When do I need to edit?
The subtitles can be used directly, but in some cases you may want to adjust them in the editor, for example to improve readability or timing.
Download multiple results as a package
- Go to My Files.
- Select multiple files of the same type in the list (Subtitles or Transcript).
- If any of the files you select are not yet transcribed, they will not be included in the export.
- If you select mixed types (both Subtitles and Transcript), the feature is not available.
- Click the green button labeled EXPORT above the list of files.
A dialog box opens. Depending on whether you selected Subtitles or Transcript, you will get different Export settings.
The settings apply to all selected files.
When you click Export in the dialog box, the files will be downloaded as a zip file to your computer.
How it works
- Choose the right transcription type
- When you start a transcription, you can choose between two types: Transcript and Subtitles. Which option is best depends on how you intend to use the result.
If you want to continue working with the text as a document, Transcript is usually the right choice. The text is then presented as continuous timestamped paragraphs with speaker changes labeled, which makes it easy to read, edit and use in, for example, analyses, reports or interviews. This format is particularly suitable for research materials or other situations where the content needs to be further processed.
Subtitles are used instead when the text is to follow audio or video. The text is then divided into shorter blocks with time codes, so that it can be displayed in sync with what is being said. This makes the format suitable for lectures, presentations and other material that is to be published with subtitles, for example in Sunet Play, on YouTube or in an LMS.
The key difference , then, is that Transcript is intended for reading and further analysis, while Subtitles is intended for playback along with media.
If you're unsure, it's usually easiest to start with Transcript. You can always create a new transcript in a different format later. Also keep in mind that subtitles sometimes require some adjustment afterwards to get a good rhythm and readability.
Subtitle Principles
Subtitles are designed to be read while watching a video. Therefore, they differ from a regular running text, both in how they are divided and how the content is presented.
The text appears in shorter blocks that remain on the screen for a limited time. To be readable in real time, each block needs to be relatively short and clearly worded. This means that longer sentences are often broken up and that the text is sometimes compressed.
Since subtitles should be read at the same time as the content is playing, they also need to have a consistent and predictable timing. The text must not disappear too quickly, but neither must it remain longer than necessary. This means that subtitles sometimes prioritize readability over accurately rendering each word.
This means that subtitles may look different than a full transcription. The goal is not to reproduce the speech in detail, but to make the content easy to follow in the moment.
Machine-generated subtitles usually work well as a starting point, but may in some cases need to be adjusted in the editor, for example to improve line breaks or timing.
What is verbatim?
In Sunet Scribe, for some languages, you can choose verbatim transcription. When you do this, you get a more literal representation of what is being said. This means that the text not only tries to correctly convey the content, but also things that are otherwise often cleared out, such as filler words such as "eh" and "uh", repetitions and interrupted sentences. The result is then closer to how the speech actually sounds.
However, it is important to understand that verbatim in Sunet Scribe does not imply a completely accurate representation in every detail, and the differences can sometimes be subtle. The transcription is still done by machine, and even in verbatim mode, there is some interpretation. Verbatim preserves more nuances in speech than in a "cleared" text, but does not create a perfect transcript of every pause or sound. The level of detail in verbatim transcriptions can also differ depending on the language model used.
Verbatim is especially useful in contexts where details of language matter. In interviews, research material or conversation analyses, hesitation, retakes and formulations can be significant. In such cases, verbatim gives a more faithful picture of what has been said.
At the same time, verbatim means that the text is often less easy to read. Filler words and repetitions can make the text appear more choppy and less coherent. If the purpose is instead to quickly get a clear and readable text, for example for lectures, meeting notes or publication, it is usually better not to use verbatim. This will automatically remove such things and make the result more suitable for reading.
If you are unsure, it is usually easiest to start without verbatim. You can always create a new transcription with verbatim if you need a more detailed rendition afterwards.
