Ado vs Vocaloid — Villain / Buriki no Dance / Senbonzakura (Cover vs Original Study)
What an amazing study that was. I learned so much, but I feel still like I haven't scratched the surface on all of this yet!
Comments (15)
So much time wasted on that first awful song.
It's so funnny but for years I didn't know bukiri no dance was a vocaloid song. I forget who, but I played a utaite's version as an osu map, and it took a while to figure out that it wasn't the original. Edit: Looked it up and it was 96neko.
On Japanese: This language is very difficult for English native speakers, and it's not just reading (5000-ish kanji and two 46 character alphabets*), the grammar and entire manner of speaking is very unintuitive to most native English speakers. There's a very real stereotype about weebs going to college, doing Japanese 101 and dropping out very fast. Also people often cite the U.S. Foreign Service Institute, where Japanese is not only listed as a Category V (the hardest category or languages) but is also considered *the* hardest language within the Category V languages. None of that necessarily to discourage you (Japanese music is also the main reason I started learning Japanese), but to know what you'd be getting into. *They're actually syllabaries, which means one symbol covers a syllable** **OK actually they're shorter than syllables, not important With that said, Japanese phonetics are VERY easy so if you just want to make sure your pronunciation is on point it isn't difficult to learn. Japanese vowels never change: a "ah" like fAther, i "ee" like "sEE", u "oo" like "cOOl", e "eh" like "Etch", o "oh" like "Over". Consonants are about how you would normally expect from their romanized spellings, with the small exception that "fu" isn't correct (No lip movement: Basically, pretend you're blowing out a candle and use your vocal cords) and Rs are actually in between an L and an R (if you know how to roll your Rs, basically try to do that but only make a single sound instead of a full trill). Standalone Ns (like seNboNzakura) also have weird rules, but they're not really important to get good enough pronunciation. The only important one is that before a p or b it's more like an M. (Sembonzakura is a more native pronunciation), which is why many Japanese words will be spelled in English with *either* an M or N in some situations.
For me personally, the original vs cover thing tends to depend on the song, but a good bit of the time I think my favorite versions tend to be covers made with other voicebanks. (Might just be me being a fan of more obscure banks, but I feel like some songs are made with Miku because she's the most popular and not because she's the best bank for the song.) It's not super common, but there are Vocaloid songs where I feel the artificial vocals give the song more of an edge and make it more interesting to listen to than human covers. For an example, my favorite cover of "Ghost Rule" by DECO*27 is a cover by akem using an UTAU bank called Kohaku Merry. Her vocals might be super rough, but it makes me feel more emotional than most of the utaite covers I've heard where they tend to be softer and quieter. (I think an Ado version could be really good, but as far as I know she's never sung the song publicly.) There are some songs where I really prefer utaite covers though, especially "Villain" (no offense to Flower but it took me a long time to warm up to her voice and it was largely because of liking Teniwoha's songs that I did). One of my favorite utaite, Kyounosuke, released a cover of it when he came out as a trans man, and it's probably tied with Teniwoha's self-cover as my favorite version.
I like Vocaloid songs, but I do prefer them being performed by real people, and as a huge Ado fan, I love her versions of the songs.
Really fun video Tim. Thanks! If I could ask, what is meant when you refer to “ethically trained voice bank”. I do understand the idea of a voice bank(on a very basic level) but what is ethically trained? Just really curious, you mentioned it a couple of times and I’d like to understand, vs unethically. If you see this and respond, appreciate it, if not, thanks for the fun video.
He is talking about the use of AI in art. The voice bank itself is a human recorded bank of vocals, but the modern software does use AI to synthesise some of the sounds from the voice bank to give more complex and detailed options to producers. AI use in creative media can often be unethical as they “steal” art online without consent from the creator/owner and use it to train the models for generative AI. Thus most ai generated art, be it drawn, video, music, etc, is unethical. However here the use of AI in the more recent voicebank software is done in a much more ethical manner.
