The Combine allows you to easily attach audio of pronunciations to words. In the upcoming release (v1.2.0), you can add Speakers to a project, associate a Speaker to an audio recording, and attach a consent file to each project Speaker!
As a consequence of this update, users will not be able to delete audio that was added before the upcoming release. Please contact us if you encounter a situation in the future where you need to delete audio in an older project.
(At this time, due to limitations in transferring audio between The Combine and FLEx, speakers cannot be associated with imported audio nor with audio that was added prior to this release.)
Associating a Speaker to an audio recording sounds good, but will the user have to choose the Speaker each time a recording is made, or will it be possible to associate the same Speaker to multiple recordings with a bulk operation, or set the default Speaker and then that Speaker is automatically associated unless changed?
Great question, @Sara_M! Yes, the user will be able to easily set/change/unset the default speaker to be automatically associated with all new recordings. (After recording, the speaker can be updated one recording at a time.)
What will the pronunciation label look like (in the LIFT file / in FLEx)?
ID first, then last name, then first name? With separators?
Will the pronunciation label be imported into FLEx in the analysis WS that is used in TC? This might not be the top listed WS in FLEx.
At this time, it is added as an English label:
Speaker #<id>: <name>
where <id> will match the filename of the associated consent file (if there is one) and <name> will be the speaker’s name exactly as entered in The Combine.
In the LIFT file, this will look like (e.g.):
<text>Speaker #65739cf35534390b675e7f96: Sra. Maria Lopez</text>
(The consent files are being placed in a
consent subdirectory of the zipped export; they are not explicitly mentioned in LIFT file nor imported into FLEx.)
Another question: on export to LIFT, will the admin be able to choose to export from one particular speaker, or will the LIFT file contain all data from all speakers?
All recordings and associated speaker info will be exported. We don’t have partial exports in view at this time.
Testing showed the following, with regard to the speaker consent files:
- the files were exported without a file extension; thus hard to know which are even audio / image. Can the extension be added?
- the audio consent files are exported as .webm, rather than .wav as per the LIFT export. Could these be .wav too?
Ah, thanks for the feedback! I’ll work on those.