Dive has two types of read-aloud features. Each of them also supports multiple languages (read-aloud audio supports 65 languages, and other languages are read aloud in English. For the list of supported languages, see Languages supported by automatic translation).
1. Read text
When you open a step, the Text pane is read aloud. In addition to the body text, Tips / Quality warnings / Safety-critical items are also read aloud (you can change which items are included in the "Note: Items included in Read text" section below). For steps other than video, Read text is turned on by default.
Features: Not synced with the video, but no timing adjustment is needed, so it is easy to edit. It can also be used for content other than video.
If the read-aloud is longer than the step's video, the video repeats until the read-aloud finishes. To learn how to stop the repetition, see The same step video repeats many times in an exported MP4.
2. Read subtitles
Subtitles are read aloud in sync with when they appear. Only the subtitles are read aloud. Tips / Quality warnings / Safety-critical items are not read aloud. When you create an SOP from a video, choose the narration target in "Step conversion settings" during creation. The default is "Subtitles" (your last selection is kept).
Features: Synced with the video, so the read-aloud content is easy to understand.
Choose the narration target
For each step, choose one of "None," "Description," or "Subtitles" as the narration target. You cannot read both subtitles and descriptions in the same step. If you want to read the tips and cautions aloud, choose "Description" for that step. In this case, subtitles are shown on screen but not read aloud, and the body text, tips, and cautions are read aloud.
On the SOP edit screen, set it up as follows.
- Video steps: Press the speaker button that appears at the top when you select a video step (it shows the current setting: "Subtitles," "Description," or "None"). This opens [Common audio settings], and you can change the narration target for all steps made from the same video at once.
- To change only one video step: In that step's [Audio] section, turn on "Define individually," then choose the narration target.
- Steps other than video: In that step's [Audio] section, turn "Read text" on or off.
- To change multiple steps at once: In the step list, hold down Ctrl (command on Mac) or Shift while selecting steps, then press [Bulk edit] on the right. On the [Audio] tab, choose the narration target and apply it. (If you open [Bulk edit] from the [Other] menu, all steps are targeted.)
Note: About the reading speed of subtitles
For subtitle read-aloud, the reading speed (tempo) is adjusted automatically based on how long each subtitle is displayed (its frame) while editing. When the audio does not fit in the frame, the speed goes up and the read-aloud becomes fast.
To keep the read-aloud from being too fast, extra time is secured in the following order.
- If there is a gap between a subtitle and the next one, the reading uses the time up until the next subtitle starts.
- If there is a shape freeze in between, its freeze time is also used for reading aloud (the reading keeps going during the freeze without stopping).
- If it still does not fit, the audio is generated slightly faster at the time it is created (the sound is less distorted than when the audio is sped up afterward).
You can check which subtitles are read too fast in the Subtitle list of the video editor. Subtitles that do not fit in their frame show "Fast ○x" (this is the speed when the freeze time of shapes is also used for reading aloud).
If you want a more natural speed, split the sentence into shorter parts, or leave more space before the next subtitle (for the last subtitle of a step, the end of the subtitle is extended). For details, see For short steps, subtitle read-aloud is too fast to follow.
Note: Volume of read-aloud audio
Subtitle read-aloud is made by creating the audio one sentence at a time and joining them together. Because the volume of each sentence is matched before joining, the volume does not change from sentence to sentence within the same SOP. The volume is almost the same across SOPs, and even when you switch to a translated language.
Note: Items included in Read text
With Read text, tips, quality warnings, and safety items are read aloud in addition to the body text. If the read-aloud gets long, you can choose whether to include each item in [Items to read aloud] under "Playback & display settings" > "Advanced display settings" in [Basic information] on the SOP edit screen.
The body text is always read aloud. Items you turn off are left out of the audio, but the display on screen does not change. When you switch, the read-aloud audio is regenerated.
This setting only applies to steps whose narration target is "Description." Steps whose narration target is "Subtitles" read only the subtitles, so switching this setting does not change them. If you want to read the tips and cautions aloud, set that step's narration target to "Description" using the method described in "Choose the narration target" above.
Note: Language used as the basis for read-aloud
Translation and read-aloud are based on the Original language set for the SOP. The default is Japanese.
For SOPs created in a language other than Japanese, set the Original language in "Playback & display settings" in [Basic information] on the SOP edit screen. If you play the SOP as it is, without setting this or translating it, the read-aloud audio may not be generated correctly.
The team-wide default can be set with "Language of SOPs you create from now on" in the [Terminology Glossary] of the dictionary (Team Administrator or higher). It does not affect SOPs that have already been created.
Note: When read-aloud audio is regenerated
Read-aloud audio is regenerated by the operations listed below. If the read-aloud specification changes, existing SOPs will reflect it from the time their audio is regenerated.
- When you rewrite and save subtitles or text
- When you add, delete, or change the length of a shape freeze within the subtitle read-aloud range
- When you change the Original language
- When you register a word in the Pronunciation Dictionary (it may take a few minutes for existing audio to reflect the change)
- When you run Bulk replace on the text
- When you change the Items to read aloud
The behavior of using the shape freeze time for read-aloud also applies to SOPs that have already been created. For steps with a freeze within the subtitle read-aloud range, the read-aloud audio is regenerated automatically the next time you open the SOP online.