So, speaking on covers, for me personally it varies song to song. This makes me think of giga and reol a lot. For the song hibikaze, Reol's cover is definitely great. But the song just *fits* miku's voice so well that I really only ever listen to the Miku version. It's tuned so well, the synths work so well with her voice, she sits in the mix perfectly. You can tell the song was made for miku. This is my preference for most songs, though the preference isn't always as strong as this. I prefer the Miku version of Godish to Ado's cover for example, but don't feel as strongly about it as hibikaze. On the other hand, you have Drop Pop Candy on the same album as Hibikaze, which was made with Rin and Luka. It's a great song. But the lyrics feel clearer and more heartfelt with Reol. Reol's singing somehow carries the beat more, and her ornamentation seems to add a lot to the song that feels missing from Rin and Luka's version. There's other songs where I feel a human's vocals really elevated the song. Iya Iya Iya, Traveling mood. Some songs just really want a soft heartfelt voice I guess, like traveling mood. Where others like Iya Iya Iya, I feel like the original tuning wasn't quite meeting the production of the song where it was at. For a third path, a good example is Birdbrain by Jamie Paige. In the original she uses Kasane Teto. It's such an amazing song, I love it. Second favorite Jamie Paige song. She manages to have a unique voice and tuning style with teto, and bring so much emotion into the song. However, I'm also very fond of Nerissa Ravencroft's cover. It has a playfulness, and a different tone and musicality to it. They both bring something great to the song, and I find myself listening to both pretty often. This is also true of 'No Title' the namesake of the same album as Hibikaze and Drop Pop Candy. It's also how I feel about a lot of covers in different genres from the original, such as jazz covers. Also, english trans-lyric covers, now that I think of it. Such as Will Stetson, and the aforementioned Nerissa Ravencroft. I find I tend to lean towards vocaloid more often. That might be because i'm a hipster, Vocaloid has been my musical identity since early highschool, or that I hear the vocaloid version first usually. I'm not sure. But regardless, the vocaloid version usually remains more inspiring to me. It feels like if I tried hard enough, I could make that song. Whereas I could never sound like ado (and have no ambition to). There's also the fact that the song probably wouldn't exist without the vocaloid.
For me the original vocaloid version and the cover version is like 2 different songs in my head. It's the same for me when a band makes a cover of a popular song, I might love both versions but it's 2 different songs in my head even though the origin of them is the same.
A fun aspect we overlooked was that Ado's Senbonsakura cover was rearranged by Teniwoha! A bit of a shame you didn't get a third party lyric vid for the Villain cover. The one on the MV is a bit overly literal.
As Chris has pointed out the preference comes largely down to what people hear first. For me personally I tend to enjoy both the cover and vocaloid song. For each vocaloid song i love to find covers. Once I have a prefered cover I tend to come back to that cover even if I enjoyed other covers. I also sometimes crave the vocaloid sounds so I will listen to the vocaloid version
When you talk about the preference of original vocaloid song VS covers it often comes down to which version people heard first. Some people will prefer the vocaloid song to the cover automatically because of that or the other way around. I personally like the arrangement of the original song of Senbonzakura more than the arrangement of Ado's cover, though the cover adds it's very personal tone to it. The Ado cover is absolutely fantastic and I like it, but even then the cover wouldn't probably even crack the top 3 of Senbonzakura covers I heard in my life since it released like a decade later than most. A great cover imo is something that not only connects to the original song and conveys the feelings of the producer that made it, but it also adds on top emotional connectivity to the singer as well and those personal notes/layers that are deliberately done, which then can elevate songs to completely new heights. I can always listen to any version I would say, but yeah this varies for people, it also comes down to how the vocaloid was tuned and all and other little details in the instrumentation of the producer. Personally my favorite cover of Senbonzakura is probably the one from Wagakki Band, it's legendary there is a reason why it is the most viewed version of the song at 176 million views. The arrangement is so cool and unique.
I agree about Wagakki bands version of Senbonzakura. It is also affected by the fact that I heard that version first.
There is a very interesting documentary on YT regarding Senbonzakura in relation to Fukushima. Wagakki band is very interesting musically.
Can’t wait to watch this video. 3 banger songs. I love all of them
