Recording Transcriptions

Artigo
07/05/2006

Use the Recording Tool in Speech Prompt Editor to record prompt transcriptions as .wav files and store them in a prompt database.

Recording .wav Files

Before recording transcriptions, enter transcriptions and extractions in the Recording and Extraction panes. This process is described in Entering Transcriptions. By default, Speech Prompt Editor does not display the Recording Tool.

To open the Recording Tool

Open Speech Prompt Editor and add transcriptions.
On the Speech Prompt Editor toolbar, click Record All.
- or -
On the Prompt menu, click Record All.

The Recording Tool displays in a pane at the bottom of the Visual Studio .NET 2003 environment. It shares this space with the Prompt Validation tool and the Visual Studio .NET 2003 Output window. Click a tab to select which window displays.

To record transcriptions

Open the Recording Tool.
Click Record.
Speak the transcription into the microphone, then click Stop.
Click Play to review the recording, click Record to re-record the transcription, or click Next to select the next unrecorded transcription for recording.

Speech Prompt Editor automatically generates alignments when it records the file. If automatic alignment generation is not needed, in Speech Prompt Editor click the Tools menu and then choose Options. Expand Speech Application SDK and click Prompt Editor. If automatic alignment fails, or if alignment is disabled, Wave Editor displays a series of equally spaced alignments and associated labels when the .wav file is opened.

To re-record transcriptions

Open an existing prompt database in Speech Prompt Editor.
On the Speech Prompt Editor toolbar, click Record All.
Click the appropriate transcription in the Recording window, and then on the Recording Tool, click Record.
Speak the transcription into the microphone, and then click Stop.

Changing Recording Options

The Options dialog box for Speech Prompt Editor contains important settings that control how the Recording Tool records .wav files.

To open the Options dialog box for Speech Prompt Editor

Open a prompt project.
On the Tools menu, click Options.
In the tree view pane on the left, expand the Speech Application SDK item, and then click Prompt Editor.

Has Wave Icons

One of these three icons appears in the Has Wave column for each transcription.

Icon	Description
	The red x icon indicates there is no .wav file in the database for this transcription, and alignments do not exist.
	The blue icon with asterisk indicates that an unsaved .wav file exists in the database.
	The blue wave icon indicates that a saved .wav file exists in the database.

Importing .wav Files

Importing individual .wav files is a simple operation. Open a file browser and choose a file to import for the selected transcription.

Note Only transcriptions which do not already have a recording are imported. If a .wav file is in an unsupported format, it is not imported. A message appears and points to the log for details. Speech Prompt Editor automatically generates alignments when it imports a file. If automatic alignment generation is not needed, in Speech Prompt Editor click the Tools menu and then choose Options. Expand Speech Application SDK and click Prompt Editor. If automatic alignment fails, or if alignment is disabled, Wave Editor displays a series of equally spaced alignments and associated labels when the .wav file is opened.

Note Stereo .wav files are converted to monaural when imported.

To import individual files

In the Speech Prompt Editor, click Record All on the Prompt Editor Toolbar to open the Recording Tool.
In the Recording Tool, click Import....

To import a folder

In the Speech Prompt Editor, click Import Script on the Prompt Editor Toolbar.
In the Import Recordings dialog box, select Import wave data and browse to the folder containing the files.

When Speech Prompt Editor imports a set of .wav files, it attempts to match the file names with the Wave property in the Recording window. The following actions may occur:

If file names match, the alignment process continues.
If a match fails, the .wav file is not imported. A message appears after the import operation if there are transcriptions that did not match .wav files.

Understanding Alignments

Speech Prompt Editor associates .wav files with transcriptions in a process called alignment. Alignment matches words in the transcription to words in the .wav file. During alignment, the prompt engine marks the end of each word in the .wav file.

Note If the speech recognition engine cannot form alignments, then a red x icon appears in the Has Alignments column. Add alignments manually to match the words in the transcription, or re-record the transcription and import it again.

Mix and match words marked as extractions in the Recording window with other extractions to build new prompts.

To view alignments in Wave Editor

Double-click either the Has Wave or Has Alignments icon for any transcription with a .wav file.

To return to the Recording window

Close Wave Editor, or in Solution Explorer double-click the prompt database.

To improve the sound quality of a new prompt

Edit word alignments as described in Tuning Alignments and Extractions.

Speech Prompt Editor re-creates alignments every time a user edits transcription text. Brackets, braces, and tag text are not considered transcription text for this purpose. Changes to markup and tags do not trigger a re-creation of alignments. Adding or deleting any characters that are neither inside braces, nor are actually braces or brackets, causes a message to appear indicating that text has been changed and alignments must be re-created; Speech Prompt Editor then re-creates alignments.

Partilhar via