“Very reliable voice AI software, works exactly as advertised in the DAW and one of the few options for offline processing and a one time purchase price. The voices are varied and high quality. I'm now able to sing on my own demos and make them sound like a real singer.”
“I don’t use it entirely to replace the voice, but rather as a combination — and in certain cases, it sounds great, just like using an AI engine to replace the voice.”
“Stunning effect. I managed to turn a solo song into a beautiful male/female duet.
However file handling is poor and I would recommend to render the effect down as soon as possible.”
“The point of difference with this software is that you don't have to go in and out of your DAW. Once you've "captured" your source vocal, you can browse the models and get your desired voice. Vocalistai was the front-runner but this is now. Well done. Real-time processing would be nice, but I get how complex that would be.”
“I am owning a perpetual license and it works great. Sound quality is very good for my needs. The time required to learn the tool as well as the time the tool needs to apply the voice model is very good. Only one thing remains open at this time, but the support team is already working on that: For me the tool "forgets" the link to the processed data. It still realizes that there is an audio available, but I need to "reconnect" this so that I can hear the voice model instead of my own voice.”
“Setting VoiceAI up as a plugin in my DAW (Reaper) worked similarly to a VST instrument rather than a normal EQ/effect, meaning that the source track is read in it's entirety, transformed and a replacement WAV file written out for the DAW track. This took a bit of time to complete on my Intel I7/64gb system, but seemed to work consistently enough. The first issue I had was in establishing the chain sequence of effects on the vocal track - it had Melodyne already in place and this seemed to cause VoiceAI a challenge - probably because Melodyne also reads the whole track and creates a "fixed" file - and VoiceAI seemed to read the non-Melodyned version, which wasn't my intent. In the end I had to put VoiceAI on the BUS/Folder track that the vocal was inside. Then I had to reorder my BUS effects (I had Nectar also on the BUS) so VoiceAI was first in the chain. The second issue was that some of the notes on the vocal track were glitchy after VoiceAI processing, which may well be because they were not in the "Best Key" recommendation for the VoiceAI model selected. I appreciate that the "Best Key" recommendation is there to ensure the optimum outcome, but vocals aren't that easy to align with the plugin's needs, so would be good to see the processing improved on each of the models so that a wider range of keys can be accomodated without glitching. I hope this will improve with development - but for now the product seems a bit finicky.”
