SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1607.05666 · 2016

Trainable Frontend For Robust and Far-Field Keyword Spotting

arXiv · PDF · Open in the Atlas

Code that ran

8 of 10 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 1 violated it and 7 have no contract verdict; 0 failures of our instrument; 2 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/1607.05666.json · badge: /api/ran/1607.05666.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
asteroid-team/asteroid-filterbanks community (classifier label)
classifier basis not recorded on the link
8 ran · 2 no recorded run
FunctionStatusWhere it lives
batch_packed_1d_conv Ran
ran on a synthesized input
asteroid-team/asteroid-filterbanks/asteroid_filterbanks/enc_dec.py
code served (permissive licence) · get_code("e27f2fb58bd944e4")
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 Tensor, ndim 3, dtype family float, shape (2, 3, 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 torch
import torch.nn as nn
from torch.nn import functional as F

torch.manual_seed(0)

# Create input tensor: shape (batch=2, sequence_length=4)
# This will be reshaped to (2, 1, 4) in the function for conv1d
inp = torch.randn(2, 4)

# Create filters tensor: shape (out_channels=3, in_channels=1, kernel_size=2)
# conv1d expects filters of shape (out_channels, in_channels, kernel_size)
filters = torch.randn(3, 1, 2)

stride = 1
padding = 0

ARGS = {
    "inp": inp,
    "filters": filters,
    "stride": stride,
    "padding": padding,
}
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 Tensor, ndim 3, dtype family float, shape (2, 3, 3) — shape (2, 3, 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
erb_scale_2_freq_hz Ran
ran, violated the contract we drafted
asteroid-team/asteroid-filterbanks/asteroid_filterbanks/multiphase_gammatone_fb.py
code served (permissive licence) · get_code("39ebbabf630b2f81")
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 1, dtype family float, shape (4,) 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

# The function takes a single parameter: freq_erb
# It expects a numeric value or array-like that can be used with np.exp and arithmetic operations

ARGS = {
    "freq_erb": np.array([1.0, 2.5, 5.0, 10.0])
}
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 1, dtype family float, shape (4,) — shape (4,) 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
gammatone_impulse_response Ran
ran on a synthesized input
asteroid-team/asteroid-filterbanks/asteroid_filterbanks/multiphase_gammatone_fb.py
code served (permissive licence) · get_code("7ee826bfe8bb4bd1")
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 1, dtype family float, shape (8000,) 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 math
import numpy as np

ARGS = {
    "samplerate_hz": 16000,
    "len_sec": 0.5,
    "center_freq_hz": 440.0,
    "phase_shift": 0.0
}
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 1, dtype family float, shape (8000,) — shape (8000,) 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
generate_mpgtf Ran
ran on a synthesized input
asteroid-team/asteroid-filterbanks/asteroid_filterbanks/multiphase_gammatone_fb.py
code served (permissive licence) · get_code("779d3152b94b1fb5")
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 2, dtype family float, shape (48, 8000) 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 math
import numpy as np

# All parameters for generate_mpgtf
ARGS = {
    "samplerate_hz": 16000,
    "len_sec": 0.5,
    "n_filters": 48
}
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 2, dtype family float, shape (48, 8000) — shape (48, 8000) 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
mark_deprecated Ran
ran on a synthesized input
asteroid-team/asteroid-filterbanks/asteroid_filterbanks/deprecation.py
code served (permissive licence) · get_code("8de8d30f605a4889")
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 warnings
from functools import wraps

class VisibleDeprecationWarning(UserWarning):
    """Visible deprecation warning."""
    pass

def mark_deprecated(message, version=None):
    """Decorator to add deprecation message.

