☰
MSST Bible
☾
INDEX
edit. 17.10.26
Last updates and news
BS/Mel-Roformer UVR beta patch
Full list of model architectures supported by UVR: MDX-Net (a.k.a. v2), MDX23C (archs by kuielab), VR (voice-remover by tsurumeso, versions: 4, 5 [UVR fork], and 5.1), Demucs (by Meta; v. 1-4, only models trained on OG code, not MSST ones), BS-Roformer, Mel-Roformer (arch by Bytedance & implementation from paper by lucidrains; issues on Linux explained later), SCNet, Apollo (in Tools; for upscaling, no DirectML acceleration), BandIt (SFX, no DirectML support). It has also a feature of ensembling models of various archs. If some models don't appear on the ensemble list reach the appropriate section below (or rich the document outline in options). Demudder functionality added in newer UVR patches (currently not on Linux and MacOS).
Current instructions for UVR Roformer patches and installing custom models
Overlap comparisons for Roformers
Most common issues in UVR
Ensembling impossible - model not visible in vocal splitter in UVR
Memory issues - chunk_size table for combatting “RuntimeError”
'use_amp' “Key error”
“”’norm’”” attributeError using e.g. unwa beta 5e model in UVR
Layers errors (general)
TypeError: (...) freqs_per_bands
mlp_expansion_facfor
skip_connection
torch._dynamo.polyfills.fx
[WinError 2] The system cannot find the file specified
num_bands
System error / file not found
Python39.dll error
RuntimeError: ""
Fixed in the beta patch #3 and #4 for all platforms
Older news
The best models
for specific stems
Recent bleedless models
Higher fullness (but with more noise)
Higher bleedless (not so full)
Special purpose models
Ensembles
Models for older archs (outdated models)
for vocals
Tips for separating vocals
Ensembles
RVC models choice with AI Hub advice (subject to change;
Fast inference models for general use
Other services (multi-purpose)
Speech
Can’t find model link?
Models repository
Debleeding/cleaning vocals/instrumentals/inverts
Aligning for bad inverts
How to check whether a model in UVR5 GUI is vocal or instrumental?
Karaoke
MVSEP’s BS Roformer by MVSep Team (SDR: 10.41)
Ensembles
Advanced chain processing chart (image)
Keeping only lead vocals in a song (backing vocals extractor)
Harmonies
For two singers in a duet from one song
Various speakers' isolation (from e.g. podcast or movie)
4-6 stems (drums, bass, others, vocals + opt. guitar, piano):
Chained separation order
Drums
Drumsep (separating parts of drums)
Bass
Strings
Electric guitar
Acoustic guitar
Separating electric and acoustic guitar
Separating two electric guitars
Lead and rhythm guitar
Wind instruments and wind noises
Piano
Synths
Organs
Idiophones
De-crowd
SFX
Any other stem/instrument/sample if not listed above
De-reverb
RX11's dialogue isolate - probably the best for de-echo for now
“Specifically for mono vocal room reverb.” as most vocals are recorded in mono.
Free apps for de-reverb/de-echo/denoise
De-echo
De-noising (vinyl noise/white noise/general)
Plugins (different types of noise)
Bird sounds
De-breath
Mixing/mastering
AI audio upscalers list
Blending with RVC model
(old) More descriptions of models
MDX settings & ens. explanations in UVR5
Ensemble algorithm explanations
4-5 max ensemble models rule
Q: Why I shouldn’t use more than 4-5 models for UVR ensemble (in most cases)
MDX v2 parameters (e.g. HQ_1-5, Kim inst, Inst 1-3, NET, Crowd)
Overlap
>>Tips to enhance separation results<<
SDR leaderboard & explanation
Bleedness and fullness leaderboard
Instrumental models sorted by instrumental bleedless metric:
Vocal models/ensembles sorted by instrumental bleedless metric:
Other ensembles for UVR5
Mateus Contini's methods
Karaoke / Backing Vocals
Table of content
50 models sorted by SDR
Similarity/Phantom Center/Mid channel Extractor
Separating people in recording
UVR5 GUI (MDX, VR, Demucs 2-4 and UVR team models)
GUI FAQ & troubleshooting for UVR (PT. 2)
Chunks may alter separation results
(older) UVR & x-minus.pro updates/news (2021-2023)
Colabs (all current ones)
MSST / MSST-GUI by ZFTurbo | pymms
MVSEP models from UVR5 GUI explained
Manual ensemble
Manual ensemble GUI beta by jarredou
Auto ensemble Colab
Manual ensemble Colab
Joining frequencies from two models
DAW ensemble
Manual ensemble in UVR5 GUI
Model fusion
UVR’s VR architecture models
(settings and recommendations;
VR Colab by HV
VR settings
VR models settings and list
VR ensemble settings
VR Colab troubleshooting
MDX-Net Colab by HV (March 2025)
First vocal models trained by UVR for MDX-Net arch:
(the old) Google Colab by HV
(defunt) Upd. by KoD & DtN & Crusty Crab & jarredou, HV (12.06.23)
Demucs 3
Demucs 4 (+ Colab) (4, 6 stem)
Gsep (now GAudio) (2, 4, 5, 6 stem, karaoke)
dango.ai
music.ai
Iterative Ensemble Colab by IntroC
MDX23 by ZFTurbo (jarredou fork) - 2, 4 stems
What are BigShifts?
