AI Hibari Misora -- Arekara（あれから）

For the 2019 Kohaku Utagassen last year, one of the most anticipated segments on NHK's annual New Year's Eve special was the "return" of the legendary kayo kyoku singer Hibari Misora（美空ひばり）, who had passed away in 1989 at the age of 52, through the assistance of computer technology. There was a CG-represented Misora on the Shibuya stage with her voice created from Vocaloid technology and a wealth of voice samples via her huge discography to sing a totally new song, "Arekara" (Since Then).

From what I've read online through YouTube and one article regarding the Kohaku appearance and the earlier debut of AI Hibari Misora last September has been mixed, to say the least. I've seen comments rejoicing this return with many people admitting that they cried a good bucket of tears on hearing the song and seeing her. On the other hand, there were people who criticized the finished rendition of Misora and the song. Even Tatsuro Yamashita（山下達郎）gave his own very blunt opinion, calling the Kohaku performance and the use of the technology in this manner "Blasphemy!".

My take on it? I was only half-impressed with the execution. The CG did look pretty clunky considering all of the hype surrounding the creation (I'm only half-joking when I say that I had been expecting something along the lines of "Star Trek" holographic characters). One YouTube commenter compared it to PlayStation 2-age characters, and sure enough, I also thought that it seemed more like a computer-designed figure created circa 2000 which was when the PS2 came out.

However, the aspect that I found intriguing and far more successful was the creation of Misora's voice. Dyed-in-the-wool fans may take issue with how her voice was emulated in the computer, but I thought it was pretty authentic and nothing sounding like Miku Hatsune（初音ミク）singing like Misora. I realize that this obviously isn't the real person herself, but listening to "Arekara", I could pick up on that familiar voice which enthralled people from the late 1940s to the late 1980s. Strangely, I didn't think that her performance was devoid of soul; that Misora warmth did come out. Listening to it a few times now, yep, perhaps I can pick out certain points where the Vocaloid was peeking through, but picking through it like that is kinda like me spending an hour doing a taste comparison between real butter and "I Can't Believe It's Not Butter!" spread. And I do love both...in moderation!

Ultimately, I think that how the voice came out is the main thrust. The creation of the not-so-great physical Misora character was just there so that the audience didn't have only beloved photos of the singer flashing on the screen as "Arekara" was playing.

Following the Kohaku performance, I started thinking about the benefits and implications on what had been attained with the near-perfect creation of the voice of a long-dead singer. Could that mean now that software exists to pull this off, will there be new songs by a "resurrected" Misora being released in shops or online? If so, who would be making profits? Would the software maker or the songwriter be getting the majority of the money? Would the estate of the original singer get a cut as well? Moreover, what if an unscrupulous entity steals the software with the ability to emulate any singer's voice authentically without permission? Using Misora's voice as an example, would that person end up uploading videos or music files with the kayo kyoku legend singing profanity-laced lyrics in a genre far from enka or even saying things that would end up demeaning Misora? I think that might be the issue that Yamashita was hinting at in his opinion. He certainly doesn't want an AI version of himself gallivanting away after he's gone, but how would other singers feel?

As for "Arekara", the single came out last December. The song is very reminiscent of Misora's swan song hit from 1989, "Kawa no Nagare no you ni"（川の流れのように）, and its delivery sounds as if the lady herself came back to Earth to greet and give thanks to the fans once more. Lyricist Yasushi Akimoto（秋元康）and composer Yoshinori Sato（佐藤嘉風）did a good job at creating this very late-age Misora addition to her discography, although I'm not certain how well "Arekara" did on the charts.

Thrilled? Disturbed? Oblivious? Whatever the reaction, I think this Pandora's Box is now fully open. Will it be possible to hear an AI Yukiko Okada（岡田有希子）or an AI Yutaka Ozaki（尾崎豊）with new material in the next five to ten years?