Speech recognition

OpenAI Whisper (has nice C++ & CoreML ports), ocotillo & SpeechBrain are nice.

Notes

• Voice assistants don't seem to stick for most people is that they're actually command line interfaces, but even less discoverable because they don't provide any visible feedback at all.

Links

• HN: Facebook open-sources a speech-recognition system and a machine learning library (2018)
• DeepSpeech - Open source Speech-To-Text engine, using a model trained by machine learning techniques, based on Baidu's Deep Speech research paper. (Examples)
• Online speech recognition with wav2letter@anywhere (2020)
• wav2letter++ - Fast, open source speech processing toolkit from the Speech team at Facebook AI Research built to facilitate research in end-to-end models for speech recognition.
• Kaldi - Speech Recognition Toolkit.
• Building an end-to-end Speech Recognition model in PyTorch (HN)
• Real-Time Voice Cloning - Clone a voice in 5 seconds to generate arbitrary speech in real-time.
• Kaldi Active Grammar - Python Kaldi speech recognition with grammars that can be set active/inactive dynamically at decode-time.
• SpecAugment with PyTorch - PyTorch Implementation of GoogleBrain's SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition.
• Dragonfly - Speech recognition framework for Python that makes it convenient to create custom commands to use with speech recognition software.
• Gentle - Robust yet lenient forced-aligner built on Kaldi. A tool for aligning speech with text.
• Porcupine - On-device wake word detection powered by deep learning.
• Eesen - End-to-End Speech Recognition using Deep RNN Models and WFST-based Decoding.
• Ask HN: Is there any work being done in speech-to-code with deep learning? (2020)
• Silero Models - Pre-trained STT models and benchmarks made embarrassingly simple. (HN)
• High-quality pre-trained speech-to-text models now available on Torch Hub (HN)
• Wavenet For Speech Denoising - Neural network for end-to-end speech denoising, as described in: "A Wavenet For Speech Denoising".
• Vosk - Speech recognition toolkit with state-of-the-art accuracy and low latency in Rust.
• Voicegain - Speech-to-text Platform and APIs. Speech Recognition.
• LibreASR - On-Premises, Streaming Speech Recognition System. (HN)
• WORLD - High-quality speech analysis, manipulation and synthesis system. (Web)
• ESPnet - End-to-end speech processing toolkit. (Docs)
• Speaker Diarization - Process to answer the question of 'who spoke when?' in an audio file.
• SpeechRecognition - Local auto speech recognition project based on Kaldi and ALSA.
• Athena - Open-source implementation of sequence-to-sequence based speech processing engine.
• PyTorch end-to-end speech recognition
• Cheetah - On-device streaming speech-to-text engine powered by deep learning.
• WaveRNN - PyTorch implementation of Deepmind's WaveRNN model from Efficient Neural Audio Synthesis.
• Conformer - PyTorch implementation of Conformer: Convolution-augmented Transformer for Speech Recognition.
• A Review of End-to-End Architectures for Speech Recognition (2021)
• libfvad - Voice activity detection (VAD) library, based on WebRTC's VAD engine.
• ASR with PyTorch - Experimental code for speech recognition using PyTorch and Kaldi.
• YSDA Speech Processing Course
• Paper List for Speech Translation
• Deep Contextualized Acoustic Representations For Semi-Supervised Speech Recognition (2020) (Code)
• Lyra: A New Very Low-Bitrate Codec for Speech Compression (2021)
• Parrot.PY - Computer interaction using audio and speech recognition.
• SpeechBrain Toolkit - PyTorch-based Speech Toolkit. (Web)
• Vosk API - Offline open source speech recognition toolkit. (Rust API)
• Lyra - Very Low-Bitrate Codec for Speech Compression.
• lasr - PyTorch Lightning implementation of Automatic Speech Recognition.
• Speech Recognition from Scratch
• Common Voice - Mozilla's initiative to help teach machines how real people speak.
• FullSubNet: A Full-Band and Sub-Band Fusion Model for Real-Time Single-Channel Speech Enhancement (2021) (Code)
• DeepSpeech2 in PyTorch using PyTorch Lightning
• Speech and Language Processing Book (2021) - Introduction to Natural Language Processing, Computational Linguistics, and Speech Recognition. (2020 Version)
• voice2json - Command-line tools for speech and intent recognition on Linux. (Web)
• wav2vec Unsupervised: Speech recognition without supervision (2021)
• Online Speech recognition using RNN-Transducer
• Openspeech - Open-Source Toolkit for End-to-End Speech Recognition.
• Unsupervised Speech Decomposition via Triple Information Bottleneck (2020) (Code)
• AudioCLIP: Extending CLIP to Image, Text and Audio (2021) (Code)
