Shift

VST version of Shift is strongly desired. Depending on the recording environment, there is a desire to equalise or gate the sound before inputting it into Shift, and after audio conversion in Shift to make it more audible by equalising the sound. To achieve this currently requires a complex structure like DAW=>Shift=>DAW. Also, if Shift were to become a VST, I would be happy to benefit from ASIO and have a little less latency. I was very surprised by Shift's low processing load and clean conversion. We look forward to the further evolution of Shift! (I am writing this English text using machine translation because I am a Japanese speaker, so I apologise if the text sounds strange.)

The current version of Shift sounds very good with vocals but outputs glitchy, incomplete sounds with non-vocal noises such as laughter or coughing with all the voices. Quieter sounds such as breathing tend to be eliminated entirely, but laughter/giggling is a common and important part of streaming and content creation so I think this aspect of the AI needs to be improved to make it usable for a wider variety of expressions and emotions.

213

VST?????

After updating to 0.1.3, the voice change sometimes fails when I laughed, and the sound is choppy and robotic. This did not happen in 0.1.2. I liked the old version better, so if you can revert it, please do.

I would like to be able to use it in places where the internet is not stable, so it would be great if it could work offline.

An adjustable compressor would be very useful on Shift. A simple compressor slider/dial to turn it up and down. Essentially reducing the dynamic range of voices when speaking into it and acting as a limiter. I notice when talking into shift, it is quite easy to clip audio as when talking, naturally you speak louder and quieter sometimes which means you have to be careful not to speak too loud and clip, but also not too quiet so no one can hear you. I can provide more information or examples as and when needed.

There are still bugs and parts that are out of reach, but it's amazing to be able to use this quality AI voice changer for free in the beta version. As a request, I would like the recommended environment and required environment for the software to be described in a little more detail. It is necessary in order to know in advance how much leeway there is when running at the same time as a game or the like.

I'd like to see a feature where you can upload custom AI models, like pth/index files to use as a custom voice. Models that have already been trained where it uses Supertone Shift technology to run the models on the program. If it's possible anyways, I know pth/index models can be pretty unoptimized,slow and delayed on some programs to use efficently. Give it some fine-tuning features like a wider pitch-shifting range like +50/-50 instead of +12/-12. Thanks for reading this suggestion!

I have already purchased Clear and I am very satisfied. “Shift” has a lot of potential that can improve over time. The voice changer is accurate and believable, but sometimes it does not recognize some words or pronunciations in Italian: the words sound a bit strange. It’s a fun tool, but if the pronunciations become more accurate and the expression of emotions more similar to the speaker’s voice, it could become a very useful tool. I could use it professionally to create audio stories with various characters. I hope that the cost of the final version will be affordable for me. Please note: last week I wanted to use it, but all the characters were unusable, the voice editor had turned gray. Now it’s working again. I tried to reinstall the software, but it kept not working. In my opinion, it is important to warn users in advance of possible changes that could prevent the use of the software; perhaps warn a couple of days in advance, especially out of respect for the work of those who intend to use it professionally. Thank you for your work, it’s very nice. Greetings from Italy.