Google Doc

edit. 28.08.26

Sat, 29 Aug 2026 20:20:11 GMT

edit. 28.08.26 deton24’s Instrumental and vocal & stems separation & mastering (UVR 5 GUI: VR/MDX-Net/MDX23C/Demucs 1-4, and BS/Mel-Roformer in beta / MSST/pymms MVSEP-MDX23-Colab/Drumsep/SCNet/Apollo/MedleyVox x-minus.pro (uvronline.app | nextgen)/mvsep.com/Colabs Gaudio/Dango.ai/Audioshake/Iterati

Last updates and news

Sat, 29 Aug 2026 20:20:11 GMT

Last updates and news - (uvronlone) “Added BS-RoFormer Mag v2 vocal model: https://nextgen.uvronline.app/#model=bs_roformer_mag_v2_anvuew” - Aufr33 - New best open ensemble for instrumental was found by nextgen.uvronline.app user, and approved by dca: Mel Inst_GaboxFv9 + BS SW (Max Spec) Can be used

- BS/Mel-Roformer UVR beta patch For GPU acceleration, UVR currently supports: a) CUDA (NVIDIA GPUs) b) DirectML (AMD and Intel GPUs; previously misnamed as OpenCL) c) MPS (Mac M1 [ARM]/x86-64) And even old CPUs are supported (AMD A6-9225 Dual-Core or Intel Core 2 Quad [models with SSE4.1 tested]), but for at least MDX-Net HQ (v2) models speeds will be good on such CPUs without GPU acceleration. DirectML acceleration is not supported for Apollo, Bandit (also incompatible with MPS), SCNet and probably Demucs 2 archs - CPU will be used automatically. Minimum reasonably good enough NVIDIA GPU for Roformers might be: desktop RTX 3050 6GB, 2304 CUDA cores (at best 8GB variant with more 2560 CUDA). E.g. Colab’s Tesla T4 11GB (older RTX 2000 gen) has 2560 (which without TTA [not implemented for Roformers in UVR] is just alright - see separation times). The old 980 Ti 6GB with 2816 CUDA will be rather slower than these due to older architecture. I'd refrain from getting a mobile RTX 3050 - the Ti variant has 2560 CUDA cores, but both Ti and regular have 4GB, and using shared memory is slower - Roformers generally use more than that. In FP16 inference, if we “convert” CUDA cores performance to match the Blackwell: RTX 5060 Ti 4608 | 4608 | 100% RTX 4070 Ti 7680 | ~4800-5120 | ~105-110% RTX 3080 8704 | ~3200-3500 | ~70-75% Plus, you can always convert current Roformers to FP16 if some model uses FP32 (more below). Performance comparisons: 1 | 2 | 3 | 4 Using CUDA, even with currently the biggest 2GB size inst model, Rifforge, doesn’t reach 6GB VRAM usage on RTX 2000 (but its separation time on T4 using TTA in Colab takes a long 36 minutes for a 5 minute file), so consider min. 6GB VRAM as bare minimum for reasonable separation times with around 2560 CUDA cores on at least RTX 2000 (newer gens should have faster cores in inferencing). Unless you want to use the 53 stem model. For AMD/Intel GPUs, if you want to use default high chunk_size with certain models, 16GB VRAM is recommended, so you won’t have to decrease it manually (DirectML is way less memory-efficient than CUDA). Minimum AMD GPU capable for Roformers using DirectML might be RX 6600 XT or RX 7600 (might be even faster than 3050, DML will probably use its new RDNA3 AI cores).

