Greg | 5 March 2025

The Audio Project Manager development team is pleased to offer the 4.1 release of the Audio Project Manager. This version updates the dependencies to the latest libraries and adds three AI experimental features: noise reduction, voice conversion, and speech recognition. These features are enabled when setting up the team or on the Team settings dialog.

The first two are available on the record step and the speech recognition is available on the transcriber tool (both for vernacular and for back translations). For the noise reduce and voice change, it is possible to select a portion of the wave form and then the noise reduce or voice change is only applied to the selection.

For the speech recognize, if the verses have been marked in a previous step, then the transcriptions are done within the verse. One verse shows up at a time. This will sync to Paratext with the verse markup. If in the future, a change is made to the recording, the verses would need to be marked again. But you can then pull the previous transcription from Paratext, remove the verse that has been revised and the speech recognize will add the transcription for the missing verse (or verses) to the end for the transcriber to cut and paste where they would like them in the transcription. Here is a video where Nathan demonstrates these AI features.

We also added a copy sheet function back which we used to have but was removed when the performance of large sheets was improved. This is helpful for setting up another project using the sample passages (for example when in a cluster). We fixed an issue with the recorder where the microphone started out very “hot” (high volume) and became quieter as the recording went along. We also made improvements for publishing stories to Akuo.

Here is a detail list of the changes:

  • Add SplitButton for ASR

  • Add ai for noise removal, voice change, and transcribe (#1815)

  • Add logic to choose a related language

  • Avoid network error for large noise cancel files (#1865)

  • Call Transcription via mediafile. Find task id in segments (#1836)

  • Don’t show ASR if offline

  • TT-6125 change audio capture options

  • TT-6135, TT-6168, TT-6240 Handle desktop exit better

  • TT-6150 Copy selected rows from sheet (#1851)

  • TT-6208 Add a way to show general projects as stories in Akuo (#1811)

  • TT-6216 No alert for download

  • TT-6226 set min width for compare select

  • TT-6229 Show note shared resource title instead of reference (#1819)

  • TT-6241 set default progression (#1828)

  • TT-6243 Add Settings button to personal projects (#1859)

  • TT-6244 Handle large audio for AI (#1826)

  • TT-6249-6253 Only recognize speech on transcriber steps when transcribe role is transcribe

  • TT-6253 configured button

  • TT-6260 no special chars in age or gender

  • TT-6265 update rights collection (#1831)

  • TT-6269 Return to same page on reload

  • TT-6274 Allow user to do other things while transcribing (#1840)

  • TT-6294 force not override existing transcription (#1854)

  • TT-6297 show blue dot immediately on close of progress (#1843)

  • TT-6303 Ask about propagating to section and movement (#1844)

  • TT-6304 Check for Online before AI (#1849)

  • TT-6307-6150 Create a Copy to Clipboard menu item (#1856)

  • TT-6307-6150b use multi row selection (#1860)

  • TT-6309 check bibleid and bibleName before Yes enabled (#1857)

  • TT-6314 check if needs bibleid (#1862)

  • TT-6320 no offline ai features

  • TT-6322 No duplicate name (#1864)

  • Update to build on Mac

  • Update voice permission statement

  • add publishing explicitly

  • AI adding work for hire

  • feature flags

  • mobile team screen

  • obthelps required bible (#1814)

  • offer romanize for other non roman scripts

  • project menu and personal choice

  • return multiple tasks if timing available (#1841)

  • treat Chinese special case