• Wav2vec: Semi and Unsupervised Speech Recognition (HN)
• WeNet - Production First and Production Ready End-to-End Speech Recognition Toolkit. (Docs)
• Why Hasn’t the iPhone Moment Happened Yet for Voice UIs (2021)
• LeBenchmark: a reproducible framework for assessing SSL from speech
• INTERSPEECH 2021
• WER are we? - Tracking states of the art(s) and recent results on speech recognition.
• GigaSpeech - Large, modern dataset for speech recognition.
• Coqui STT - Deep learning toolkit for Speech-to-Text, battle-tested in research and production. (Docs) (Rust lib)
• Coqui - Startup providing open speech tech for everyone. (GitHub)
• Open Speech Corpora - List of accessible speech corpora for ASR, TTS, and other Speech Technologies.
• An Overview of Multi-Task Learning in Speech Recognition (2020)
• Coqui Inference Engine - Library for efficiently deploying speech models.
• PDF to Speech - Deep-learning powered accessibility application which turns PDFs into audio files.
• ASV-Subtools - Open Source Tools for Speaker Recognition.
• VoiceFixer - General Speech Restoration.
• speechmetrics - Wrapper around speech quality metrics MOSNet, BSSEval, STOI, PESQ, SRMR, SISDR.
• Silero VAD - Pre-trained enterprise-grade Voice Activity Detector, Language Classifier and Spoken Number Detector.
• A New AI Lexicon: Voice (2021) - The Legacies and Limits of Automated Voice Analysis.
• Octopus - On-device speech-to-index engine powered by deep learning.
• Open Audio Search - Full text search engine with automatic speech recognition for podcasts.
• HuBERT: How to Apply BERT to Speech, Visually Explained (2021)
• Happy Scribe - Audio Transcription & Video Subtitles.
• Speech Recognition Papers
• Steerable discovery of neural audio effects (2021) (Code)
• audapolis - Editor for spoken-word media with transcription.
• Shennong - Python toolbox for speech features extraction.
• Paderbox - Collection of utilities for audio / speech processing.
• Icefall - Speech recognition recipes using k2. (Docs)
• k2 - FSA/FST algorithms, differentiable, with PyTorch compatibility.
• ViSQOL (Virtual Speech Quality Objective Listener) - Objective, full-reference metric for perceived audio quality.
• Espresso - Fast End-to-End Neural Speech Recognition Toolkit.
• UniSpeech - Large Scale Self-Supervised Learning for Speech
• NISQA: Speech Quality and Naturalness Assessment
• Optimization techniques proposed in Improving RNN Transducer Modeling for End-to-End Speech Recognition
• Conformer: Convolution-augmented Transformer for Speech Recognition (2020) (Code)
• CAT: Crf-based Asr Toolkit - Complete workflow for CRF-based data-efficient end-to-end speech recognition.
• Neural HMMs are all you need (for high-quality attention-free TTS) (2022) (Code)
• End-to-End Speech Translation Progress - Tracking the progress in end-to-end speech translation.
• EfficientTTS: An Efficient and High-Quality Text-to-Speech Architecture (2020) (Code)
• S3PRL - Self-Supervised Speech Pre-training and Representation Learning Toolkit.
• pyannote-audio - Neural building blocks for speaker diarization: speech activity detection, speaker change detection, overlapped speech detection, speaker embedding.
• DiffSinger: Singing Voice Synthesis via Shallow Diffusion Mechanism (2021) (Code)
• Speech recognition polyfill - Polyfill for the SpeechRecognition standard on web, using Speechly as the underlying API.
• Speech-to-Text Benchmark
• Hyperion - Speaker Recognition Toolkit based on PyTorch and numpy.
• textlesslib - Library for Textless Spoken Language Processing.
• FastSpeech 2: Fast and High-Quality End-to-End Text-to-Speech (2021) (Code)
• HuggingSound - Toolkit for speech-related tasks based on HuggingFace's tools.
• hear - macOS speech recognition via the command line.
• PaddleSpeech - Easy-to-use Speech Toolkit including SOTA ASR pipeline, influential TTS with text frontend and End-to-End Speech Simultaneous Translation.
• BYOL for Audio: Self-Supervised Learning for General-Purpose Audio Representation (2021) (Code)
• Edinburgh Speech Tools
• rVADfast - Python library for an unsupervised, fast method for robust voice activity detection.
• NeuralSpeech - Research project in Microsoft Research Asia focusing on neural network based speech processing, including automatic speech recognition (ASR), text to speech (TTS), etc.
• Speech Super-resolution Evaluation and Benchmarking
• Real Time Speech Recognition with Gradio (HN)
• Assem-VC: Realistic Voice Conversion by Assembling Modern Speech Synthesis Techniques (2021) (Code)
• CoVoST: A Large-Scale Multilingual Speech-To-Text Translation Corpus
• Learning Audio-Visual Speech Representation by Masked Multimodal Cluster Prediction (2022) (Code)
