| A_CELL | Transition arc of HMM state |
| A_CELL2 | Transition arc holding cell |
| ADIn | A/D-in work area |
| APATNODE | Patricia binary tree node, to search related pointer from string |
| BACKTRELLIS | Whole word trellis (aka backtrellis) generated as a result of 1st pass |
| BASEPHONE | Basephone information extracted from hmminfo |
| BMALLOC_BASE | Information of allocated memory block for mybmalloc() |
| CD_Set | Context-dependent HMM set (called "pseudo") for a logical context |
| CD_State_Set | |
| CMEAN | Structure to hold sentence sum of MFCC for realtime CMN |
| CMNWork | Work area for real-time CMN |
| CN_CLUSTER | Word Cluster for confusion network generation |
| DeltaBuf | Cycle buffer for delta computation |
| DFA_ARC | Transition arc of DFA |
| DFA_INFO | Top structure of a DFA |
| DFA_STATE | State of DFA |
| DS_BUFFER | Down sampling data |
| DS_FILTER | Down sampling filter |
| ENERGYWork | Work area for energy normalization on real time input |
| FBankInfo | Workspace for filterbank analysis |
| FSBeam | Work area for the first pass |
| GCODEBOOK | Gaussian mixture codebook in tied-mixture model |
| GMMCalc | Work area for GMM calculation |
| GRAMLIST | List of grammars to be read at startup |
| GS_SET | Set of HMM states for Gaussian Mixture Selection |
| HMM | HMM state sequence |
| HMM_basephone | List of all basephone in hmminfo |
| HMM_CDSET_INFO | Top structure to hold all the HMM sets |
| HMM_Logical | Logical HMM to map logical names to physical/pseudo HMM |
| HMM_STATE | HMM State |
| HMMWork | Work area and cache for HMM computation |
| HTK_HMM_Data | Top HMM model, corresponds to "~h" macro in hmmdefs |
| HTK_HMM_Dens | HMM Gaussian density (or mixture) data |
| HTK_HMM_INFO | Top HMM structure that holds all the HTK HMM definition |
| HTK_HMM_Options | HMM Option |
| HTK_HMM_State | HMM state data |
| HTK_HMM_StreamInfo | Stream information (although current Julius supports only single stream) |
| HTK_HMM_Trans | HMM transition table |
| HTK_HMM_Var | HMM variance data |
| HTK_Param | Input speech parameter |
| HTK_Param_Header | Parameter types and extraction conditions |
| Jconf | Configuration parameters (global) |
| JCONF_AM | Configuration parameters (AM) |
| JCONF_LM | Language models (N-gram / DFA), dictionary, and related parameters |
| JCONF_SEARCH | Search parameters |
| LM_PROB_CACHE | LM cache for the 1st pass |
| LMFunc | User LM function entry point |
| MFCCCalc | Instance for a parameter vector computation |
| MFCCWork | Work area for MFCC computation |
| MIXCACHE | A component of per-codebook probability cache while search |
| MULTIGRAM | Grammar holder |
| NEXTWORD | 第2パスの次単語候補 |
| NGRAM_INFO | Main N-gram structure |
| NGRAM_TUPLE_INFO | N-gram entries for a m-gram (1 <= m <= N) |
| NODE | 第2パスの文仮説 |
| OptionStr | Structure for decoding/encoding parameter type code |
| Output | Recognition result output structure |
| PATNODE | Another patricia binary tree node, to search integer value from string |
| PROCESS_AM | Instance for an AM |
| PROCESS_LM | Instance for a LM |
| RealBeam | Work area for realtime processing of 1st pass |
| Recog | Top level instance for the whole recognition process |
| RecogProcess | Instance for a decoding, i.e |
| S_CELL | Element of successor word list for LM factoring computation |
| SEGTOKEN | Token definition for viterbi segmentation |
| Sentence | Output result structure |
| StackDecode | Work area for the 2nd pass |
| TERM_INFO | Information of each terminal symbol (=category) |
| TOKEN2 | Token to hold viterbi pass history |
| TRELLIS_ATOM | Word trellis element that holds survived word ends at each frame on the 1st pass |
| USEROPT | User-defined option |
| Value | Mfcc configuration parameter values |
| WCHMM_INFO | Whole lexicon tree structure holding all information |
| WCHMM_STATE | HMM state on tree lexicon |
| WCHMM_WORK | Wchmm-specific work area |
| WORD_INFO | Word dictionary structure to hold vocabulary |
| WordGraph | 単語グラフ上の単語候補 |
| ZEROCROSS | Work area for zero-cross computation |
1.5.4