Hello guys, I am new to CMU Sphinx. I am using this api in Java language. When
I was trying out HelloWorld program I had found out that .gram file (grammar
file) is important for running the program and the program will only show
words that are in grammar file. I had a question in my mind that is there any
way by which I can print values which are coming from dictionary file while I
am speaking something without the dependence of grammar file.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Sphinx can either use "grammars" or "language models". A language model is a
set of rules for what words can follow other words that comes from analyzing
enormous amounts of text, so a language model of English will recognizer
pretty much any "sensible" phrase or sentence from a matching dictionary, like
"today I saw a blue car". The language models that come with Sphinx, like
hub4.5000.DMP are built from transcribed news broadcasts, and, believe it or
not, work pretty well for casual conversation. But they won't recognize highly
task dependent phrases that wouldn't occur in the news. Even if the words
"excise", "lateral", and "hematoma" are in the dictionary, the sequence
"excise lateral hematoma" is probably not covered by the language model.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello guys, I am new to CMU Sphinx. I am using this api in Java language. When
I was trying out HelloWorld program I had found out that .gram file (grammar
file) is important for running the program and the program will only show
words that are in grammar file. I had a question in my mind that is there any
way by which I can print values which are coming from dictionary file while I
am speaking something without the dependence of grammar file.
Sphinx can either use "grammars" or "language models". A language model is a
set of rules for what words can follow other words that comes from analyzing
enormous amounts of text, so a language model of English will recognizer
pretty much any "sensible" phrase or sentence from a matching dictionary, like
"today I saw a blue car". The language models that come with Sphinx, like
hub4.5000.DMP are built from transcribed news broadcasts, and, believe it or
not, work pretty well for casual conversation. But they won't recognize highly
task dependent phrases that wouldn't occur in the news. Even if the words
"excise", "lateral", and "hematoma" are in the dictionary, the sequence
"excise lateral hematoma" is probably not covered by the language model.