Module | Name | Default | Description |
UPDATEMEANS | F | Update means | |
HCOMPV | SAVEBINARY | F | Load/Save in binary format |
MINVARFLOOR | 0.0 | Minimum variance floor | |
NSTREAMS | 1 | Number of streams | |
SAVEASVQ | F | Save only the VQ indices | |
HCOPY | SOURCEFORMAT | HTK | File format of source |
TARGETFORMAT | HTK | File format of target | |
SOURCEKIND | ANON | Parameter kind of source | |
TARGETKIND | ANON | Parameter kind of target | |
SAVEBINARY | F | Load/Save in binary format | |
HEREST | BINARYACFORMAT | T | Load/Save accumulators in binary format |
| HSKIPSTART | -1 | Start of skip over region (debugging only) |
HSKIPEND | -1 | End of skip over region (debugging only) | |
HHED | TREEMERGE | T | After tree splitting, merge leaves |
TIEDMIXNAME | TM | Tied mixture base name | |
HPARSE | V1COMPAT | F | Enable compatibility with HTK V1.X |
REFLEVEL | 0 | Label level to be used as reference | |
TESTLEVEL | 0 | Label level to be scored | |
STRIPCONTEXT | F | Strip triphone contexts | |
HRESULTS | IGNORECASE | F | If enabled, converts labels to uppercase |
NISTSCORE | F | Use NIST fomatting | |
PHRASELABEL | SENT | Label for phrase level statistics | |
PHONELABEL | WORD | Label for word level statistics | |
SPEAKERMASK | NULL | If set then report on a per speaker basis | |
HVITE | RECOUTPREFIX | NULL | Prefix for direct audio output name |
RECOUTSUFFIX | NULL | Suffix for direct audio output name | |
HLSTATS | DISCOUNT | 0.5 | Discount constant for backoff bigrams |
HLIST | AUDIOSIG | 0 | Audio signal numberfor remote control |
SOURCERATE | 0.0 | Sample rate of source in 100ns units | |
TRACE | 0 | Trace setting |