• 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.
lIlI
Reaction score
5,376

Joined

Profile posts Latest activity Postings Media Albums Awarded medals Usernames About

  • in the interest of science, I will now be dividing video essays into categories

    • The Problem with Society Caters to Your Existing Beliefs
    • Making Fun of Teenage Girls Under the Pretence of Moral Concern
    • Here's Why Popular Media is Actually Bad
    • Using TikTok to Prove Feminist Issues are Other Women's Fault, Actually
    • Here's Why All Media You Consumed Age 15 Was Actually Genius
    • Conventionally Handsome Guy Explains Topic He Learnt Yesterday
    and my personal background noise of choice:
    • Video Games are ...HAPPENING
    I enjoy how the lack of timestamps on Tumblr increases the longevity of posts. Every so often getting a few dozen reblogs on a joke from 8 years ago. The out-of-body experience of seeing a post you have no memory of, that at some point in the distant past, you had already clicked like on. Who were you, lili of the past that laughed at this meme...
    This video is great for understanding why English vocal synthesis needed AI to sound 'real': sample based voicebanks weren't built to account for how frequently our pronunciation changes based on context, as well as the huge number of phonetics we slur and drop in different instances. The way we speak has very few consistent rules, and a phonetic alphabet can only scratch the surface of their complexity. It simply isn't fair to programmers trying to create a vocal synth when the way we pronounce words is so often, technically, wrong. The reason AI voicebanks sound more natural is because the program has finally been trained on our mistakes.

    To tune a concatenative voicebank to vocalise the same way as a real person would require a PhD in linguistics!

    An interesting sentiment I've seen a few times from Japanese fans is that the non-Japanese SynthV's are higher quality, and wishes that vocals like Stardust and Solaria were easier to buy (several people hoping Eclipsed Sounds would translate their website to Japanese). As well as encouraging people to make Japanese songs with the non-Japanese voices to advertise their potential.

    Most curiously, it's non-Japanese fans that are most self-conscious about SynthVs having an accent in Japanese; I've only seen complaints about vocalists sounding to American from English-speakers. Likewise, I rarely see English-speakers saying Tokyo6's SynthVs are lower quality.
    Fun online observation: people will always believe someone they see debunking a post, even if the debunk has no credentials and doesn't provide a source. So, theoretically, you could go around the internet writing 'Actually, this isn't true because [insert vaguely convincing bollocks here]' under every piece of information you see, and cause chaos.

    *edit* Just remembered they made an entire website dedicated to people who do this for a hobby. Would ya look at that!
    Review of The Banshee's of Inisherin based on its Wikipedia plot summary: yeah this is about what I imagine is happening any time I see two locals arguing outside the pub
    • Like
    Reactions: WyndReed
    lIlI
    lIlI
    Review of Succession based on its Wikipedia plot summary: bro this is NOT what tumblr made me think this show was about.
    lIlI
    lIlI
    Review of Wenesday: I should have read the Wikipedia plot summary before watching this. The liars and jezebels of twitter tricked me believing it was a quirky opposites-attract romantic comedy. Instead I had to sit through not one, but three YA love triangles with not a wlw in sight. fun mystery 7/10
    Realising that we won't hear Asterian used as often in covers: most vocal synth songs are created with higher registers in mind, meaning they are arranged and mixed so there's a gap in the middle to high frequencies for the voice to sit. Putting Asterian straight into the mix will result in his voice clashing with the instruments currently occupying the lower frequencies, and an empty space left in the upper frequencies. So to most effectively use Asterian, you would need to completely rearrange the instrumental, or apply careful sidechain and eq.

    But far from making him less useful, I think this illustrates how important Asterian is. With a bass singer, you have much more room to compose with instruments in the middle to high ranges, allowing producers to create songs that were previously completely impossible with modern vocal synths.
    morrysillusion
    morrysillusion
    the ranges of typical vocsynth songs has certainly been trouble for me and other male voices to of other kinds too-- esp recently i had a few Yuma covers i wanted to do but the normal pitch was way too high by the chorus, and sadly for him, a bit too low when lowering it an octave-- for asterian is the same kinda deal but kinda more frusterating because he can handle low ranges super well, but like you say its more work to make him fit into it. definitely defines why i couldnt for a bit figure out what was wrong- he can singer super low! but the songs dont sound right bc of those issues out of my control.

    still, its kind of encouraged me to seek out more music to cover. all my covers right now are non-vocsynth songs with singers similar to Asterian. its way more work to wanna cover with him but i hope more people branch out w their ranges in music w a voice like him available for the community. i definitely look forward to people using him in originals (the few i have seen have been great too so!).
    • Like
    Reactions: lIlI
    had the thought 'Azari is the Billie Eilish of Vocaloid' and now I'm trying to match other vocaloPs to pop stars
    Dreamtonics could probably release my dream voicebank and I wouldn't realise because they'd only have one short demo that doesn't show off the voice's unique qualities. Every Dreamtonics voicebank I've bought or considered buying was one I dimissed based on initial impressions, then was impressed by when I heard someone else use them.

    This status was bought to you by: Cong Zheng really can pull off rock!
    morrysillusion
    morrysillusion
    ive been hearing a lot of good cong zheng covers as well-- and like i still proooobably wouldnt be convinced enough to buy her just due to my own uses, but the covers i have been hearing make her sound so much better thatd id still react like 'yeah that sounds like a good voice' instead of 'oh, another voice'
    • Like
    Reactions: lIlI
    Looking back at the old prediction thread really made it hit home how much the vocal synth scene as changed this year for English speakers. With improved AI, cross-lingual, and people outside the VOCALOID niche finally discovering vocal synths as a tool, we may be reaching the cusp of that theoretical Western vocal synth singularity: the moment when our community experiences the same explosive growth that occurred in Japan, all the way back in 2008.

    The world has changed a lot since then, and we're living in a very different digital landscape. I can only hope that this growth, if it comes, will be as positive and artistically empowering as it was for independant musicians on old NicoNicoDouga.
  • Loading…
  • Loading…
  • Loading…
  • Loading…
  • Loading…
  • Loading…
  • Loading…