• We're currently having issues with our e-mail system. Anything requiring e-mail validation (2FA, forgotten passwords, etc.) requires to be changed manually at the moment. Please reach out via the Contact Us form if you require any assistance.
morrysillusion

morrysillusion

v flower enthusiast 25 From Socal
Messages
776
Reaction score
4,375
Medals
5
Joined
Last seen

Profile posts Latest activity Postings Awarded medals About

  • morrysillusion
    ive shared a few lil tests of the diffsvc voice i trained off flower bc its fun rendering for myself but i dont really plan to do anything else w it, but its no surprise ppl keep asking me for it. and i just dont rly get it. like its not even hard to train so you can do it yourself p easily anyways, but also? i know its so not legal to distribute! ive faced years of ppl asking me for my sonic/shadow utaus. im not giving in now lol.. pls have some awareness
    morrysillusion
    there is a lot of excitement in like... ease of production, when it comes to stuff like diff svc-- ive always had a back and forth with the ways to make your own singing voices, in the sense i just dont have the time or energy to produce vcv/cvvc/vccv utaus for whatever language... or perhaps tackle a NNSVS voice, etc. even if i were to sit down and do it, it would take quite a while. and so im excited to find ways i can use diff svc to assist my music production. BUT i still have a lot of love for the making of utaus, nnsvs, etc. the feeling of constructing a voicebank, tuning it, etc is a whole other unique process. i think its important to know what that feels like. and knowing when i want to use that. hell, just going through the work anyways of making one can still be its own labor of love. i still want to, and plan to, produce some utau banks. id still love to make a nnsvs bank!
    morrysillusion
    the conflict for me w dreamtonics' vbs is like. theyre not all perfect of course but the more they make them i feel like the other issues aside from general samey-ness is that the demos too are just super lacking. i hear other people's covers on release and they convince me far more that they might be worth a purchase-- their quality is usually consistent and in general the way i see them used by others in the community shows their strength far better. i dont give in to it bc i know they have released SO many and in the end they add up to all being similar in ways that makes me know i dont want like 10 of them lol. but seriously the official demos i feel like turn me off to them, and the fan works make me see that often times, these voices are fine and have potential in their usage. its just a lot of the same thing.
    morrysillusion
    unless i properly practice voice acting i dont think i could make a diff-svc model of my own voice and want to release it... itd be a super useful tool to me, and my own work/music! but i wouldnt wanna give it out. its kinda.... scary how accurate it can be with the right, clean samples tbh.
    • Like
    Reactions: ___ and TheStarPalace
    ___
    ___
    nah i feel ya, i wouldn't even feel comfortable releasing utau vbs made with my voice :LOL: the thought of people being able to play back the raw recordings of me going "aaiaueaa~ 🎶 " makes me cringe rtyui
    morrysillusion
    im too excited about this Asterian cover im gonna do so i just want to say what it is even tho i have nothing to show: im gonna cover the oogie boogie song from nightmare before christmas. if i month passes and i dont post it come bug me to finish it
    morrysillusion
    oh wow diff-svc got a lot of updates whle i was away. you can easy pick pretrained stuff in training settings now??? i may retrain the model i was doing for a test since i couldnt figure that out at all before. i wass only at 36k so...
    morrysillusion
    i made a last minute decision to go to MFF (like literally yesterday) bc i just havent been anywhere since the pandemic started and loads of my friends are going. but that means i wont get to use Asterian on release day....
    morrysillusion
    hmmm. i dont join a lotta random discord and partake in them, but had little options with diff-svc. and while the people have definitely been fine, and the help channel hasnt been bad (no ones been an outright asshole), there is like this Culture of communication i see there, that seems especially common with younger ppl or sites i dont go (like tiktok. which is often younger ppl than me). and its like definitely one i dont understand, or read as friendly?? it reminds me a lot of ppl younger than me who act 'aggressively friendly' in that like... 'im gonna mke jabs and jokes at you even tho we arent close and act like its cool bc we're friends'. lots of emoji reactions in the discord are laughing emojis and memes, to people making mistakes and asking for help-- and while i certainly laugh at myself about these annoyances bc id rather laugh than get mad.... theres something just a little off about it. like idk these people? maybe its the autism, idk. like im not friendly enough w these strangers to know the intents of these funny emojis are good ya know
