RE: Narration and transcription

Hello TT folks,

-- Charles Wiltgen asks:
Do these have unique requirements?

George replies:
On the Timed Text side of things, I think that they are not unique,. but
I should be explicit.

Timed Text is required to support:
--Word level synchronization
--Identification of text by reference that is independent of the timing
information.

For example, the full text of a book should be able to be used in the
SMIL presentation. Both the audio and the text are synchronized at  a
level as granular as by word. It is expected this would be done by a XML
element ID, but there may need to be something else. I would not expect
that the full text would need to have each word marked up in the text.

Here are two instances that conforms to the DAISY 2.02 specifications,
which is an application of SMIL 1 and XHTML. The SMIL ref points to text
contained in an ID referenced element. This allows for display in
context or just the current text.:
Example 1 is a paragraph of text and a series of audio clips that is the
narration of the text.

        <par endsync="last" id="rgn_par_0005_0003">
          <text src="daisynews2.html#rgn_cnt_0049"
id="rgn_txt_0005_0003"/>
          <seq>
            <audio src="dtb_0005.mp3" clip-begin="npt=55.105s"
clip-end="npt=70.364s" id="rgn_aud_0005_0007"/>
            <audio src="dtb_0005.mp3" clip-begin="npt=70.364s"
clip-end="npt=86.495s" id="rgn_aud_0005_0008"/>
           </seq>
         </par>

Example 2 shows a single text item and its associated single audio clip.

            <par endsync="last">
                <text src="klee2.html#badw_000b" id="badw_000b"/>
                <seq>
                    <audio src="badw000a.mp3" clip-begin="npt=0.000s"
clip-end="npt=2.599s" id="phrs_0001"/>
                </seq>
            </par>

Best
George

Received on Sunday, 9 February 2003 13:40:28 UTC