• State Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 1 answer
  • Subscribers 101 subscribers
  • Views 3687 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TMDX5535EZDSP: I want to know how to create a new dictionary engine using "_Fil" ?

Abhinandan BR
Intellectual 945 points

Part Number: TMDX5535EZDSP

Hi,

I want to create a dictionary for _Fil. I want know how to create a new dictionary engine using  "_Fil" ?

The already generated speech dictonary engine model by TI uses this _Fil. But I'm not able to understand how it is created.

Please help.

Regards, 

  • 0
    TI__Guru 65405 points
    Hi Abhinandan,

    I've forwarded this to the c55x software experts. Their feedback should be posted here.

    BR
    Tsvetolin Shulev
  • 0
    TI__Mastermind 30365 points

    Abhinandan,

    I have forwarded this question to someone with better knowledge of the TIesr inner workings, and how the _Fil model was created.

    Lali

  • 0 in reply to Lalindra Jayatilleke
    Intellectual 945 points

    Hi,

    Any updates on this ?

  • 0 in reply to Abhinandan BR
    TI__Prodigy 745 points

    Abhinandan and Lali,

    The C55 demo reference design uses the TIesr recognizer with a dictionary and model set that includes the special word "_FILL" which is defined as pronounced by using a special _fil "phone model".  The "_FILL" word can be used in developing a user's specific TIesr grammar network text file for recognition given the model and dictionary set supplied with the demo. The documentation describes how to create a TIesr grammar network text file.

    To create a new _fil "phone model", for example for another language, requires detailed knowledge of the TIesr recognizer and support tools. One must create a new model and dictionary set with support for a _fil model, and train the _fil model over a large speech corpus. Creating such sets are described in the documentation and examples that accompany the open source TIesr recognizer ( https://gforge.ti.com/gf/project/tiesr ). The exact training procedure and form of the _fil model must be defined by the user. Implementing such training assumes that the TIesr user is familiar with automatic speech recognition processing, and development of such systems.

  • 0 in reply to Lorin Netsch
    Intellectual 945 points

    Hi Lorin,

    Thanks for the reply. As per my understanding _fil is modelled just like an other alphabet or phone. So I should consider _fil as a word and a phone in my dictionary and model it with AUDIO recordings for it. Am I correct?

  • 0 in reply to Abhinandan BR
    TI__Prodigy 745 points
    Hi Abhinandan,

    There are a large number of ways to create a filler/garbage model capability. One way that I did this was to add a word, "_Fill" to the dictionary with the pronunciation using a single "_fil" phone. I modeled the _fil phone as an HMM with a single emitting state, and a number of Gaussian mixture components (hence a Gaussian mixture model). To train the _fil phone, I did a special training in HTK where in the training master label file I replaced all phone models with _fil, and then ran several passes of training. The idea was that this would train the _fil model to represent the broad acoustic classes of all speech in the database. Then to perform recognition in TIesr I could place the _Fill word in the grammar network wherever I wanted to perform out-of-vocabulary discrimination.

    Regards,
    Lorin
  • 0 in reply to Lorin Netsch
    Intellectual 945 points

    Hi Lorin,

    I completely get it now. That s a very detailed answer. Thanks a lot. And could you tell me whats the accuracy % for tests conducted using TIESR?

  • 0 in reply to Lorin Netsch
    Prodigy 175 points

    Hi Lorin,

    I want to creat a new _fil "phone model" for Chinnese and i do not know how to do this.Following your suggestion,i went to https://gforge.ti.com/gf/project/tiesr  for help.But after creating an account i still did not have access to TI Embedded Speech Recognizer documents.

    How can i get the documents and examples?Must i join the project?

    Also i submitted an application to join the project,but did not get a response.What should i do to get the documents and examples?

    Best regards,

    Eric

  • 0 in reply to eric li
    Intellectual 945 points

    Hi Eric,
    Even I tried the same. Even I dont have access to those files. I think the documents you're looking for are in this link:
    www.ti.com/.../TIDEP0066
    Download TIDEP0066 Software ZIP package at the end of the page. It contains various documentation. Hope this is what you're looking for.

  • 0 in reply to Abhinandan BR
    Prodigy 175 points

    Hi Abhinandan，

    Thank to remind me.

    I just want to find the ducocuments that could guide me  how to creat a new _fil "phone model" for Chinnese.

    Anyway i will check the TIDEP0066 and hope to find something useful for me.

    Regards,

    Eric

  • 0 in reply to eric li
    Intellectual 945 points

    Hi Eric,
    As for the documentation for _fil model creation, there is none. You just need to follow the tips and leads given by Lorin in this post thread.