KaraFan by Captain FLAM
Ripple/Capcut/SAMI-Bytedance/Volcengine/BS-RoFormer (2-4 stem)
Ripple on Windows or MacOS
Drumsep - single percussion instruments separation
Mel-Roformer MVSEP drumsep models
BS-Roformer MVSEP 5 stem Drumsep model by jarredou
SCNet MVSEP drumsep models
(newer) MDX23C 5 stem drumsep by jarredou
(older) MDX23C 6 stem drumsep by jarredou/Aufr33
BS-Roformer MVSEP 4 stems
Older drumsep by Inagoy
Moises.ai drumsep
FactorSynth
Regroover
UnMixingStation
LarsNet
VirtualDJ 2023/Stems 2.0 (kick, hi-hat)
RipX DeepAudio (-||-) (6 stems [piano, guitar])
Spectralayers 10
[Sample or prompt-based separation]
USS-Bytedance (any; esp. SFX)
Zero Shot (any sample; esp. instruments)
SAM-Audio (prompt-based separation)
AudioSep
Medley Vox (different vocalists)
About other services:
Spleeter
Izotope RX-8/9/10
phonicmind
melody.ml
ByteDance Ripple/CapCut
real-time
Serato
Stems 2.0
Acon Digital Remix
FL Studio
pymss-ara
djay Pro 5.x
Neutone VST
Peel Stems
Fadr.com from SongtoStems.com
Apple Music Sing
Voxless
Ozone 11 Master Rebalance
algoriddim djay
Music to MIDI transcribers/converters
Piano2Notes
Audioshake
Lalal.ai
DeMIX Pro V3
Hit'n'Mix RipX DeepAudio
Moises.ai
___(old) VR settings for ensemble (section deprecated, see the section here)__
(old) ___Good UVR accapella models______
__How to remove artefacts from an inverted acapella?_____
__Sources of FLACs for the best quality for separation process__
Various versions of the same song
Sites and rippers
Dolby Atmos ripping
from Tidal (with orpheus_dl_tidal installed over orpheusDL; max 5.1[.4?])
from Tidal (with https://github.com/exislow/tidal-dl-ng)
360 Reality Audio FAQ
___AI mastering services___
___Best quality on YouTube for your audio uploads____
___Best quality from YouTube and Soundcloud - how to squeeze out the most from the music taken from YT for separation___
__Custom VR models (old)__
__Repository of old Colab notebooks__
__Google Colab troubleshooting (old)_
Error of authorisation during mounting:
Repository of stems/multitracks from music to create your own dataset
SFX
List of cloud services with a lot of space
or for temporary storage
Training models guides
Local SDR testing script
Best ensemble finder for a song script
Universal function to make different types of ensembles by ZFTurbo
Volume compensation for (old) MDX v2 models
UVR-MDX parameters & hashes decoded by Bas Curtiz
UVR model hash decode
Voice Cloning
Stable Audio Open Gen
Anjok’s interview on YT
Demudder
Simpler demudder
UVR Demudders released in the beta Roformer patch
Phase fixer/swapper
Phase-fixed output used in UVR’s Manual ensemble
Pitch shifting/stretching algorithms' comparison
Restoring hi-end in pitched-down tracks - click
TL;DR
Section with miscellanies
Support us
Index
·
Google Doc ↗
MSST Bible
›
___Best quality from YouTube and Soundcloud - how to squeeze out the most from the music taken from YT for separation___
__Google Colab troubleshooting (old)_
Page 203 of 228 ·
Open the original GDoc ↗
← __Repository of old Colab notebooks__
Error of authorisation during mounting: →
MSST Bible