Alex Serdiuk, CEO of Respeecher, answered questions from the audience in a live AMA session we recorded on YouTube. This interview is part of a series of four interviews that will cover topics like the ethics of deepfakes, the synthetic media industry, voice resurrection for Hollywood movies, Respeecher in the context of war, and more.
If you haven’t read the first parts, you can do so here:
Watch the full video of the AMA session here:
[Q] Does Respeecher have job openings for producers with 20 years of experience? If so, how can I apply?
[AS] Yeah, we usually have openings in our delivery team for sound engineers people with golden ears who are able to work with our voice cloning and AI voice generator technologies and know what sound is. You can always send us your CV to info@respeecher.com or reach out directly to Vitalii. He's the one who is in charge of growing the delivery team and it's really cool to work with sound folks who have an understanding of speech synthesizer capabilities.
We have so many talented sound engineers and that's also very exciting that we can help them bring their skills and expertise to the next level to make them part of big projects as we had with Obi-Wan Kenobi and The Book of Boba Fett - also sound engineers from a small city near Mariupol who had limited access to jobs because there are not that many productions happening in Ukraine. Please write to us. If you're looking for sound jobs, I'm sure we have right now openings in the delivery team. So please reach out.
[Q] Can you tell the story about Respeecher? When it was founded? How did you get the idea?
[AS] The company itself, Respeecher Inc, was founded back in 2018, early 2018. But before that, for several years we were playing with the technology and the concepts in pet project mode. And initially, the idea came up during a Grammarly hackathon.
One of the co-founders of this feature, Dmytro, participated with a small team, and they picked this idea to apply machine learning to make one voice sound like another voice. And the hackathon was won. We started to look into some potential applications of the technology. We thought that there is an interesting niche for a product and services built in the speech-to-speech domain.
We heard some initial feedback from filmmakers, mostly from Ukrainian filmmakers, in the initial stages. We understood that they are looking into synthetic voices, because of all the limitations of the solutions they tried. The inability to control performance and all those things make speech-to-speech technology stand out from many text-to-speech technologies in the market. So we thought there was an opportunity here.
We also found out that it's an exciting technical challenge to make this happen, and it unites great minds. Therefore, we were able to create an initial team of very talented people from the deep learning and machine learning area in order to solve this problem. And then in 2019, we already had our first Hollywood movie.
We became a part of Techstars Leaf Labs, Comcast Accelerator. We raised a few rounds of investments. We grew our team. Now we are 42. We keep growing. And now more than ever, we are very much excited about the technology applications and how it changes not only the content creation world but many other amazing use cases, including healthcare, including ensuring that small creators with very limited budgets are able to use our Hollywood grade technology in their projects. That all drives us to grow faster no matter what.
[Q] Do you have any new releases you can announce?
That's a common theme in high-profile content creation. If you’re a Hollywood studio, if you're creating a triple-A video game or TV series, or feature film, you don't want this information to leak out and not be under your control as a creator, because many things depend on that. So when we were in charge of making the voice of young Luke Skywalker in The Mandalorian, you are not allowed obviously, to talk about that. And we were not able to even talk about that for eight months after the release of our episode in season two.
This can be annoying because it's really cool to share amazing projects we are part of. But also it has an interesting bright side because of the way we can surprise the audience, and then see their raw first reaction.
People didn't notice any synthetic sounds in our synthetic speech. It just sounded like it was recorded in front of a microphone. And it's very exciting to see that the technology we invested a lot in, in terms of time, in terms of money, in terms of efforts of a very talented team is capable of doing that.
We can say that in a couple of weeks, a very big game will be released and our work will be featured there. So you will see it once you play the game. We are also working right now on a project for a big YouTube channel with double-digit million subscribers. It's going to be educational and entertaining.
We started working on an interesting project in the audio restoration domain, where we will restore content produced by a famous scientist. Most of you probably know that the quality of this content is usually not good. It was not recorded properly. And here we apply a lot of fresh things from our R&D in order to do not just voice conversion but Nero restoration.
We will use the original, damaged corrupted content to drive the model of the voice from other pieces of content they created back then to restore it, to make it sound good and pleasant to listen to.
We have several activities in our Support Ukraine campaign. This campaign is about enabling famous people to say their words of support for Ukraine in the Ukrainian language. It's very meaningful work for us. We are calling celebrities and famous people to say something to our nation, to people who fight for their existence and defend some values, and to the world.
The same world shares democracy, like the right to choose, the right to decide, and the right of defining their future. And the project is about enabling someone who does not speak Ukrainian, to speak Ukrainian. We already released some videos in the Support Ukraine campaign, and a couple more videos are in the plan.
You might have seen the actress from the Orange is the New Black series, Abby Savage, saying encouraging words to Ukrainians.
Maye Musk, Elon Musk’s mother, also became one of the first supporters of this campaign.
We also have a couple of anonymization projects for documentaries. Those are cases when you need to hide the speaker's identity. And in cases of investigative journalism, it is really important because you might need to interview a victim of a crime and they wouldn't feel comfortable revealing their identity.
Usually, filmmakers just hide them visually by using some contour lights. So you wouldn't be able to recognize the person. And people in documentary film making, investigatory journalism used to apply very heavy voice morphing techniques which do not sound good. They hide a lot of content, and a lot of emotions.
With our technology, speech-to-speech voice cloning could be applied to use the original voice of those people and get it converted into another voice. So, their voice identity would be completely hidden, but the way they say things, and their testimonies, will be retained.
For all the investigative journalists, all the documentary makers that are doing work on pieces featuring victims of the war crimes committed in Ukraine through Russian aggression - we are committed to doing pro bono work to help you conceal the victims’ identities.
We are also working on a big musical biopic. It's very exciting. The release will be in 2023, but our technology will be used across the whole film. It's a really cool and exciting team.
Probably one of the biggest projects we are working on right now.
We are also focusing a lot on the Voice Marketplace. We were obsessed with Voice Marketplace for the whole of 2022. We want to bring it to the next level.
We want more small creators to be able to use this amazing technology. And it's really encouraging to see all those projects that are being done with the Voice Marketplace. They do amazing things. They voice over educational content, and they voice over animation. They do it from their basement.
It's getting to the point where creators will start competing with their ideas, not with their budgets.
I guess this was it. I don't have any more questions from our audience. Thank you for your time. Sorry for any glitches, this was our first live session. We hope to do more in the future, and we also hope to feature more people from our team in the future. People that have a deeper knowledge and understanding of voice cloning technology or sound than I do. We'll announce it soon and feel free to subscribe to our newsletter and please follow us on social media. We post many exciting things and really cool projects there. Thank you so much. Have a good one.