> For the complete documentation index, see [llms.txt](https://support.audacityteam.org/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://support.audacityteam.org/au3/music/isolating-or-removing-vocals-from-a-song.md). # Isolating or removing vocals from a song This page describes some methods to try to isolate vocals in stereo tracks. ## Using AI to separate vocals {% hint style="info" %} This functionality is only available on Windows and Linux {% endhint %} You can use the Intel OpenVINO Music Separation plugin to separate vocals from music. To do that, first download and install the plugin as described here. {% embed url="" %} Download and installation instructions {% endembed %} Once installed and Audacity is restarted, you can find it in **Effect -> OpenVINO AI Effects -> OpenVINO Music Separation**.

OpenVINO plugin

Here, choose the appropriate separation mode. The 2 stem - Instrumental and Vocal works well on most songs, while the 4 stem version can be useful to extract only drums and bass, if a song happens to feature these instruments. The Inference Device setting lets you choose what processor should do the work: * CPU always works, but is typically slowest * GPU is typically fastest on high-end computers * NPU can only be used on modern Intel processors, and is similarly fast as GPU. {% hint style="info" %} Note: This effect takes a lot of computing power, as do all AI effects. It tends to be faster on a second run, after the model has warmed up. {% endhint %} ## Manually separating the stereo field A faster, but typically less reliable method of removing vocals is to take advantage of their typical position in a stereo field 1. Select **Split Stereo to Mono** from the stereo audio track dropdown menu\ ![](https://2387260374-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-MhmBVzGzh8SctWQ6jPR%2Fuploads%2FnDItvaOtrBfQZe81qwEm%2FRemove%20vocals%20-%20Split%20Stereo%20to%20Mono.png?alt=media\&token=589c289a-f94a-41e9-84a1-d04449d4bb2f) 2. After splitting the stereo track you will end up with 2 mono tracks similar to this:![](https://2387260374-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-MhmBVzGzh8SctWQ6jPR%2Fuploads%2FGHdMHszfsCVI2KvHO8pl%2FRemove%20vocals%20-%202%20Mono%20tracks.png?alt=media\&token=c9096aee-aa6c-4284-9c78-0b6e3b0914aa) 3. Invert one of the mono tracks by **selecting** it and then choosing **Effect -> Invert** 4. Play back the result. This will remove everything panned in the center, not just vocals and returns a dual mono result (both channels have the same audio). In some music this could mean removing instrumental parts. Removal of the vocals can often be incomplete leaving artifacts behind; this is especially true where there are backing vocals or where reverb (echo) has been applied as this spreads sound sources and makes them very hard to extract from each other. ## Alternative AI models {% hint style="info" %} **Note:** This is an experimental feature not part of the normal Audacity installation. {% endhint %} To use AI models in Audacity, you first need to download the current alpha with this feature from Once you have installed this version, you can download and apply AI models via **Effects → Deep Learning Effects**. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://support.audacityteam.org/au3/music/isolating-or-removing-vocals-from-a-song.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.