    Args:
        message: Migration steps to be given to users.
    """

    def decorator(func):

        @wraps(func)
        def wrapped(*args, **kwargs):
            from_what = 'a future release' if version is None else f'asteroid v{version}'
            warn_message = f'{func.__module__}.{func.__name__} has been deprecated and will be removed from {from_what}. {message}'
            warnings.warn(warn_message, VisibleDeprecationWarning, stacklevel=2)
            return func(*args, **kwargs)
        return wrapped
    return decorator


# Define a simple test function to decorate
def dummy_function():
    return "test"


# Test the decorator with different argument combinations
ARGS = {
    "message": "Please use new_function instead. It has a better API.",
    "version": "0.4.0"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
perfect_synthesis_window Ran
ran on a synthesized input
asteroid-team/asteroid-filterbanks/asteroid_filterbanks/stft_fb.py
code served (permissive licence) · get_code("5d963eb6c86771b1")
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 1, dtype family float, shape (512,) 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 simple analysis window
# Using a Hann window as a realistic example
window_length = 512
analysis_window = np.hanning(window_length)

# Create a hop size that's a reasonable fraction of window length
# For STFT, hop_size is typically 1/4 or 1/2 of window length
hop_size = 128

ARGS = {
    "analysis_window": analysis_window,
    "hop_size": hop_size,
}
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 1, dtype family float, shape (512,) — shape (512,) 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
script_if_tracing Ran
ran on a synthesized input
asteroid-team/asteroid-filterbanks/asteroid_filterbanks/scripting.py
code served (permissive licence) · get_code("a84548a7c681b48d")
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 torch
import functools

torch.manual_seed(0)

def simple_function(x, y):
    """A simple function to be used with script_if_tracing"""
    return x + y

def another_function(a):
    """Another simple function"""
    return a * 2

ARGS = {
    "fn": simple_function
}
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
square_ola Ran
ran on a synthesized input
asteroid-team/asteroid-filterbanks/asteroid_filterbanks/torch_stft_fb.py
code served (permissive licence) · get_code("a518652d7056ad7f")
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 Tensor, ndim 3, dtype family float, shape (1, 1, 8) 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 torch

torch.manual_seed(0)

# Analyze the function:
# - window: torch.Tensor, used with .pow(2), .view(1, -1, 1), .repeat(1, 1, n_frame)
# - kernel_size: int, used in fold operation
# - stride: int, used in fold operation
# - n_frame: int, used in repeat and fold calculations

# The function applies overlap-add with a squared window
# window needs to be 1D tensor of size kernel_size
kernel_size = 4
stride = 2
n_frame = 3

# Create a small window tensor (1D)
window = torch.hann_window(kernel_size, periodic=False)

ARGS = {
    "window": window,
    "kernel_size": kernel_size,
    "stride": stride,
    "n_frame": n_frame,
}
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 Tensor, ndim 3, dtype family float, shape (1, 1, 8) — shape (1, 1, 8) 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
griffin_lim Not yet run
unverified (harvested, no recorded run)
asteroid-team/asteroid-filterbanks/asteroid_filterbanks/griffin_lim.py
code served (permissive licence) · get_code("8923f688906e1436")
How it ran

no invocation recorded.

misi Not yet run
unverified (harvested, no recorded run)
asteroid-team/asteroid-filterbanks/asteroid_filterbanks/griffin_lim.py
code served (permissive licence) · get_code("93bfd5f9c97e18fb")
How it ran

no invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 1 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 1 later) and 9 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
erb_scale_2_freq_hz 1 other paper
1 later
1910.11615 (2019-10)
A Multi-Phase Gammatone Filterbank for Speech Separation via TasNet

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

Robust and far-field speech recognition is critical to enable true hands-free communication. In far-field conditions, signals are attenuated due to distance. To improve robustness to loudness variation, we introduce a novel frontend called per-channel energy normalization (PCEN). The key ingredient of PCEN is the use of an automatic gain control based dynamic compression to replace the widely used static (such as log or root) compression. We evaluate PCEN on the keyword spotting task. On our large rerecorded noisy and far-field eval sets, we show that PCEN significantly improves recognition performance. Furthermore, we model PCEN as neural network layers and optimize high-dimensional PCEN parameters jointly with the keyword spotting acoustic model. The trained PCEN frontend demonstrates significant further improvements without increasing model complexity or inference-time cost.

For agents

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

get_harvested_code_for_paper("1607.05666")
get_code_for_paper("1607.05666")
have("1607.05666")

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

curl https://syntology.ai/api/ran/1607.05666.json

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

[![Syntology run record](https://syntology.ai/api/ran/1607.05666.svg)](https://syntology.ai/paper/1607.05666)

Connect an agent — have() is free.