

Voice cloning can not replace a blogger in a video frame, but if you watch any third-person video reviews, it can ultimately save you from studio work for a while.

Any member of your team can dub audio content and then have their voice easily cloned to match yours. The best part is that all this can be done while you enjoy a well-deserved seaside vacation.

Saving time and resources on producing new content

Managers of large YouTube projects know that their audience is ready to consume significantly more content than they are able to release. To a large extent, restrictions on production are explained by the fact that a channel is tied to one or two hosts. And people are always a risk factor.

Voice cloning services can help bring a large team together for content production. This is because production teams no longer have to depend on an actor’s physical presence in a studio.

Coming up with new video ideas

This is perhaps one of the main problems for those who run YouTube channels. How else to surprise the audience?

With speech synthesis, a third party is now able to speak using the voice of a YouTuber. Remember that the host of a channel can speak using other people's voices as well. Imagine a famous blogger suddenly speaking with the voice of an equally famous YouTuber from another channel. Or, for example, using the voice of an iconic movie hero or villain, a child or grandmother, crying like a baby, or roaring like a lion. 

Respeecher frees creators from tedious and time-consuming processes while providing them with more space to generate new ideas for content and their implementation. Try it to see for yourself!