    • Like
    Reactions: WyndReed, Mika, MillyAqualine and 5 others
    View previous comments…
    morrysillusion
    morrysillusion
    yeah im really never a fan of discord being used like a help forum (i hate when so many things in development these days also just go to discord for all updates and info in general), and now discord even....has a forum feature??? i guess? may be nitro/boosted servers only. but id much rather have a forum where i can give loads of details in a big post. on discord it feels like youre being a nuisance when you provide a 1k word support post with five screen shots of the problem but im just trying to give detail!! and then some doesnt respond and just reacts with a doge face emoji so
    MillyAqualine
    MillyAqualine
    Oh yeah like the ones cursing and throwing lmao every two words, at one point it blurs completely the information
    morrysillusion
    morrysillusion
    right, its the kinda casual rough banter id expect like, in my own friend server maybe? but not really on a discord full of strangers, especially in a help channel
    morrysillusion
    my main troubles with diff-svc is just that i am having issues getting it to continue off of my past training;; i only manage to get so far before i leave it over night (trying to take precautions to keep it active) and it ends so, this has been very annoying lol.
    morrysillusion
    probably one of my biggest drawbacks for v5/6 is all the rendering it does when you make changes.... i felt like V6 was actually a bit slower but that was at release so, idk if it got improved in any bug fixes. but i remember it being very very annoying for every tweak i made. i think if it didnt do that id be much more eager to use it for things
    morrysillusion
    rip having troubles with diff-svc and i have had no solutions in the discord lol. explained my problem and all i got was "huh". but i rendered something w a 12k step model for now and the fact its understandable is wild!!
    • Like
    Reactions: MillyAqualine and ___
    View previous comments…
    ___
    ___
    When things go smoothly then diff svc is relatively easy ( aside from the babysitting when it trains which is pain ) but troubleshooting it is on whole other level 💀
    morrysillusion
    morrysillusion
    YEAH... like its a breeze otherwise when i have hit these hiccups. rendering the audio, no issues. training and everything was fine until i obvs ran out of gpu to use and got cut off....hopefuilly i will figure out the problem lol
    ___
    ___
    I ran into so many issues before I even began training it....then it worked fine when I returned to it few hours later uytre hopefully it's similar situation for u too.
    morrysillusion
    tho as intrigued as i am by vocalo changer/similar it brings me back to that point i made about like. the point of "covering" a song by just taking the original vocalists' vocals and putting your voice filter on it. i think its best use (which is what i have mainly seen) is using existing vocal synth renders to "cover" instead considering thats one of the few ways you could "tune" your generated render.
    morrysillusion
    morrysillusion
    i think for me the interest and appeal is ease of use. making a "voice filter" in a sense, and not a whole bank of bits and pieces.... is easier for me, and then getting to tune it second hand in a different software like synth v (which again, easier for me than say, utau) is great. of course theres the nuances of it-- slapping your AI voice filter on a render of another voice will never make it purely You, or sound 100% like you at least in the current state. perhaps itll get so good it wont be like that. but i dont think its too different from how Utau renders/alters the output of your voice.
    morrysillusion
    i actually really like synth v Natalie, she can handle a lower range really well- much better than solaria and theres some stunning uses of her- but that official demo really.... not good lol. i think it did no favors- she isnt the most flawless, unique bank but she really shines with the right use... she hits that "mature" tone with actually being suitable for a lower range and i like her!
  • Loading…
  • Loading…
  • Loading…
  • Loading…
Top Bottom