The jconf file has notations in it which are probably really useful but as a new user to voice recognition most of the terminology is beyond my grasp.

Is there a good breakdown where the options are discussed in greater depth that someone could point me to?

I'd like to get Julius to return longer strings, and there are some options in there that seem like they may be able to help with background noise and map unknown words to a common word. That would be amazing as I could then use a null word like "foo" and filter the "foos" out with the API I've written.

Thank you, Tom J

--- (Edited on 11/21/2020 12:24 pm [GMT-0600] by Tom J.) ---

I've read through everything and totally destroyed my jconf with changes twice now. The answer should be in here:

https://julius.osdn.jp/juliusbook/en/desc_options.html#desc_opt_global_misc

.... But I cannot find it. Can anybody offer some clues?

--- (Edited on 11/22/2020 4:45 pm [GMT-0600] by Tom J.) ---

>I'd like to get Julius to return longer strings

Julius uses two approaches to recognition: Grammar based and a Statistical Language Model. 

Grammar files can be designed to return as long a string as you want.

ken

--- (Edited on 12/1/2020 10:47 am [GMT-0500] by kmaclean) ---