What do you think singthesis will look like in 10 years?

Discussion in 'Multiple / Other Voice Synths' started by partial, Mar 5, 2017.

  1. partial

    partial Guest

    What do you think will be the leading vocal synthesis program? Why? Do you ever think UTAU will surpass Vocaloid in terms of popularity? Do you expect another new, clear method for recording Japanese to catch fire and take off? I'm curious, guys.
     
  2.  
  3. Kiyoteru

    Kiyoteru Local Sensei Supporter Defender of Defoko

    Messages
    3,071
    Likes Received
    3,922
    Trophy Points
    161
    My hopes/predictions in that Vocaloid will remain popular, but so too will Alter/Ego gain users. More music producers will turn to synthesized vocals as another element in their work. UTAU is probably going to stay a niche thing, unless there's suddenly another update that makes it unicode friendly and gives it an API complex enough to handle Arpasing in a presamp-like way. I obviously don't expect to hear vocalsynth on the radio, but it's going to reach beyond the anime fascination and become part of the general internet music consciousness. Like vaporwave?
     
  4. Arissa

    Arissa Ritsu's Renegades Defender of Defoko

    Messages
    646
    Likes Received
    908
    Trophy Points
    117
    I'm not sure. I think there'll be more producers using utau in the future, like Police Picadilly with Yamine Renri. I don't think utau will necessarily outshine Vocaloid, what with marketing and popular figures like the Hatsune Miku and the other cryptonloids. Vocaloid gets constant updates,(V3,V4, etc.) But utau hasn't from Aya/Ame in a while.

    I remember reading a thread about the future of utau or something, and that if the program isn't being updated, in time it'll end up being obsolete or unusable, but idk if that'll happen. Some popular utau might end up like the ones that used to have their long-standing fame, (Momo Momone, etc.) and so I think that utau's still long ways from surpassing vocaloid, giving those various factors.
     
    partial likes this.
  5. AykadUTAU

    AykadUTAU Momo's Minion

    Messages
    5
    Likes Received
    2
    Trophy Points
    4
    I think singing synthesis will improve a lot in 10 years. I think we will have a lot of engines using HMM or GAN like CeVio is currently doing. I also think we will not have to refine anything in the voice in terms of tuning, parameters... Since it will be all automatic. Just put the notes and input lyrics and magic.
    Actually, Techno-Speech, Inc. and the Nagoya Institute of Technology are working in this new technology to be implemented in CeVio. Here is a demo: https://www.techno-speech.com/news-20181214a-en
    The other thing that I think we will see is that many singing synthesizers will be invented ( Mainly Chinese IMO ), and many other languages being supported ( French, German, Italian, Portuguese, Tagalog... ).
    In addition, this is my crazy opinion, but I think if UTAU is not supported in 2026, some voicebanks may turn into VOCALOIDs or Synth V or some other HQ synthesizer. I can definitely see Gahata Meiji, Kohaku Merry, Namine Ritsu, Yokune Ruko... Being in VOCALOID.
     
  6. VocAddict

    VocAddict The Voice Within Us Defender of Defoko

    Messages
    189
    Likes Received
    584
    Trophy Points
    125
    People keep referencing this article but no where has it been mentioned that this new technology would be implemented in CeVIO. This was just a collaborated research project done between Techno-Speech and Nagoya Institute of Technology that was presented at the Acoustical Society of Japan earlier this year.
     
    Last edited: May 19, 2019
    Kiyoteru and YukitoYuki like this.
  7. YukitoYuki

    YukitoYuki Pronouns: They/He Supporter Defender of Defoko

    Messages
    801
    Likes Received
    2,150
    Trophy Points
    131
    My future visions are that unless Yamaha do something, Vocaloid’s popularity is decreasing. Don’t get me wrong but V5 update so far has not done well than previous updates. For now, it’s STILL buggy and unstable even after many updates and almost a year has passed from intial release!

    HMM is outdated (it’s the method that Sinsy and CeVio uses). I think that newer method Deep Learning will be used more and more in future programs and eventually we will have someday a software that completely uses that: Kanru Hua’s Synth V already uses that as a part of synthesis method!

    UTAU might get an update in next 10 years as UTAU’s developer has become active this year on Twitter twiiting new things that they’re experimenting for possible update. If not, I’m sure that they’ll be a close alternative available in future and don’t forget that we’ll have already very stable - both free and paid - virtual mashines where wej can run old Win and Mac OS in a “virtual bubble”. So even if UTAU’s “””replacement””” is not coming in the next decade, we’ll have still ways to run current version of it and I think that many older users will.
     

Share This Page