Auto-Tune changed the music world forever. Some artists achieved fame thanks to that audio processor, while others claim it destroyed the industry entirely. Now we've entered the era of artificial intelligence, and history seems destined to repeat itself. On one side, there's absolute rejection of algorithms; on the other, services like Voice-Swap emerge, specifically designed for artists and producers who want to change the voices in their own songs.
Anyone closely following the writers' and actors' strike in the United States knows that artificial intelligence is one of the main points of conflict. Generated stories, three-dimensional models, complete face replacements... many people are furious, scared, and worried. The profit-driven mentality of the big production houses doesn't help, leaving technology in a very bad position with the public.
However, there are always those who see these tools as a new opportunity, and that's where Voice-Swap points. Its goal is clear: provide resources so musicians and producers can change/replace the voice in their songs, using models from artists who have already joined the platform, offering special licenses to legitimize the process.
Voice-Swap: Is There Room for Artificial Intelligence Among Singers?
Before you begin, Voice-Swap makes it clear that it is not a tool for general consumption. The musician or producer must have all the corresponding legal rights to their samples. Voice-Swap uses ACRCloud to detect unauthorized use, and its technical limits tell us about WAV files with a maximum size of 30 megabytes. The alternative is to connect the microphone and sing directly so the system can record it.
The second step is to choose one of the eleven available models, and finally adjust the pitch so the tone is more intense. Voice-Swap takes a few seconds to make the transformation (generally no more than 30), and the last options are downloading the result in WAV or storing it in our audio profile.
Needless to say, Voice-Swap doesn't perform miracles. Its official email warns from the beginning: "Rubbish in, rubbish out". The audio must be clean, in a single layer (that is, no more than one voice at a time), without distortion, delays, or reverb... nothing. In fact, it recommends a session with Auto-Tune or GSnap to improve its condition before uploading.
My tests were very negative, and I imagine a professional's demands will be much higher, but time is on Voice-Swap's side, and any algorithm can be optimized. Finally, the trial provides only a handful of credits. You can buy more for $8, or enter one of its plans, starting at $6 per month.
Official site:Click here