Sat, 29 Aug 2026 20:20:11 GMT

- BS/Mel-Roformer UVR beta patch For GPU acceleration, UVR currently supports: a) CUDA (NVIDIA GPUs) b) DirectML (AMD and Intel GPUs; previously misnamed as OpenCL) c) MPS (Mac M1 [ARM]/x86-64) And even old CPUs are supported (AMD A6-9225 Dual-Core or Intel Core 2 Quad [models with SSE4.1 tested]),

DirectML is slow nevertheless, so if you want some serious GPU acceleration boost on AMD GPUs, use MSST/pymms with ROCm instead. DirectML gives tremendous overhead over CUDA. For older GPUs than AMD Vega (which doesn't support ROCm PyTorch or only potentially with ZLUDA with modded libs) or Intel GPUs with Vulkan support, instead of MSST use: https://github.com/chenmozhijin/BSRoformer.cpp (CLI, C++ ready build, model conversion required besides Deux and FV6, allows quantization for faster separation, Vulkan backend faster than ROCm 5+ZLUDA on Windows) or also: https://github.com/pymss-project/pymss-mnn (CLI, requires building and also model conversion, Vulkan backend too, potentially a bit slower)

Sat, 29 Aug 2026 20:20:11 GMT

DirectML is slow nevertheless, so if you want some serious GPU acceleration boost on AMD GPUs, use MSST/pymms with ROCm instead. DirectML gives tremendous overhead over CUDA. For older GPUs than AMD Vega (which doesn't support ROCm PyTorch or only potentially with ZLUDA with modded libs) or Intel GP

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).

Sat, 29 Aug 2026 20:20:11 GMT

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 & implementati

Models for Roformers/SCNet/Bandit arch are located altogether in the MDX-Net menu.

Sat, 29 Aug 2026 20:20:11 GMT

Models for Roformers/SCNet/Bandit arch are located altogether in the MDX-Net menu. Apollo audio upscaler is located in Tools. Don’t forget to enable GPU Conversion - if it works, it speeds up separation hugely. Even Tiger Lake iGPUs are capable of working with at least MDX-Net HQ (v2) models. - Min.

Ensembling impossible - model not visible in vocal splitter in UVR

Sat, 29 Aug 2026 20:20:11 GMT

Ensembling impossible - model not visible in vocal splitter in UVR

- If user-imported Roformers aren't recognized in "instrumental/vocals" in ensemble or in vocal splitter, but are in "multi-stem ensemble": “The .yaml associated with the model usually needs to be updated to match UVR's stem naming conventions. For example, if your config shows the instruments as "other" and "vocals", it will need to be updated to "Instrumental" and "Vocals" (case-sensitive)” - Anjok

Sat, 29 Aug 2026 20:20:11 GMT

- If user-imported Roformers aren't recognized in "instrumental/vocals" in ensemble or in vocal splitter, but are in "multi-stem ensemble": “The .yaml associated with the model usually needs to be updated to match UVR's stem naming conventions. For example, if your config shows the instruments as "o

Memory issues - chunk_size table for combatting “RuntimeError”

Sat, 29 Aug 2026 20:20:11 GMT

Memory issues - chunk_size table for combatting “RuntimeError” dim_t (old patches) to chunk_size conversion (new patches) for Roformers with hop_length = 441 in the model’s yaml (patch #1-2 and Linux uses dim_t instead of chunk_size - more) Useful if you see insufficient memory error, so you can dec

'use_amp' “Key error”

Sat, 29 Aug 2026 20:20:11 GMT

'use_amp' “Key error” using the GH repo above, and referencing (separating) some models: - “add: use_amp: true in the training part of [models’ yaml] config file (it's missing)”

“”’norm’”” attributeError using e.g. unwa beta 5e model in UVR

Sat, 29 Aug 2026 20:20:11 GMT

“”’norm’”” attributeError using e.g. unwa beta 5e model in UVR - a) Ensure you installed UVR Roformer patch (5.6.1), and you're not using the old 5.6 version (but 5.6.1 is reported once you open the app) b) You could pick wrong model architecture in Install model option (so not Roformer, and general

Layers errors - general

Sat, 29 Aug 2026 20:20:11 GMT

Layers errors - general a) You didn’t install the newest patch and still use e.g. beta 2 with some newer model b) You could check Roformer v2 instead of v1 during installing of the custom model. c) Model trainer didn't clean the weight using this Python script (you can do it by yourself) Usage of th

TypeError: (...) freqs_per_bands

Sat, 29 Aug 2026 20:20:11 GMT

TypeError: (...) freqs_per_bands You probably set Mel-Roformer model type instead of BS-Roformer when it was necessary. Go to MDX-Net, pick the model>Edit model config>Choose parameters. There you should change Model type.

mlp_expansion_facfor

Sat, 29 Aug 2026 20:20:11 GMT

mlp_expansion_facfor You probably use outdated codebase for Linux or old UVR version incompatible with some newer Roformers like the SW. We have some workarounds for it, but at least below not confirmed to work so far: "The first issue is that the BS Roformer implementation in UVR5 does not accept t

"The size of tensor a (484864) must match the size of tensor b (485100) at non-singleton dimension 2"

Sat, 29 Aug 2026 20:20:11 GMT

"The size of tensor a (484864) must match the size of tensor b (485100) at non-singleton dimension 2" Regarding the second issue, chunk_size should be a multiple of stft_hop_length." - anvuew Q: Hop is 512 and chunk is 588800. That's a ratio of 1150. A: Then you need to find out where this 485100 co

skip_connection

Sat, 29 Aug 2026 20:20:11 GMT

skip_connection Remove that entire line from the yaml too if you started seeing this error after the above. But it might still not help.

torch._dynamo.polyfills.fx

Sat, 29 Aug 2026 20:20:11 GMT

torch._dynamo.polyfills.fx Some of the newer Roformer models fail with the UVR RTX 5000 patch giving that error, while the non-RTX 5000 patch works as usual. You need to go to model’s yaml and edit the line: use_torch_checkpoint: True to False, or delete it. To find out which yaml is associated with

[WinError 2] The system cannot find the file specified

Sat, 29 Aug 2026 20:20:11 GMT

[WinError 2] The system cannot find the file specified - Sometimes 32/64 bit float output set can trigger it - Or you can also reencode your input file and name it intput.wav, and choose wav mode. - Setting FLAC output might also work (it seems to happen with mp3 input and output). (can’t remember i

num_bands

Sat, 29 Aug 2026 20:20:11 GMT

num_bands “Most of these models are BS[band-split]/Mel-Band Roformer. [It] can be distinguished by their config, bs uses `freqs_per_bands`, Mel-band uses `num_bands`” So you most likely installed this model with wrong model type chosen during installation of the model. Go to Edit model config and se

System error / file not found

Sat, 29 Aug 2026 20:20:11 GMT

System error / file not found > Use WAV output quality - UVR will only process files with English characters - some complicated names/paths give “System error” during separation (can’t remember if it still exists in the latest patches - check if your UVR is up to date as well) E.g. for RuntimeError:

Python39.dll

Sat, 29 Aug 2026 20:20:11 GMT

Python39.dll Probably it happened after installation of some patch on (too) dirty UVR installation (probably already patched before or some older than 22_30. You must reinstall UVR using only the latest required patches. More troubleshooting

RuntimeError: ""

Sat, 29 Aug 2026 20:20:11 GMT

RuntimeError: "" Traceback Error: " If you have these two lines without any text at the beginning of the error log as above using AMD GPU on every attempt of separation with GPU Conversion option turned on in UVR, for all archs and models - you probably use outdated GPU drivers and/or Windows not co

- Stable 5.6 OpenCL (DirectML) version of UVR 5 GUI for Windows

Sat, 29 Aug 2026 20:20:11 GMT

- Stable 5.6 OpenCL (DirectML) version of UVR 5 GUI for Windows

Supporting AMD and Intel GPUs acceleration but no Roformers yet

Sat, 29 Aug 2026 20:20:11 GMT

Supporting AMD and Intel GPUs acceleration but no Roformers yet https://github.com/Anjok07/ultimatevocalremovergui/releases/download/v5.6/UVR_v5.6.0_setup_directml_old.exe Mac: https://github.com/Anjok07/ultimatevocalremovergui/releases/ (newer beta Roformer [with “roformer” in the installer name] s

- (no longer necessary) Fork of UVR GUI and How to install - support for AMD and Intel GPUs appeared (works only for VR and MDX architectures), Besides W11, also W10 confirmed working, MDX achieves speeds of i5-4460s using 6700 XT, while for VR, speeds are v. fast and comparable to CUDA, so CPU processing might be slower in VR, but for MDX you might want to stick with the official UVR5 GUI.

Sat, 29 Aug 2026 20:20:11 GMT

- (no longer necessary) Fork of UVR GUI and How to install - support for AMD and Intel GPUs appeared (works only for VR and MDX architectures), Besides W11, also W10 confirmed working, MDX achieves speeds of i5-4460s using 6700 XT, while for VR, speeds are v. fast and comparable to CUDA, so CPU proc

The best models

Sat, 29 Aug 2026 20:20:11 GMT

The best models

for specific stems

Sat, 29 Aug 2026 20:20:11 GMT

for specific stems There's no such thing like the best model. It depends on a song, even in specific genre, mixing, effects, etc. You need to test the best models posted at the top here, and see what fits the best for your song or cut it into pieces and/or check ensembles. For constant vocal buzzing

- Advanced chain processing chart (image)

Sat, 29 Aug 2026 20:20:11 GMT

- Advanced chain processing chart (image) It’s a method utilizing old models, and e.g. Kim Vocals 2 can be potentially replaced by unwa’s BS/Mel-Roformer models in beta UVR (or other good method for vocals) or ensembles mentioned in this document. Check the best current methods for vocals in one ste

Chained separation order

Sat, 29 Aug 2026 20:20:11 GMT

Chained separation order With single stem models below, feel free to experiment with different orders of sequential stem separation to enhance separation results: #1 1) Well sounding result from instrumental model/s first 2) drums or bass 3) piano or guitar 4) strings or horns Note: If your song has

Plugins (different types of noise)

Sat, 29 Aug 2026 20:20:11 GMT

Plugins (different types of noise) Free - Guide for classic denoiser tools in DAW, e.g. for debleeding (Bas Curtiz): Tips & Tricks - Bertom Denoiser Classic (or paid Pro) - Accusonus ERA 6 (released for free after FB acquisition) - bundle with also de-esser, voice auto-EQ, voice leveller (better tha

“UVR BVE v2 model [currently on x-minus] is actually full band. There is, however, a small nuance. This model uses MDX VocFT preprocessing, which is not full band. MDX VocFT model is rebalancing the song. The music is slightly mixed with the vocals (25% music + 100% vocals). This mix is then processed by the BVE model. A small amount of music can help the model better understand the context (it's important for harmony separation). We train the model on a rebalanced dataset. It contains 25% of music.” aufr33

Sat, 29 Aug 2026 20:20:11 GMT

“UVR BVE v2 model [currently on x-minus] is actually full band. There is, however, a small nuance. This model uses MDX VocFT preprocessing, which is not full band. MDX VocFT model is rebalancing the song. The music is slightly mixed with the vocals (25% music + 100% vocals). This mix is then process

MDX settings & ens. explanations in UVR5 (and also Demucs/VR/MDX v2/23C inferencing parameters)

Sat, 29 Aug 2026 20:20:11 GMT

MDX settings & ens. explanations in UVR5 (and also Demucs/VR/MDX v2/23C inferencing parameters) In one of the pre-5.6 UVR updates, the following min/avg/max features for single models got replaced by a better automated alternative, and you might still get cleaner results of e.g. voc_ft with max_mag

>>Tips to enhance separation results<<

Sat, 29 Aug 2026 20:20:11 GMT

>>Tips to enhance separation results<< If you cannot achieve good separation, you can conduct the following experiments. Some tricks might be outdated when using Roformers. 1. De-bass Turn down all the bass to stabilize the voice frequencies of your input song (example EQ curves: 1 and 2). Male sett

24. arigato78 method for lead vocal acapella

Sat, 29 Aug 2026 20:20:11 GMT

24. arigato78 method for lead vocal acapella 1) Try to make the best acapella (using mvsep.com site or using UVR GUI). I recommend the MDXB Voc FT model for this with an overlap setting set to at least 0.80 (I used 0.95 for this example). The overlap for this model at mvsep.com is set to 0.80. Speak

SDR leaderboard & explanation

Sat, 29 Aug 2026 20:20:11 GMT

SDR leaderboard & explanation Tested on multisong dataset https://mvsep.com/quality_checker/leaderboard2.php?&sort=instrum (some models/AIs/methods are not public, or only on MVSEP, all others you will find in UVR's and/or download center if you can't find some models, some only after using VIP code

Bleedness and fullness leaderboard Python evaluation script by jarredou (prob. mirror), Torch version (with Bas Curtiz), used on Quality Checker

Sat, 29 Aug 2026 20:20:11 GMT

Bleedness and fullness leaderboard Python evaluation script by jarredou (prob. mirror), Torch version (with Bas Curtiz), used on Quality Checker Librosa version added to ZFTurbo training repo. More detailed and reliable method of evaluation on multisong dataset than the popular SDR metric. (old) Bas

50 models sorted by SDR

Sat, 29 Aug 2026 20:20:11 GMT

50 models sorted by SDR (from the public ones - so only those available to download and offline use) (last updated 07.10.2024) These are basically the top single models for now (conventionally after these, additional vocal residues kick in, especially if not a vocal model) Based on Multisong dataset

UVR5 GUI (MDX, VR, Demucs 2-4 and UVR team models)

Sat, 29 Aug 2026 20:20:11 GMT

UVR5 GUI (MDX, VR, Demucs 2-4 and UVR team models) GUI provides more functionalities and models/AIs compared to Colabs, incl. custom model import: https://github.com/Anjok07/ultimatevocalremovergui/releases Official app Win 11 installation tutorial: https://youtu.be/u8faZW7mzYs MacOS build: https://

GUI FAQ & troubleshooting for UVR (PT. 2)

Sat, 29 Aug 2026 20:20:11 GMT

GUI FAQ & troubleshooting for UVR (PT. 2) Start with reading information about Roformer patch and its common issues section - “If you enable the "enable help hints" setting” “you can hover parameters with the mouse, [and] you'll get [settings] info hints (...) if [it’s] not activated by default)” -

- Chunks may alter separation results

Sat, 29 Aug 2026 20:20:11 GMT

- Chunks may alter separation results (update: chunks are now replaced with batch mode on even 4GB cards, feature was introduced in one of beta patches and is available in v. 5.6, and you cannot use chunks experimentally in this version if batch mode gives you some vocal pop-ups vs 11GB GPUs which i

Manual ensemble

Sat, 29 Aug 2026 20:20:11 GMT

Manual ensemble - Make sure you use lossless files of the same sample rate (or when lossy, potentially from the same format and source), otherwise the files will begin in different places, e.g. if one is lossy or in different lossy format, and then the process will require manual alignment in DAW be

UVR’s VR architecture models

Sat, 29 Aug 2026 20:20:11 GMT

UVR’s VR architecture models (settings and recommendations; mostly outdated arch for all vocals and instrumental models) Available on Colab, HF, UVR, UVR old CLI, MVSEP VR Colab by HV (old) https://colab.research.google.com/github/NaJeongMo/Colaboratory-Notebook-for-Ultimate-Vocal-Remover/blob/main/

First vocal models trained by UVR for MDX-Net arch:

Sat, 29 Aug 2026 20:20:11 GMT

First vocal models trained by UVR for MDX-Net arch:

9.703 model is UVR-MDX-NET 1, UVR-MDX-NET 2 is UVR_MDXNET_2_9682, NET 3 is 9662, all trained at 14.7kHz

Sat, 29 Aug 2026 20:20:11 GMT

9.703 model is UVR-MDX-NET 1, UVR-MDX-NET 2 is UVR_MDXNET_2_9682, NET 3 is 9662, all trained at 14.7kHz (instrumental based on processed phase inversion) List of all (newer) available MDX models at the very top. I think main was 438 in UVR 5 GUI at some point. At least now it's simply main_438 (if i

