CMU Sphinx vs Kaldi

CMU Sphinx is a free, open-source toolkit for developers building speech applications. Its maintained components are PocketSphinx, a lightweight C speech recognizer library, and SphinxTrain, tools for acoustic model training. The toolkit is designed for efficient recognition on low-resource platforms, and its documented interfaces support keyword spotting, grammar and language-model searches, and word-level segmentation. Recognition requires acoustic and language models. The toolkit is language-independent, with prebuilt models listed for English, Chinese, French, Spanish, German and Russian. SphinxTrain supports adding a language through data collection and cleaning, model training and testing. Noise handling uses MFCC features, noise tracking and spectral subtraction. The BSD-like license permits commercial distribution. Possible applications include voice control, language learning, transcription, closed captioning, speech translation and voice search. Decoders require PCM input, so encoded audio must be converted before processing. The project documents a PocketSphinx Android library and demo, but cautions that its Android tutorial may no longer work with current tools. The homepage announces PocketSphinx 5.1.0.

Kaldi is a free, open-source C++ toolkit for speech recognition aimed at researchers and professionals. Its code and scripts cover methods including linear transforms, MMI, boosted MMI, MCE discriminative training and feature-space discriminative training. The toolkit integrates finite-state transducers through OpenFst and includes a matrix library that wraps standard BLAS and LAPACK routines. Example recipes support building speech recognition systems, and downloadable model archives cover speech recognition, speaker identification, speech activity detection, diarization and language modeling. The models may be used for any purpose. Kaldi is released under Apache License 2.0, which permits commercial use and redistribution. It runs on Linux distributions and Mac OS X, supports Android cross-compilation, and can be compiled to WebAssembly for browser execution. Windows compatibility is not actively maintained, and example scripts do not work there. Users obtain the current development version by cloning the GitHub repository and following its INSTALL file. Documentation is expert-oriented; the project directs help requests to its mailing list and bug reports or feature requests to GitHub issues.