| Name | Remarks | 
|---|
      | CVSROOT |  | 
      | s3.2 | 
	    Location: $CVSROOT/s3.2Fast Sphinx-3 decoder using lextree organization:
	      
		5-10x real time speed on large vocabulary tasksContinuous density acoustic models onlyBatch-Mode operation onlygausubvq: Sub-vector clustered acoustic model building
		Needed for fast acoustic model evaluationDocumentation:
	      
	     | 
      
      | libutil | 
	    Location: $CVSROOT/libutilMiscellaneous utilities needed by s3.2 (some of them by Eric Thayer
	      and Paul Placeway):
	      
		Platform-independent data typesCommand-line arguments parsingHash tablesHeap structures (for sorting)Memory allocationCPU usage profilingError reportingNot Sphinx specific | 
      | s3 | 
	    Location: /net/alf20/usr2/rkm/s3Original Sphinx-3 decoderSlow; 50-100x real time speed on large vocabulary tasksAny kind of acoustic model (discrete, semi-continuous, continuous, others)Major applications:
	      
		s3decodeands3decode-anytopo: Speech-to-text Decodings3align: Forced alignments3allphone: Allphone decodings3astar: A* search, nbest generations3dag: Shortest-path searchOther utilities:
	      
		stseg-read: State-segmentation binary file readersen2s2: Sphinx-II "sendump" file creation from Sphinx-3
		  acoustic modelDocumentation:
	      
	   | 
      | s2 (fbs8) | 
	    Location: Open source (search for
	      "CMU Sphinx")Sphinx-II decoderReal-time operationSemi-continuous, Sphinx-II acoustic models only (Sphinx-II format)User applications support:
	      
		Compiled into a library with a straightforward API for building
		  speech-enabled applicationsContinuous-listening supportDynamic language model loading and switchingSeveral test applications:
	      
		Basic dictation with and without "push-to-talk"Basic audio recording and playbackAudio segmentation using the continuous listenerAdditional recognition modes:
	      
		Forced alignmentAllphone decodingA* search, nbest generationShortest-path search | 
      
      | lm3g2dmp | 
	    Location: $CVSROOT/lm3g2dmpConversion from "arpabo" format language model file to binary ("dump")
	      format used by all decoders | 
      
      | SphinxOCX |  | 
      
      | data | 
	    Location: /net/alf20/usr/rkm/SHAREDAn attempt at collecting all available data and models under one roof (not
	      entirely successful):
	      
		Cepstrum filesControl filesHMMs (acoustic models)Language modelsDictionariesImplemented mainly via symbolic links, rather than physical copies |