AI is becoming more and more powerful. Now you can even cover songs directly with other voices. The “Replay” I will introduce in this article is a free tool that can cover all songs with AI. The operation is very simple. , read the song into the YouTube video URL, select the vocal you want, then press Process, and the voice will start to change. Even Chinese songs are supported.
In addition, if you want to make accompaniment music, that is, remove the vocals, this tool can also do it.
Advertisement (Please continue reading this article)
Replay uses AI to cover your favorite songs, separate vocals and accompaniment, and supports YouTube
Replay supports Windows and Mac operating systems. Click the link above to go to the official website to download the installation file. If there are no requirements during the installation process, just click Next. After installation, open it and you will see this screen, press CONTINUE:
Advertisement (Please continue reading this article)
Before use, you need to download the AI model data. The file is quite large. In my impression, it should require at least 5~6GB of space. It is recommended to keep more than 10GB of space to be safer:
Another thing to note is that this is a local AI calculation, so it will use your CPU and GPU resources. For people with insufficient hardware specifications, it may be a bit difficult to run, which means it will take a lot of time, and It will take up a lot of resources. It would be better to have a GPU, as the computing speed is much faster than a CPU. It can be run without a GPU, but it just takes longer.
Will be downloaded twice:
The RAM memory part requires at least 8GB, and 16GB is the recommended specification. If it is less than 16GB, it may occasionally crash:
After entering the main screen, there are three main areas. The top right part is the part for importing songs. You can manually upload, record your own voice, or paste the YouTube URL (a sample audio file will be provided for free by default, click X can be canceled). Below is the sound model you want to apply. There are many copied voices of singers, as well as some cartoon and animation character voices. By default, not all will be downloaded. When you need to, click the download icon on the right to download it to your computer:
Settings settings are some more advanced adjustments. It is recommended to use the default settings at the beginning. Once you are familiar with them later, you can slowly try each setting.
I tried to lose the YouTube URL:
After successfully reading, there will be a preview image, and then I will cover the vocals of this song with AI animation sounds. After selecting, press CREATE SONG at the bottom and the AI cover will start:
As for how long it takes to complete, it depends on the specifications of your computer. Those with NVIDIA graphics cards will be much faster. For example, I used an RTX 3080. This Mayday song was completed in less than a minute, which is very fast.
After it is complete, you can listen to it. The effect is pretty good, and you can cover it again (REMIX AGAIN), and the sound will be different every time. It’s a pity that there is no download function, but you can create a sharing link.This is the audition file I sang using AI：
In addition to songs, any sound file can be sung or narrated by AI. For example, if you have some sound files that you don’t want to use the original voice, you can use this to change it. In fact, it is a bit like a voice-changing function.
In advanced settings, Stem Only is a function that separates vocals and accompaniment: