SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2003.01531 · ICML · 2020

Voice Separation with an Unknown Number of Multiple Speakers

Lior Wolf, Eliya Nachmani, Yossi Adi

arXiv · PDF · Open in the Atlas

Code that ran

12 of 13 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (5 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 1 violated it and 8 have no contract verdict; 2 failures of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2003.01531.json · badge: /api/ran/2003.01531.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
muhammad-ahmed-ghani/svoice_demo community
Papers with Code archive
7 ran · 1 no recorded run
Mack189/gdprnn community (classifier label)
classifier basis not recorded on the link
4 ran · 0 no recorded run
facebookresearch/svoice marked official in Papers with Code
Papers with Code archive
1 ran · 0 no recorded run
FunctionStatusWhere it lives
ByPass Ran
ran on a synthesized input
muhammad-ahmed-ghani/svoice_demo/svoice/models/swave.py
pointer only (licence: NOASSERTION) · get_code("173650e5ae028b2f")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, muhammad-ahmed-ghani/svoice_demo (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

DPMulCat Ran
ran on a synthesized input · constructed an object; no result was computed
muhammad-ahmed-ghani/svoice_demo/svoice/models/swave.py
pointer only (licence: NOASSERTION) · get_code("3b1d338215096fe7")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, muhammad-ahmed-ghani/svoice_demo (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Decoder Ran
ran on a synthesized input · constructed an object; no result was computed
muhammad-ahmed-ghani/svoice_demo/svoice/models/swave.py
pointer only (licence: NOASSERTION) · get_code("5f72b86aea22a846")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, muhammad-ahmed-ghani/svoice_demo (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Encoder Ran
ran on a synthesized input · constructed an object; no result was computed
muhammad-ahmed-ghani/svoice_demo/svoice/models/swave.py
pointer only (licence: NOASSERTION) · get_code("820a2cedf7915ddd")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, muhammad-ahmed-ghani/svoice_demo (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

MulCatBlock Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/svoice/svoice/models/swave.py
pointer only (licence: NOASSERTION) · get_code("92748185c4ff9d16")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, facebookresearch/svoice (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Separator Ran
ran on a synthesized input · constructed an object; no result was computed
muhammad-ahmed-ghani/svoice_demo/svoice/models/swave.py
pointer only (licence: NOASSERTION) · get_code("facca3bf4fbfc471")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, muhammad-ahmed-ghani/svoice_demo (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

cal_SISNR Ran
ran, violated the contract we drafted
Mack189/gdprnn/evaluate.py
code served (permissive licence) · get_code("891e417fc38edeb0")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create two 1D numpy arrays of the same length
# These represent reference and output signals
np.random.seed(0)
ref_sig = np.random.randn(100)
out_sig = np.random.randn(100)

ARGS = {
    'ref_sig': ref_sig,
    'out_sig': out_sig,
    'eps': 1e-08
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
cal_SISNRi Ran
ran, honoured the contract we drafted
Mack189/gdprnn/evaluate.py
code served (permissive licence) · get_code("ca6c3339d98f2a2c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Seed for reproducibility
np.random.seed(0)

# Create small synthetic audio signals
# src_ref: [C, T] - reference sources (2 channels, 100 time samples)
src_ref = np.random.randn(2, 100).astype(np.float32)

# src_est: [C, T] - estimated/separated sources (2 channels, 100 time samples)
src_est = src_ref + 0.1 * np.random.randn(2, 100).astype(np.float32)

# mix: [T] - mixture signal (100 time samples)
mix = src_ref[0] + src_ref[1]

ARGS = {
    "src_ref": src_ref,
    "src_est": src_est,
    "mix": mix,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
capture_init Ran
ran; our contract draft was wrong, not the code
muhammad-ahmed-ghani/svoice_demo/svoice/models/swave.py
pointer only (licence: NOASSERTION) · get_code("ccacfa5207d7243d")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, muhammad-ahmed-ghani/svoice_demo (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

capture_init Ran
ran on a synthesized input
Mack189/gdprnn/src/models/swave.py
code served (permissive licence) · get_code("ecdad0e168981f5d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned function. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import functools
import torch
import torch.nn as nn

def capture_init(init):
    """
    Decorate `__init__` with this, and you can then
    recover the *args and **kwargs passed to it in `self.kwarg`
    """

    @functools.wraps(init)
    def __init__(self, *args, **kwargs):
        self.kwarg = (args, kwargs)
        init(self, *args, **kwargs)
    return __init__


class SimpleModel(nn.Module):
    @capture_init
    def __init__(self, hidden_dim, num_layers=2, dropout=0.1):
        super().__init__()
        self.hidden_dim = hidden_dim
        self.num_layers = num_layers
        self.dropout = dropout
        self.linear = nn.Linear(hidden_dim, hidden_dim)
    
    def forward(self, x):
        return self.linear(x)


torch.manual_seed(0)

instance = SimpleModel(hidden_dim=64, num_layers=3, dropout=0.2)

ARGS = {
    "init": instance.__init__.__wrapped__,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
overlap_and_add Ran
ran; our contract draft was wrong, not the code
muhammad-ahmed-ghani/svoice_demo/svoice/models/swave.py
pointer only (licence: NOASSERTION) · get_code("9c926905c0e7a1c7")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, muhammad-ahmed-ghani/svoice_demo (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

pad_list Ran
ran on a synthesized input
Mack189/gdprnn/preprocess.py
code served (permissive licence) · get_code("1e3f6a2103e513ce")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 3, dtype family float, shape (2, 2, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create a list of arrays with different lengths to test padding
xs = [
    np.array([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]], dtype=np.float32),  # shape (2, 3)
    np.array([[7.0, 8.0, 9.0]], dtype=np.float32),  # shape (1, 3)
]

ARGS = {
    "xs": xs
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family float, shape (2, 2, 3) — shape (2, 2, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
SWave Not yet run
unverified (harvested, no recorded run)
muhammad-ahmed-ghani/svoice_demo/svoice/models/swave.py
pointer only (licence: NOASSERTION) · get_code("8967a79d23072fba")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, muhammad-ahmed-ghani/svoice_demo (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 3 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 3 later) and 10 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
cal_SISNRi 4 other papers
3 dated earlier, 1 later
1809.07454 (2018-09)
Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separation
1910.06379 (2019-10)
Dual-path RNN: efficient long sequence modeling for time-domain single-channel speech separation
1910.14104 (2019-10)
End-to-end Microphone Permutation and Number Invariant Multi-channel Speech Separation
and 1 more not named here
cal_SISNR 3 other papers
2 dated earlier, 1 later
1809.07454 (2018-09)
Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separation
1910.06379 (2019-10)
Dual-path RNN: efficient long sequence modeling for time-domain single-channel speech separation
2007.13975 (2020-07)
Dual-Path Transformer Network: Direct Context-Aware Modeling for End-to-End Monaural Speech Separation
capture_init
brick of this paper's assembly: helper
2 other papers
2 later
2007.06833 (2020-07)
Sudo rm -rf: Efficient Networks for Universal Audio Source Separation
2211.12232 (2022-11)
AERO: Audio Super Resolution in the Spectral Domain

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We present a new method for separating a mixed audio sequence, in which multiple voices speak simultaneously. The new method employs gated neural networks that are trained to separate the voices at multiple processing steps, while maintaining the speaker in each output channel fixed. A different model is trained for every number of possible speakers, and the model with the largest number of speakers is employed to select the actual number of speakers in a given sample. Our method greatly outperforms the current state of the art, which, as we show, is not competitive for more than two speakers.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2003.01531")
get_code_for_paper("2003.01531")
have("2003.01531")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2003.01531.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2003.01531.svg)](https://syntology.ai/paper/2003.01531)

Connect an agent — have() is free.