ElevenLabs, a startup dealing in generative audio AI, has introduced a groundbreaking tool that can generate a broad range of sound effects sourced from a written text prompt. This development portends significant implications for independent game developers, film students, and any individual in need of a quick sound effect without the financial backing typically associated with such services.
The tool named Text to Sound Effects is the latest audio model from ElevenLabs; it is designed to generate sound effects of diverse formats spanning from short instrumental tracks to character voices and soundscapes. The only information required from the user is a text prompt and the AI effortlessly brings the sounds to life.
The operation of the tool is straightforward. Users simply type in their prompt, hit the generate button, and in a matter of seconds, they have their sound ready. For each prompt, the program generates four sounds, granting the user a variety to select from. If unsatisfied with the results, the user can adjust their prompt until the desired sound is achieved. Additionally, prompts for the tool can be up to 450 characters long.
Despite its exciting prospects, the tool’s quality is inconsistent. The resulting audio tends to differ, with their duration hinged on the prompt. For instance, a prompt like “monkey falling down a long staircase” would yield outputs varying between 6 to 16 seconds, while “an elephant suddenly achieving enlightenment” would result in shorter audio typically between 2 to 5 seconds.
To further enhance the user experience, any generated sound effect can be directly downloaded from the program onto the user’s device. However, to perform this action, users will have to register a free account with the service. While this tool’s potential seems boundless, the company has put checks in place to prohibit inappropriate use. It will not produce just any sound. All prompts are regulated by its Prohibited Content and Uses Policy, which forbids any content, pattern or use that could be deemed violent, hateful, illegal, or abusive.
Eager users can test the tool on the ElevenLabs site to learn more about and experience its capabilities first-hand. Hence, with such a highly innovative product, ElevenLabs has painstakingly attached a new dimension to audio AI, blurring the boundaries of possibility in the space.