HiFiPLN v2
Info
Multispeaker Community Vocoder model for DiffSinger
Trained with ~97 hours of varied singing data.
The goal of our vocoder model is to provide more quality possibilities that can be brought out of DiffSinger acoustics. The vocoder can be used with any voice.
We would like to send huge thanks to the Western Diffsinger community for providing datasets for training! Without them, this model wouldn’t be possible.
Code used to train the vocoder is avaible on HiFiPLN Github.
Pretrained checkpoint for finetuning is available at the bottom of this page.
- Model Developed & Trained by Scarfmonster
- Data coordination by PixPrucer
License
Copyright © 2025, Ryszard Goń, Oliwier Ziembla.
https://utau.pl/vocoders/hifipln-v2/
All model weights in the HiFiPLN Community Vocoder, including model weights available to download on this page, are provided under the Attribution-NonCommercial-ShareAlike 4.0 International license.
TERMS OF REDISTRIBUTION:
- Do not sell this vocoder, or charge any fees from redistributing it, as prohibited by the license.
- Include a copy of the CC BY-NC-SA 4.0 license, or a link referring to it.
- Include a copy of this notice, or any other notices informing that this vocoder is provided by Ryszard Goń and Oliwier Ziembla, that this vocoder is licensed under CC BY-NC-SA 4.0, and with a complete acknowledgement list as shown above.
- If you fine-tuned or modified the weights, leave a notice about what has been changed.
- Leave a link to the official release page of the vocoder, and tell users that other versions and future updates of this vocoder can be obtained from the website.
How To Use
- Download your vocoder of choice
- Drag and drop the downloaded .oudep file onto the OpenUtau window.
- Change the
dsconfig.yamlconfiguration file of your chosen voice and setvocoder:to a proper value.
For your convenienve the setting is listed in the table.
Save the file, restart OpenUtau if you had it opened.
Download
By downloading the Vocoders you agree to the Terms of Service.
| Name | Download | Version | dsconfig.yaml | Notes |
|---|---|---|---|---|
| hifipln | 193 MB | 2.0 | vocoder: hifipln_2.0 | sample_rate: 44100, n_mels: 128, hop_length: 512 |
Singing Voice Contributors
Public
AdoVoc Pro,
Cantoria Dataset,
Makam Acapella,
Singing Database,
VocalSet
Community
Ambision,
Beatrix,
Elpatata,
Felipe Souza,
FukaOni,
Galanist,
Gianloop,
HoodyPisDed,
Kisa,
LUNAI,
NebulaMeadow,
Peeslubn,
Rainygardens,
Ryouichi,
SpoopyAce,
Tigermeat,
Tomo,
UniverStars
UtaUtaUtau,
vieri,
Violin,
Vyreed,
Wik,
ariika,
funhouse,
hq_png,
julieraptor,
liure,
nobodyproducer,
postTEENIDOL,
rev,
vahntanabe,
xiel exalt