Demucs 3

Sat, 29 Aug 2026 20:20:11 GMT

Demucs 3 for 4 stems (SDR 7.7 for 4 stems, it’s better than Spleeter (which is SDR 6.5-7), or better than MDX 4 stem. In most cases, it’s even better than Audioshake - at least on tracks without leading guitar) Accompanied by MDX-UVR 9.7 vocal model, it gives very good 4 stem separation results (For

Demucs 4 (+ Colab) (4, 6 stem)

Sat, 29 Aug 2026 20:20:11 GMT

Demucs 4 (+ Colab) (4, 6 stem) 4 stem, SDR 9 for vocals on MUSDB HQ test, and SDR 9 for mixdowned instrumentals (5, 6 stem - experimental piano [bad] and guitar) https://github.com/facebookresearch/demucs (all these models available in UVR 5 GUI or MVSEP [just x-minus doesn’t have ft model for at le

Gsep (now GAudio) (2, 4, 5, 6 stem, karaoke)

Sat, 29 Aug 2026 20:20:11 GMT

Gsep (now GAudio) (2, 4, 5, 6 stem, karaoke) https://studio.gaudiolab.io/ Paid (20 minutes free in mp3 - no credit card required) 7$/60 minutes 16$/240 minutes 50$/1200 minutes Electric guitar (occasionally bad), good piano, output: mp3 320kbps (20kHz cutoff), wav only for paid users, accepted input

dango.ai

Sat, 29 Aug 2026 20:20:11 GMT

dango.ai (2 or more [up to 6+] stems, paid only, 30 seconds free preview of mp3 320 output, 20kHz cutoff) drums, vocal, bass guitar, electric guitar, acoustic guitar, violin, erhu “10 tracks = €6.33 + needs Alipay or WeChat Pay” max 12 minutes input files allowed Now the site has English interface C

music.ai

Sat, 29 Aug 2026 20:20:11 GMT

music.ai Paid - $25 per month or pay as you go (pricing chart). In fact, no free trial. Good selection of models and interesting module stacking feature. To upload files instead of using URLs “you make the workflow, and you start a job from the main page using that custom workflow” [~ D I O ~]. Alle

Iterative Ensemble Colab by IntroC

Sat, 29 Aug 2026 20:20:11 GMT

Iterative Ensemble Colab by IntroC (for instrumentals) https://colab.research.google.com/github/Qupci/Iterative-Ensemble/blob/main/Iterative_Ensemble.ipynb By default it uses models from the end of 2025 (V1e, Resurrection, Revive3e, SCNet), but it's a subject for further development. It performs man

MDX23 by ZFTurbo (jarredou fork) - 2, 4 stems

Sat, 29 Aug 2026 20:20:11 GMT

MDX23 by ZFTurbo (jarredou fork) - 2, 4 stems (2-4 stems, max 32-bit float output) As of October 2025, Colabs are defunct due to Google’s runtime changes (possible fix - make the dependencies look like here). v2.5 (broken mirror to change repo), v2.5 /w HQ_5 (experimental - muddiness, residues - set

KaraFan by Captain FLAM

Sat, 29 Aug 2026 20:20:11 GMT

KaraFan by Captain FLAM (2 stems) It was made before Roformers models (potentially deprecated) Colab w/ more models (AI Hub fork, also fixed), fixed org. Colab, org. Colab (slow), GUI, GH documentation How to install it locally (advanced), alt. tutorial, or easy instruction Should work on Mac with S

Ripple/Capcut/SAMI-Bytedance/Volcengine/BS-RoFormer (2-4 stem)

Sat, 29 Aug 2026 20:20:11 GMT

Ripple/Capcut/SAMI-Bytedance/Volcengine/BS-RoFormer (2-4 stem) (Ripple is discontinued since 31 January 2026) Output quality in Ripple is: 256kbps M4A (320kbps max) and lossless (introduced later). 50MB upload limit, 4 stems Min. iOS version: 14.1 Ripple is only for US region (which you can change,

