Autores Portugueses

1
Title

Author DETUA - Universidade de Aveiro


LEET Processamento Digital de Voz
2
Scenario

• Present the scenario that you choosed
• Consider using figures to easier understanding

3
Grammar

• Include your gram file
• And comments if relevant
• What were the simplifications ? ...

4
Word list

• Present word list
• And information about them
• How many?
• Minimum length, max lenght ...

5
Diccionary

• Your final version of beep or dict

6
Monophones list

• Your monophones0 or 1

7
Recording

• How you get the prompts list for training and
  test
• How many sentences for traing and test
• How you created the recording batch

8
Recording (part 2)

• Include the script you used for doing the
  recording

9
Recording conditions

• Who was the speaker ?
• Info about sound board and microphone used

10
System training

• Present (briefly) all the steps you performed to
  train the model
• Include your batch files an explanations about
  what they do
• You can include each batch in a differente slide
  ...

11
Results

• Present the results of testing with the test set
  (see section 3.4.1 of the tutorial
• (if done) present info about using the recognizer
  in live mode (section 3.5)
• Seems to work ? It works with other person (not
  the one used for training) ? ...
• Include a screen dump showing it running ...
• NOTE you will show me the system after ?

12
Discussion

• What are your impressions from the work?
• What was more difficult?
• What didnt work?
• Was interesting ?