Mirelo, a Berlin-based artificial intelligence company that automatically generates sound effects for videos, has secured $41 million in seed funding.
The round was invested by Index Ventures and Andreessen Horowitz (a16z), with participation from Atlantic.vc and TriplePoint Capital.
Founded in 2023, the startup develops fundamental models to generate sound effects synchronized to video content. Mirero It says it addresses a gap in generative AI development.
Mirello said that while AI is rapidly advancing in the generation of text, images and video, audio technology is lagging behind. The company says video creators currently spend hours searching through stock libraries and manually syncing sound effects to visuals.
Mirelo claims to generate audio that matches your video “in seconds.”
CJ Simon Gabriel“Think of the difference between a talkie movie and a silent movie. A video without sound has far less emotion and atmosphere,” said Mirero co-founder and CEO.
“Our larger mission is to be the audio layer for all visual content, including video, games, social media, and movies.”
CJ Simon GabrielMirero
“Mirello's first step is to democratize access and allow anyone to create the sound they deserve for their (AI) videos. But at the same time, we take care of the tedious tasks like synchronization, while empowering professionals to rework audio, do more of what they love, and be more expressive and imaginative to achieve it. Our larger mission is to become the audio layer for all visual content: videos, games, social media, movies, and more.”
Founder Simon-Gabriel and CTO Florian WenzelI met him when I was working as an AI researcher. Amazon Web Services Lab. Simon-Gabriel holds a PhD in machine learning and causal inference. Max Planck Institutewhere he studied under a computer scientist Bernhard schelkopf. He completed postdoctoral work Ethereum zurichWenzel received his PhD in deep learning. humboldt university worked before google brain.
Recently released Mirero Mirero SFX v1.5a video-to-sound effects model accessible via API and web applications. Mirero studio. According to Mirelo, the model requires “50 times less computation” than typical large-scale language models, while delivering “better quality than any of our competitors to date, according to external evaluations.”
“There is a deep affinity between music and engineering,” Wenzel said. “Maybe that's why so many of the Mirello team are musicians, and why musicians have always been early adopters of new technology.”
“There seems to be something at the intersection of mathematical precision and expressiveness that attracts people to both fields.”
“There's a deep affinity between music and engineering. Maybe that's why so many of Mirello's team are musicians, and why musicians have always been early adopters of new technology.”
Florian WenzelMirero
georgia stevenson“Sound is too often an afterthought in video production, but it's what determines whether a video or game truly resonates with an audience,” said the partners at Index Ventures, who led the investment. “Mirelo gives creators a new form of expression, enabling them to move faster and sound better.”
“The team led by CJ and Florian is combining cutting-edge AI expertise with an unparalleled focus on the emotional power of audio. This combination will reshape the way the world experiences sound.”
“Sound can often be an afterthought in video production, but it can determine whether a video or game truly resonates with the audience.”
georgia stevensonIndex Ventures
Index Ventures has also invested in other music technology companies, including a futuristic musical instrument company. Loli and concert discovery services. song kick. In 2019, it participated in a funding round for a San Francisco-based crowdfunding membership platform. Patreon.
Guido AppenzellAndreessen Horowitz, partner at Andreessen Horowitz, also commented on the investment in Mirero, saying, “To date, a16z has invested in several world-leading generative models, each with a different focus. Mirero is working on a model focused on sound effects creation, one of the most technologically challenging and least explored areas of generative media.”
“CJ and Florian have assembled a research-driven team whose breakthroughs in tokenization, data curation, and conditioning rival much larger efforts. We are excited to support Mirelo as they scale their technology for next-generation video models.”
“Mirello is working on a model specifically for sound effects creation, one of the technically most challenging and least explored areas of generative media.”
Guido AppenzellAndreessen Horowitz
This marks a16z's latest investment in the AI space. Venture capital firm invested in ChatGPT developer OpenAIElon Musk's AI venture xAI, a data and AI software infrastructure company data bricklaunch AI audio eleven lab.
Last year, a16z United States Copyright Office (USCO) Training an AI with copyrighted material is fair use and does not constitute intellectual property theft. “When training an AI model on a copyrighted work, the goal is not to preserve the potentially copyrightable content (i.e., protectable expression) of the work being trained. Rather, the training algorithm is designed to use the training data to extract facts and statistical patterns from a wide range of content examples, i.e., non-copyrightable information.”
world music business
