Media Pool

The Media Pool brings audio files within reach as part of the interactive workflow giving you unprecedented access to begin, or continue, content creation directly in Wwise. For the first time you’ll be able to easily find and navigate to audio files referenced in the project while also unlocking direct access to other audio file sources. Sound libraries that contain a large number of files can be added as databases in the Media Pool where innovative search tools can find the sounds you are looking for quickly. You can also create filters as search criteria to further refine the list of results, along with the ability to toggle them on and off, to preserve common search criteria for later use. Once you find the files you need, you can audition them, modify the playback properties while auditioning, select regions, and import your final selections into your Wwise Project.  

The Media Pool is where you can search, organize, and preview your audio files to import into Wwise. This is invaluable if you have personal sound libraries, downloaded Strata libraries or third-party libraries. These libraries often contain large numbers of files, which can make it time-consuming to find what you need. You can add these library folders to Media Pool databases, and then search for and audition your files directly in Wwise. You can also apply filters as search criteria to further refine the list of results. After you find the files you want, you can audition them in the Transport area of the Media Pool. Additionally, you can modify the playback properties while auditioning the files and optionally apply these changes when you import them. The Media Pool allows you to import your final selections into your Wwise project, completing the workflow all in the same view, in Wwise.

Similar Sound Search

Similar Sound Search is an audio-to-audio and text-to-audio search functionality specifically designed for sound effects. This collaboration between Sony AI and Audiokinetic brings together cutting-edge research and industry-leading tools to support audio professionals across game development and interactive media. The search capability is powered by an AI model trained on professionally licensed, professionally-created sound libraries, ensuring production-ready results. The sound libraries were provided by Pro Sound Effects, BOOM Library, and Strata. This feature is offered as an optional way to surface additional results as part of the Media Pool in Wwise.

Think of Similar Sound Search as having a dedicated, reliable assistant for your audio library.  It leverages a deep learning model to place audio and text into a shared search space where sounds are grouped by both acoustic and semantic similarity. Similar Sound Search allows sound designers to search for sounds using either a text query or a reference audio clip. This enables the discovery of new, unexpected matches that would be difficult to find using conventional text-based search methods. For example, a sound labeled “smashing fruit” might surface as a match for “footsteps in mud.”

In particular, Similar Sound Search allows for:

• text to audio retrieval, where a description of the sound is used to find a sample based on its audio content only. This feature is a new way to find audio that would match the description, even if the sound was recorded with a different purpose. e.g. a sound tagged and obtained by “smashing fruits” could be used to create “footsteps in mud”. Such a match would have never been possible with traditional search.
• audio to audio retrieval. Sometimes it’s easier to provide an audio example to use as a way to search for similar sounds. The provided audio example and its sonic qualities can be used as a reference to surface relevant and ranked results that are understood abstractly by the trained model. 

In both use cases, Similar Sound Search can be a valuable tool that opens your ears to sound collections in a new way. These two techniques naturally enrich the traditional filtering using metadata and work well together to extend the possibilities of existing sound libraries.

About Similar Sound Search:

What is Similar Sound Search? Similar Sound Search is an AI-powered audio-to-audio and text-to-audio search capability.

What technology does Similar Sound Search use? A non-generative AI model is used to retrieve similar audio results from user-defined databases as part of the Media Pool.

How do I get Similar Sound Search? Similar Sound Search is an opt-in addition during the Wwise Installation.

Where can I find more info about using Similar Sound Search in Wwise? Check out the Similar Sound Search Documentation.

About Similar Sound Search’s use of AI:

Does Similar Sound Search leverage Generative AI? No, the non-generative AI (CLAP) model is stored and used locally.

Does the Similar Sound Search feature connect to the Internet? No, the database used for searching is stored locally and not accessed externally.

Does Similar Sound Search collect or send my data to Audiokinetic? No, the CLAP model and database is private and Audiokinetic doesn’t collect any of this data.

What data does the Similar Sound Search use to train the AI Model? Licensed Libraries were used to train the AI Model. (see above)

Does the Similar Sound Search use my data to train the AI Model? No, the non-generative AI (CLAP) model is local and is not updated.

Does the Wwise EULA allow Audiokinetic to use my data to train AI? No, Audiokinetic will not use your data to train AI.

Summary

Wwise 2025.1 continues to improve on Wwise workflows fundamental to the creation of immersive audio experiences. Many of the changes in recent releases speak to our overarching goal of unifying workflows while leveraging common user interface skills. These changes aim to increase usability while removing friction from daily tasks. We’re excited to be working on features that can bring people closer to their audio by leveraging the possibilities of AI, Machine Learning, Neural Networks, and other emerging technologies. AI development is conducted internally using licensed data, and no user data is used as part of these processes. Read more about Responsible AI at Audiokinetic. For a summary of additional features check out the Wwise 2025.1: What's New | Beta Edition.

The Wwise 2025.1 Beta is live and now available to install from the Audiokinetic Launcher. The build can be accessed after logging into the Launcher and navigating to the Wwise Tab. From there, scroll down to "Install A New Version" and use the drop-down menu to select "Special".

Note: This Beta version cannot be used for official release and will not be supported beyond the launch of Wwise 2025.1.