Dynamic Captions Best Practices

Dynamic Captions are intended to be used on short, simple recordings that are less than 5 minutes in length ands are not intended to replace Captions.

For details on how to caption your video, please see our tutorial that covers captioning, which you can find here.

When using Dynamic Captions in Camtasia 2024 and later, here are some best practices to consider:

Dynamic Captions are tied to the source file not the media clip on the timeline.

• We decided to do this so that users wouldn't have to transcribe/edit multiple copies of captions on the timeline if they used the media multiple times.
  • One common concern from users is that by deleting the caption media, they will lose changes/edits to the transcription. Because we generate off of the source, they can happily delete the caption media and replace it with a different style without losing any of their transcription edits!
    • At the moment, there is no way to “delete” the transcription of an audio file. To accomplish this, a user will need to delete the media from their timeline + Media Bin, re-import it, and then re-transcribe it.
      
      

Dynamic Captions do not transcribe over grouped audio, similar to how Audio Visualizers work.

• This was a design choice to simplify the design, reduce scope, and get the solution out more quickly. Based on user feedback, we will decide if we should evolve that design.
  • Speaking of which, groups are a good solution to having multiple captions at the same time. If you have 2 speakers, and want a dynamic caption show up for each speaker, users will have to:
    • Group speaker 1 audio with dynamic caption 1
    • Group speaker 2 audio with dynamic caption 2
      
      

Dynamic captions will only show the transcription of the "the topmost transcribed media on the timeline".

• So, if you have a dynamic caption on track 3, speaker A on track 2, and speaker B on track 1, the dynamic caption will show the transcription for speaker A (track 2).
  
  

You can think of our AI transcription engine as a person listening to you talk and trying to transcribe what you're saying. If you use a high-quality microphone, limit the noise in your environment, and speak clearly, it will generally do a good job of transcribing what you're saying. As you reduce those factors, its ability to transcribe what you're saying will be negatively impacted.

Transcribing narration can be resource-intensive, depending on different variables.

• Transcription speeds will largely depend on your hardware environment (mostly CPU and memory at the moment).
  • As such, very long narration can result in long transcription times and increased error rates. For best results, try to keep media transcription to 10 minute segments or less. For longer transcriptions, this is a great use-case for moving to Camtasia Audiate plan level.
    • Also note: Because the SOURCE is what gets transcribed, trimming a media on the timeline to make it shorter DOES NOT help; the recording itself has to be shorter.
• For better results, consider recording your narration in chunks (i.e. ten 1-minute recordings vs. one 10-minute recording).
  
  

Long periods of silence on the timeline can sometimes cause our AI transcription model to get confused, and can increase the error rate (especially if your silent sections contain background noise).

Dynamic Captions DO support many different languages! However, our AI transcription model will begin transcribing in whichever language it initially detects. If you have an audio recording that includes multiple languages, you might want to export them as individual clips.

If you find that you need to add/remove words, you might want to stick with a style that doesn't use per-word styling.