>Drumsep - single percussion instruments separation

Sat, 29 Aug 2026 20:20:11 GMT

>Drumsep - single percussion instruments separation If you want to further separate single instruments from drums stem separated with e.g. MDX23 Colab, Mel-Roformer drums on x-minus.pro premium, MVSEP, or Demucs_ft (not necessarily BS-Roformer SW) into: hihat, cymbals, kick, snare and more, you migh

USS-Bytedance (any; esp. SFX)

Sat, 29 Aug 2026 20:20:11 GMT

USS-Bytedance (any; esp. SFX) https://github.com/bytedance/uss (COMMAND: "conda install -c intel icc_rt" SOLVES the LLVM ERROR) You provide e.g. a sample of any instrument or SFX, and the AI separates it solo from a song or movie fragment you choose to separate. It works in mono. You need to process

Zero Shot (any sample; esp. instruments)

Sat, 29 Aug 2026 20:20:11 GMT

Zero Shot (any sample; esp. instruments) https://github.com/RetroCirce/Zero_Shot_Audio_Source_Separation (as USS Bytedance came out now, zero shot can be regarded as obsolete now, although zero-shot might is rather better for single instruments than for SFX) You provide e.g. sample of any trumpet or

SAM-Audio

Sat, 29 Aug 2026 20:20:11 GMT

SAM-Audio https://github.com/0x0funky/audioghost-ai More VRAM/performance optimized fork of SAM-Audio by Meta: https://github.com/facebookresearch/sam-audio The large model metrics. "Fish Audio has it on their website but credits are diabolocal" It’s a prompt-based separator. It can do a decent job

AudioSep

Sat, 29 Aug 2026 20:20:11 GMT

AudioSep “I decided to try AudioSep: https://github.com/Audio-AGI/AudioSep on MultiSong Dataset. I used prompt 'vocals'. I was sure it would be bad, but I didn't think it's so bad. https://mvsep.com/quality_checker/entry/8408 I also tried it on the Guitar dataset - it's even worse - negative SDR. Ma

Sources of FLACs for the best quality for separation process

Sat, 29 Aug 2026 20:20:11 GMT

__Sources of FLACs for the best quality for separation process__ Introduction Don’t use YouTube or mp3 as input files for separation. Compression decreases the quality of the output. If you need music video, Tidal has videos up to 320 kbps AAC audio, with bitrates varying by content and quality sett

_AI mastering services_

Sat, 29 Aug 2026 20:20:11 GMT

___AI mastering services___ Might be useful even for enhancing quality of instrumentals after separation (or your own mixed music) Be aware that at least some advanced mixing beforehand may cheat the content ID detection system, so your song won't be detected. If some label prevents you from uploadi

_Best quality from YouTube and Soundcloud - how to squeeze out the most from the music taken from YT for separation_

Sat, 29 Aug 2026 20:20:11 GMT

___Best quality from YouTube and Soundcloud - how to squeeze out the most from the music taken from YT for separation___ Sometimes a better source just doesn’t exist, and only YouTube audio can be used for separation in some cases. Introduction Audio on YT in most cases is available in two formats:

List of cloud services with a lot of space

Sat, 29 Aug 2026 20:20:11 GMT

List of cloud services with a lot of space

or for temporary storage

Sat, 29 Aug 2026 20:20:11 GMT

or for temporary storage Unlimited https://filegarden.com/ No info on any limits, URL shortener, browser bar player, open source, registration required. Unlimited https://imgur.gg/ 500MB/file and 5GB/file for registered users, no expiration, no registration required. Audio files previewing. Be aware

AI-killing tracks - difficult ones to get instrumentals (or vocals) - a lot of e.g. vocal (or instrumental) leftovers in current models

Sat, 29 Aug 2026 20:20:11 GMT

AI-killing tracks - difficult ones to get instrumentals (or vocals) - a lot of e.g. vocal (or instrumental) leftovers in current models (Moved to the training models doc #2 below)

Training models guides

Sat, 29 Aug 2026 20:20:11 GMT

Training models guides 1. Read mesk’s guide (new link #2), then proceed below for arch explanations and more details. 2. More thorough training guide (just not that step-by-step, condensed form, all the training content was moved from here, there, to avoid 1.02M character limitation of Google Docume