Thank you for this.

This is the best cover i ever heard

😭😭😭

👏🏻👏🏻👏🏻

Keys are so smooth., a bit confused on the first few lines. at .50

genuinely quality work here

It's a nice sound and performance!

What program do you use?

Thank you so, so much.

Aoi Same(Shark)

is there a 10 hour loop of this yet?

I don’t understand why this makes me cry every time I come back to it

Beautiful vid bro I really liked the Pippa's version of watame's lullaby

kino

Bravo to you good man! All sad girls will be saved.

I would love to get the lyrics

This is a really sweet song. ❤

I raise a glass to the Fishman.

🙏

i really hope the fishman sees this

Here's hoping you make it onto those Sad Girl streams for doing God's Will. Great listen, auto like and subscribe.

what a pleasant surprise to find in my feed this morning

Beautifully done. Jelly good show.

To clarify, because some people in the comment section of my last video were confused (and nobody reads video descriptions, lol): I recorded all of the voices in this video myself and converted them individually to the target voice using the RVC toolkit (github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI ). So essentially, what you hear in this video is *MY* singing with the use of a very sophisticated AI-based voice changer. I used absolutely no audio in this track other than what I recorded myself. This is why the AI sounds relatively "clean" and doesn't have most of the jankiness you usually see in "AI covers" on KZbin - what they usually do is take some official track, extract the vocal part with a tool called UVR and run the AI on that. Since this extraction process is somewhat imperfect, parts of the instrumental still end up in that vocal track and produce artifacts. They also usually have the problem that the extracted vocal still contains effects like reverb, compression and other audio mixing/mastering techniques, which RVC can't handle properly. The same thing does not happen when you use a clean vocal track without any effects as an input, like I did. So my full workflow was (in as much detail as possible): 1. Come up with "funny" lyrics and figure out the chords and melody of the original track (duh) 2. Improvise and record the piano track. 3. Put that piano track into Cubase (a music software that allows multi-layered recordings). 4. Come up with a "mock-arrangement" that uses electronically generated voices to figure out how I want to arrange the individual voices. 5. Sing and record all of the voices in multiple layers; one for each intended target voice, so I had layer 1 = the part sung by Pippa, layer 2 = the part sung by Kiara and so on (the part from 1:14 to 1:45 was especially tricky, took me hours to get the timing and everything just right). 6. Iron out any imperfections in my own singing using the vocal fine-tuning software Melodyne (www.celemony.com/en/melodyne/what-is-melodyne - it is pretty much the industry standard for editing all singing in electronic music, so chances are you've heard hundreds of tracks in your lifetime which used it) -> *I now had the full track with all voices, but still in my own voice.* 7. Train/acquire all the voice models (so the piece of data that enables RVC to simulate a specific voice) I needed for the track - I already had made the model for Pippa for my last video; so I trained the models for Kiara and Kirsche and downloaded the models for the other hololiveEN girls. 8. Feed each recorded voice (still my own voice) individually into RVC (I could do that because Cubase records in layers, so I could individually export each single track and give that to RVC); finding the right settings to make the voices sound as natural as possible -> This produced a bunch of solo vocal tracks in the desired target voices. 9. Swap all the vocal tracks in my Cubase project with the audio files produced by RVC -> I now had a full track with the right voices. 10. Use Melodyne again to fix the slight traces of jankiness that might still be present (optional, made very little difference in this track). 11. Mix and master the individual voices and the whole track, adding reverb, EQ, panning etc.; all the usual audio processing techniques. 12. Export a final mixdown -> This produced the finished track. (13. Assemble a video in Premiere and After Effects, but that's besides the point.) All in all, this whole project was approximately 20 hours of work that I performed over the course of three weeks.

It goes without saying that your playing is great, but the PippAI came out pretty good :o , well done with this!

Here's to the Fishman and all at Sad Girls Inc.

Amazing👏👏

Wait a minute, this isn’t Hatsune Miku!

This song just really makes me feel like everything's going to be okay! 🥰⚰🧟‍♀️🥰

Hii, can I please get the entire lyrics of this? And can I also use this music in our group work, where in we are recquired to perform a lullaby, and this majesty work of art caught my ears.

The song is too comfy :D

I love this… so much…🥲

Also: IVM7 - V7 (or vii7b5) - iii7 - vi7 - ii7 - V7 - I

watame x pippa very fluffy fluffy

alright now we just need Pippa herself to cover it

Thank you for sharing the model, you're a true bro.

What time to be alive, it really does sound like her.

this is amazing

Halfway through the song she started sounding british

So cute

pippipipiip

Need more Pippa music

Had me smiling the whole way through!

this is too fucking cute!

Pretty close, but now we need the real thing!

Oh my god it has the lisp and everything! 😨😨😨😨😨 Also that stablediffusion art is really good 🥇

❤

I need a Pipkin Backup choir