kaldi-native-fbank
Decision gist · record as of 2026-08-14
Yes, if you are building a real-time speech recognition system that needs Kaldi-compatible fbank features without external dependencies. The package is well-maintained, permissively licensed, and has broad platform coverage. However, if your use case is offline batch processing or you already have Kaldi or another audio library integrated, the aging maintenance status (309 days since last release) and lack of recent activity may warrant checking whether active alternatives exist.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires Python >= 3.8.0; audio samples must be provided as a list or array-like object compatible with the C++ backend.
- Medium install friction due to compiled wheels for multiple Python versions and architectures (cp310–cp313, x86/arm/aarch64, Linux/macOS/Windows), but pre-built binaries are available.
- Maintenance is aging: last release was 309 days ago, though the repository remains active and the project is not archived.
License · maintenance · safety
Apache-2.0 (permissive) — Licensed under Apache-2.0 (permissive), allowing commercial and private use with minimal restrictions; you must retain license and copyright notices in distributions.
last release 2025-10-09 (309 days) · last repo commit 2025-10-09 · 152 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 257,692 downloads/mo, #8,437 on PyPI
Alternatives
Verify before relying
pip install kaldi-native-fbank
import kaldi_native_fbank as knf
opts = knf.FbankOptions()
opts.mel_opts.num_bins = 80
fbank = knf.OnlineFbank(opts)
fbank.accept_waveform(16000, audio_samples)
features = fbank.get_frame(0)- Whether the package works reliably on Android as claimed in the description, or if that is aspirational.
- Performance characteristics (latency, throughput) for real-time speech recognition use cases.
- Compatibility with specific speech recognition frameworks beyond the two CMake projects mentioned.
What it is and what it does
kaldi-native-fbank is a Python wrapper around a C++ implementation of Kaldi-compatible filterbank feature extraction. It computes mel-frequency filterbank coefficients from raw audio waveforms, designed for real-time speech recognition pipelines. The package accepts audio samples at a specified sampling rate and produces frame-by-frame feature vectors without requiring external audio libraries or Kaldi itself.
The package is built as a compiled extension with pre-built wheels for Python 3.8+ across Linux, macOS, and Windows on x86, ARM, and aarch64 architectures. It has no runtime Python dependencies and is intended for integration into speech recognition systems like sherpa-ncnn and sherpa-onnx, where it handles the feature extraction stage of online (streaming) speech processing.
Use it for
- Extract mel-filterbank features from streaming audio in real-time speech recognition systems.
- Replace Kaldi's fbank computation in modern speech pipelines without installing Kaldi itself.
- Compute frame-level audio features for acoustic model inference in embedded or resource-constrained environments.
- Validate or benchmark fbank feature extraction against Kaldi's reference implementation.
- Integrate audio preprocessing into Python-based speech recognition frameworks.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, if you are building a real-time speech recognition system that needs Kaldi-compatible fbank features without external dependencies.
The package is well-maintained, permissively licensed, and has broad platform coverage. However, if your use case is offline batch processing or you already have Kaldi or another audio library integrated, the aging maintenance status (309 days since last release) and lack of recent activity may warrant checking whether active alternatives exist.
Install
kaldi-native-fbank on PyPI
Before you install
Medium install friction due to compiled wheels for multiple Python versions and architectures (cp310–cp313, x86/arm/aarch64, Linux/macOS/Windows), but pre-built binaries are available. Maintenance is aging: last release was 309 days ago, though the repository remains active and the project is not archived.
Requires Python >= 3.8.0; audio samples must be provided as a list or array-like object compatible with the C++ backend.
License in practice
Licensed under Apache-2.0 (permissive), allowing commercial and private use with minimal restrictions; you must retain license and copyright notices in distributions.
Quickstart
pip install kaldi-native-fbank
import kaldi_native_fbank as knf
opts = knf.FbankOptions()
opts.mel_opts.num_bins = 80
fbank = knf.OnlineFbank(opts)
fbank.accept_waveform(16000, audio_samples)
features = fbank.get_frame(0)Verify before relying
- Whether the package works reliably on Android as claimed in the description, or if that is aspirational.
- Performance characteristics (latency, throughput) for real-time speech recognition use cases.
- Compatibility with specific speech recognition frameworks beyond the two CMake projects mentioned.
Package facts
| License | Apache-2.0 permissive |
| Python support | Supports the current Python release >=3.8.0 |
| Install friction | Medium. Platform-specific wheel |
| Runtime dependencies | None |
| Maintenance | Aging 309 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 257,692 / month, #8,437 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Programming Language :: C++Programming Language :: PythonProgramming Language :: Python :: 3Topic :: Scientific/Engineering :: Artificial Intelligence |
Evidence: kaldi_native_fbank-1.22.3-cp310-cp310-macosx_10_15_x86_64.whl; kaldi_native_fbank-1.22.3-cp310-cp310-macosx_11_0_arm64.whl; kaldi_native_fbank-1.22.3-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.whl; kaldi_native_fbank-1.22.3-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.whl; kaldi_native_fbank-1.22.3-cp310-cp310-win32.whl; kaldi_native_fbank-1.22.3-cp310-cp310-win_amd64.whl; kaldi_native_fbank-1.22.3-cp311-cp311-macosx_10_15_x86_64.whl; kaldi_native_fbank-1.22.3-cp311-cp311-macosx_11_0_arm64.whl; kaldi_native_fbank-1.22.3-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl; kaldi_native_fbank-1.22.3-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl; kaldi_native_fbank-1.22.3-cp311-cp311-win32.whl; kaldi_native_fbank-1.22.3-cp311-cp311-win_amd64.whl; kaldi_native_fbank-1.22.3-cp312-cp312-macosx_10_15_x86_64.whl; kaldi_native_fbank-1.22.3-cp312-cp312-macosx_11_0_arm64.whl; kaldi_native_fbank-1.22.3-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.whl; kaldi_native_fbank-1.22.3-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl; kaldi_native_fbank-1.22.3-cp312-cp312-win32.whl; kaldi_native_fbank-1.22.3-cp312-cp312-win_amd64.whl; kaldi_native_fbank-1.22.3-cp313-cp313-macosx_10_15_x86_64.whl; kaldi_native_fbank-1.22.3-cp313-cp313-macosx_11_0_arm64.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “audio feature extraction fbank”
- kaldi-native-fbankExtracts Kaldi-compatible filterbank (fbank) audio features from…
- kaldiioKaldiio reads and writes Kaldi archive (ark) and script (scp) files…
- torchlibrosaProvides PyTorch implementations of librosa audio feature extraction…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also kaldifst · python_speech_features · aubio · kaldiio · kaldialign · asteroid-filterbanks · lhotse · Gammatone · torchaudio · livekit-plugins-soniox