• Real Time Speech Enhancement in the Waveform Domain (2020) (Code)
• Vosk-Browser - Opinionated speech recognition library for the browser using a WebAssembly build of Vosk.
• VocalSound: A Dataset for Improving Human Vocal Sounds Recognition
• PANNs: Large-Scale Pretrained Audio Neural Networks for Audio Pattern Recognition
• NaturalSpeech: End-to-End Text to Speech Synthesis with Human-Level Quality (2022) (HN)
• George Hotz | Programming | speech recognition (2022)
• NaturalSpeech: End-to-End Text to Speech Synthesis with Human-Level Quality (2022) (Code) (Code)
• CoquiSTT + Signal = Love (death to voice messages) (2022)
• ocotillo - PyTorch-based ML model that does state-of-the-art English speech transcription.
• SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing (2021) (Code)
• pyctcdecode - Fast and lightweight python-based CTC beam search decoder for speech recognition.
• Avocodo: Generative Adversarial Network for Artifact-free Vocoder (2022) (Code)
• Squeezeformer - PyTorch implementation of "Squeezeformer: An Efficient Transformer for Automatic Speech Recognition".
• Masked Autoencoders that Listen (2022) (Code)
• SyntaSpeech: Syntax-Aware Generative Adversarial Text-to-Speech (2022) (Code)
• Speech Enhancement and Dereverberation with Diffusion-based Generative Models
• SSAST: Self-Supervised Audio Spectrogram Transformer
• OpenAI Whisper - General-purpose speech recognition. Approaches human level robustness and accuracy on English speech recognition. (Web) (HN) (Notes) (Paper + Code walkthroughs) (Demo) (Demo Code) (Failure case)
• Whisper ASR Web service
• React-Speech-Recognition - Speech recognition for your React app.
• Stage-Whisper - Easy to use AI transcriber, powered by OpenAI's Whisper.
• Whisper.cpp - High-performance inference of OpenAI's Whisper automatic speech recognition (ASR) model. (HN)
• Real-time speech recognition using next-gen Kaldi with ncnn
• Gecko - Tool for Effective Annotation of Human Conversations. (Web)
• OpenAI Whisper - CPU - Improving transcription performance of OpenAI Whisper for CPU based deployment.
• Whispering - Streaming transcriber with whisper.
• Buzz - Transcribe and translate audio offline on your personal computer. Powered by OpenAI's Whisper.
• FastWhisper - Optimized implementation of OpenAI's Whisper for multilingual transcription.
• I record myself on audio 24x7 and use an AI to process the information (2022) (HN)
• Transcribe File - Free transcription service powered by Whisper AI. (HN)
• SpeechRecognition - Speech recognition module for Python, supporting several engines and APIs, online and offline.
• TransFusion: Transcribing Speech with Multinomial Diffusion
• Convmelspec: Melspectrograms for On-Device Audio Machine Learning (2022)
• Lightning Echo - Production-ready audio and video transcription app that can run on your laptop or in the cloud.
• Best way to transcribe audio snippets (2022)
• Whisper's transcription plus Pyannote's Diarization
• WhisperX - Timestamp-Accurate Automatic Speech Recognition.
• Speech-to-text with Whisper: How I Use It & Why (2022)
• SPTK - Speech Signal Processing Toolkit.
• Speechbox - Speech processing tools, such as punctuation restoration.
• Zero-shot Punctuation Insertion using Whisper
• Zero-shot Audio Classification using Whisper
• Whisper CLI, built with Rust
• Whisper-rs - Rust bindings to whisper.cpp.
• OpenAI's Whisper ported to CoreML
• Real Time Whisper Transcription
• Multilingual Automatic Speech Recognition with Word-level Timestamps
• Quick Caption - Transcribe and generate caption files (SRT and FCPXML) without manually entering time codes.
• ArchiSound - Collection of pre-trained audio models, in PyTorch.
• MacWhisper - Quickly and easily transcribe audio files into text with OpenAI's state-of-the-art transcription technology Whisper.
• Whisper.cpp example running fully in the browser (HN)
• Mesostructures: Beyond Spectrogram Loss in Differentiable Time-Frequency Analysis (2023) (Code)
• CLAP (Contrastive Language-Audio Pretraining) - Neural network model that learns acoustic concepts from natural language supervision.
• Speaker Diarization Using OpenAI Whisper
• WaaS - Self-host Whisper As a Service with GUI and queueing. (HN)
• Faster Whisper transcription with CTranslate2
• FunASR: Fundamental End-to-End Speech Recognition Toolkit
• Streamlit UI for OpenAI's Whisper
• Whisperer - Go from raw audio files to a speaker separated text-audio datasets automatically.
• Whisper Node - Node.js bindings for OpenAI's Whisper.
• Stabilizing Timestamps for Whisper
• TriAAN-VC: Triple Adaptive Attention Normalization for any-to-any Voice Conversion (2023)