RVC models choice with AI Hub advice (subject to change;

read their current docs too)

 

If you can separate with these models downloaded from above locally, see also here for the list of all cloud sites and Colabs.

 

“If you need to remove multiple noises, follow this pipeline for the best results:

Remove instrumental -> Remove reverb [probably on vocals] -> Extract main vocals -> Remove noise”

 

Or also Isling’s approach “gives insanely clean results”:

Vocals>De-reverb>Karaoke

 

“That’s how I get completely raw lead vocals if I don’t have the multitracks to a song:

Big Beta 6X (Unwa) -> Karaoke (Frazer & Becruily) -> De-Reverb V2 (Anvuew) -> De-Reverb Room -> (Anvuew) -> Less Aggressive Denoise (Aufr33)” - natethegratevhs

Note: The room model outputs mono natively, you might need to process every channel separately if you don't use MVSEP.

 

“I’m using deux to extract vocals. For lead vocals, I ensemble frazer becruily karaoke and anvuew karaoke [max spec], but I target the backing vocals instead of the lead. Once I get the backing vocals from the ensemble result, I open Audacity and invert the ensemble backing vocal track against the deux vocal track (that’s how I get a cleaner lead vocal, at least for my tracks). After I’ve collected the lead vocals, i run anvuew dereverb mono to remove the reverb” - neoculture

 

Recommended all-vocals models/ensembles for RVC:

- MelBand Roformer | Vocals FV4 (a.k.a. voc_fv4) by Gabox
- Gabox vocfv7beta1 (“seems to give better results than fv4”)

- Mel 2024.10 (mentioned in the MVSEP section of AI Hub, but -)

- BS-Roformer 2025.07 (now has all the metrics better)

- big beta 7 + bs_roformer_mag (max spec) - neoculture (the mag was trained with “pure magnitude spectrum loss (...) for tasks that only need magnitude, like SVC and TTS”)

- unwa beta6/x + voc_fv4 (also good for RVC)

- deux vocal stem (“has more backing vocals than gabox fv7 beta 1-3 and other models I've ever tried, but it may be rather noisy on silent parts or fadeouts” - makidanyee)

- unwa beta 6x + beta5e (avg/avg) - gilliaan

- unwa beta 4 (was better than big beta v5e (NotEddy/codename); research also voc_gabox2)

 

Potentially:
anvuew mag or mag_v2 (metallic noise, Leap xe “eliminates that noise but filters out too many high frequencies above 16kHz” - adrielmz_

 

- gilliaan_bsroformer_deeffect_v1beta | yaml | Colab | MVSEP

(gilliaan) “This is a new "work in progress" model, it's a proof of concept. (finetune of becruily's karaoke model shoutout to him, BS Roformer)

It separates lead vocal from BGVs, it acts as a De-echo/De-reverb/De-chorus/De-double, cleans out the lead. Made specifically for RVC training, it gives you the raw dry lead vocals of a mix basically.

Works on full mixes too, but isolated vocals give better results. If your source is complicated, try splitting L/R first”
As preprocessor “My pick would be one of the Gabox models since they are usually very clean, and there are many fullness levels to choose from." - musicalman

Also, “separate the instrumentals first with a better model since the deeffect model could leave some synths, trumpets, etc.” - makidanyee

E.g. phantom center then gabox flowers v10 with phase fix” - Izzy/bitterkite

 

To experiment more, visit ensembles for general vocal separation

 

Instrumentals

- MelBand Roformer | INSTV7 by Gabox
(unwa instrumental v1e+ OR Mel 2024.10 are also mentioned in their MVSEP section and Gabox Fv7z is mentioned in the x-minus)

 

De-reverb

- MelBand Roformer | De-Reverb by anvuew
(it’s probably v2 variant [also mentioned there], or also Sucial V2 (MelRoformer) mentioned in their MVSEP section [“if I'm unhappy with the results I go for Sucial

- isling”] - it probably follows the model naming scheme of UVR UI on HF, also the new mono-dereverb model is being used occasionally)

 

Backing Vocals

- Mel-Roformer-Karaoke-Aufr33-Viperx (surpassed by Becruily and Frazer Karaoke, but the first can be more consistent; anvuew's Karaoke model have fuller lead vocals; also older Model fuzed gabox & aufr33/viperx (SDR: 9.85) is mentioned in their MVSEP section)

 

De-noise

- Mel-Roformer-Denoise-Aufr33-Aggr (they mention also “Mel denoiser v2” in UVR section)

 

Restoration

- For lossy mp3/mixtures: Apollo Universal by Lew (sometimes AudioSR can be better)

- For voice: AP-BWE or ClearerVoice-Studio's Clear Voice “my favorite is the 2nd one” - codename0)