Hi
I have been using sphinx to create isolated word speech recogniton systems using phone models for which the dictionary looks like this:
asia ee s i a
Now i want an isolated word speech recognition system using word models for which the dictionary should look like this:
asia asia
but as you can see the "phones" i.e. the word models i use do not have any context(like the ones which occur in triphone models).
How do i configure sphinx to ignore building of context dependent models since the train process gives the following error:
MODULE: 30 Training Context Dependent models (2009-03-20 13:21)
Phase 1: Cleaning up directories:
accumulator... logs... qmanager... completed
Phase 2: Initialization
mk_mdef_gen Log File
WARNING: This step had 0 ERROR messages and 1 WARNING messages. Please check the log file for details.
completed
init_mixw Log File
FATAL_ERROR: "main.c", line 334: src mean n_cb (== 459) inconsistent w/ ts2cb mapping 441
FAILED
Please help me regarding this.
Regards
krishadhithya
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
It isn't the optimal dictionary for isolated words recognition task. The optimal one was used in tigits for example:
eight EY_eight T_eight
five F_five AY_five V_five
four F_four OW_four R_four
nine N_nine AY_nine N_nine_2
oh OW_oh
one W_one AX_one N_one
seven S_seven EH_seven V_seven E_seven N_seven
six S_six I_six K_six S_six_2
three TH_three R_three II_three
two T_two OO_two
zero Z_zero II_zero R_zero OW_zero
> FATAL_ERROR: "main.c", line 334: src mean n_cb (== 459) inconsistent w/ ts2cb mapping 441
FAILED
It means you have duplicate phones in your phoneset
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi
I have been using sphinx to create isolated word speech recogniton systems using phone models for which the dictionary looks like this:
asia ee s i a
Now i want an isolated word speech recognition system using word models for which the dictionary should look like this:
asia asia
but as you can see the "phones" i.e. the word models i use do not have any context(like the ones which occur in triphone models).
How do i configure sphinx to ignore building of context dependent models since the train process gives the following error:
MODULE: 30 Training Context Dependent models (2009-03-20 13:21)
Phase 1: Cleaning up directories:
accumulator... logs... qmanager... completed
Phase 2: Initialization
mk_mdef_gen Log File
WARNING: This step had 0 ERROR messages and 1 WARNING messages. Please check the log file for details.
completed
init_mixw Log File
FATAL_ERROR: "main.c", line 334: src mean n_cb (== 459) inconsistent w/ ts2cb mapping 441
FAILED
Please help me regarding this.
Regards
krishadhithya
i dont think you understood my question.
The error came when the dictionary was like this
asia asia
not when it was
asia ee s i a
can u help me out with that
>asia ee s i a
It isn't the optimal dictionary for isolated words recognition task. The optimal one was used in tigits for example:
eight EY_eight T_eight
five F_five AY_five V_five
four F_four OW_four R_four
nine N_nine AY_nine N_nine_2
oh OW_oh
one W_one AX_one N_one
seven S_seven EH_seven V_seven E_seven N_seven
six S_six I_six K_six S_six_2
three TH_three R_three II_three
two T_two OO_two
zero Z_zero II_zero R_zero OW_zero
> FATAL_ERROR: "main.c", line 334: src mean n_cb (== 459) inconsistent w/ ts2cb mapping 441
FAILED
It means you have duplicate phones in your phoneset