@Eman:
I know, I’m hijacking my own topic in the end, but I’d like to reply to your concerns. First of all, I have to admit that I’m completely new to this whole SV2 universe. Therefore, I can only speak for myself and my personal use cases. From my perspective, SV2 is a great app/plugin to achieve voice recordings which also sound natural in the end, if you’re looking for those kind of attributes. If you’re just down for all that mainstream “T-Pain” auto-tuned sounds, there are other and better ways to achieve that goal. But if you want to be able to create lead and backing vocals which hardly can’t be differentiated from real ones (provided you are willing to enter a deeper editing level - which is fully available within SV2), you can achieve almost anything you can think of. You are even able to separate the vovals and consonants of a particular voice from each other while exporting the outputted sound to separate audio tracks, meaning that you can mix and re-combine both components individually again with different levels applied. For example, in case you want to avoid an extensive use of de-essers or similar tools, you can just reduce all aspiration sounds without compromising the main voice itself. In contrast, by reducing the vovals to a certain minimum amount, you can even achieve complete whispering effects instead. And by applying the new parameter for mouth openings (in conjunction with all the other available parameters), you can completely change the way words are sung. Of course, it takes some time to experiment which all these individual parameters and to find out how these will affect the currently selected voice (this differs from voice to voice), but in the end it’s worth it. Also vocal runs ala Mariah Carey etc. can be easily achieved, and they sound naturally (in contrast to other applications I know). In general, I’m very impressed about the way a user is able to operate the software by starting at a relatively simple level already. Just entering notes and listening to the very first output - without applying any manual intervention yet - will most likely sound already how you would have asked a real singer to perform it. And if you don’t like all takes - or just some parts of it - no problem to tweak the performance until you’re satisfied. Of course, it will never be perfect; but hey, with the right effects and a groove that sits in the pocket, you get stunning results.
From my perspective you get much bang for the buck, at least for the application/plugin alone already. The individual voice purchases are relatively expensive, but at least it’s also fair to compensate the real singers behind these artificial voices for their efforts. For my personal taste, there could also be more “voice bundle extensions” available in the stores, especially if customers are willing to buy at least 3 voices or more at once. That’s something I’m missing here, and I’m not sure why this hasn’t been considered as an option to attract more customers. And yes, the time for running the offered SV2 trials is really a little bit too short in addition to the fact that the number of playable notes is anyway limited. Not much room for extensive testing. You really have to hurry up before a particular voice trial expires if you want to check out how each individual voice behaves, especially in terms of the ability to speak acceptable English. While for judging the general timbre and color of a particular voice the available audio demos on the Dreamtonics and Eclipsed Sounds homepages might already be sufficient, a meaningful verdict about the correct English pronunciation might be more tricky. Unfortunately, there are not always (acapella) song demos available in English, especially for voice providers for which English is not their native language. The only way is to load the related trial down and to test each voice individually by applying your test lyrics one after another. My personal judgement is that Chinese voice providers perform much better in English than Japanese ones, but that’s solely my own opinion after the voices I have checked out already. Don’t backfire at me because of my 2 cents.
To recap, even if there might be still bugs in the software (which is very often updated as I have experienced so far), for me it was a no-brainer to go for SV2 right from the start. There are so many apps/plugins out there which deal with AI voices. Most of them are cloud- and/or subscription based (which I completely reject), and - at least for me - they don’t even come close to the quality that SV2 finally offers (and no, I’m not from Dreamtonics or Eclipsed Sounds). In the end, it’s really up to you and your use case if you’re finally going for that option or not. For me, the investment has already paid off. I’m definitely going to buy more voices in the future in case they also fit to my forthcoming projects.
; however, it would be a great benefit to instantly hear the note which is to be actually manipulated. This could even come in very handy when entering lyrics, where you usually press the tab key in order to hop to the next note in the chain. Playing the currently selected note (for about 250ms or so) would be already enough to get proper audio feedback in order to determine if the current note and its pitch is correct before proceeding. To be clear, it doesn’t have to be a real-time interpretation of the SV voice singing that note; it would already be sufficient if that famous “default tone” could be played. Yes, I’m referring to the one which is audible in case a SV2 track has already notes in it, but no voice has been assigned yet.