DEV Community

Cover image for Annotating Video Clips With MoviePy
Dennis O'Keeffe
Dennis O'Keeffe

Posted on • Originally published at blog.dennisokeeffe.com

Annotating Video Clips With MoviePy

I write content for AWS, Kubernetes, Python, JavaScript and more. To view all the latest content, be sure to visit my blog and subscribe to my newsletter. Follow me on Twitter.

This is Day 12 of the #100DaysOfPython challenge.

This post will use the MoviePy library to annotate some captions onto a short video I recorded about my site workingoutloud.dev.

Prerequisites

  1. Familiarity with Pipenv. See here for my post on Pipenv.

Getting started

Let's create the hello-moviepy directory and install Pillow.

# Make the `hello-moviepy` directory
$ mkdir hello-moviepy
$ cd hello-moviepy
# We will write everything in main.py
$ touch main.py
# Add this stage, you will need to add you video clip - I added mine to the `hello-moviepy` directory
$ cp path/to/wol-dev.mp4 .

# Init the virtual environment
$ pipenv --three
$ pipenv install moviepy
Enter fullscreen mode Exit fullscreen mode

At this stage, we can write our script to annotate the video.

Annotating the video

As an example, here is a Gif of my video that I am editing:

Original video

The video is a short clip of my workingoutloud.dev site. It is around 9 seconds long and I want to add some text to it.

From watching the video, there are three things that are essentially in the demo:

  1. Viewing tasks and subtasks.
  2. Sorting subtasks.
  3. Opening the current goals.

These occur at the 0:00, 0:03 and 0:06 second mark respectively.

We can add in some captions for each of these easily thanks to MoviePy.

In main.py, add the following:

from moviepy.editor import VideoFileClip, TextClip, CompositeVideoClip

video = VideoFileClip("wol-dev.mp4")
Enter fullscreen mode Exit fullscreen mode

In the above code, we simply import the required modules and then we create a VideoFileClip object from the video file (relative to the project root directory).

We now want to create a TextClip object for each of the three points in the video.

According to the MoviePy TextClip docs, we can see the TextClip constructor can take some relevant arguments around the text, font, color and font size.

The TextClip instance itself then has access to methods to set the position, duration and start time. We will abstract our sensible defaults into a function and implement this:

def text_clip(text: str, duration: int, start_time: int = 0):
    """Return a description string on the bottom-left of the video

    Args:
                text (str): Text to show
                duration (int): Duration of clip
                start_time (int, optional): Time in video to start at (in seconds). Defaults to 0.

    Returns:
                moviepy.editor.TextClip: A instance of a TextClip
    """
    return (TextClip(text, font="Arial", fontsize=24, color='black')
            .set_position((20, video.h - 44))
            .set_duration(duration)
            .set_start(start_time))


# Make the text. Many more options are available.
text_clip_one = text_clip("Create tasks and subtasks", 3)
text_clip_two = text_clip("Sort subtasks by table header", 3, 3)
text_clip_three = text_clip("View goals", 3, 6)
Enter fullscreen mode Exit fullscreen mode

After writing out the helper method, we are creating three TextClip instances for each of our different annotations that we wish to add to the video.

Finally, we can compose them all together into a CompositeVideoClip object and write out the video file.

I am also writing out a .gif file to use as a preview of the video on this post.

# Overlay text on video
result = CompositeVideoClip(
    [video, text_clip_one, text_clip_two, text_clip_three])
result.write_videofile("wol_dev_edited.mp4", fps=25)
result.write_gif("wol_dev_edited.gif", fps=8)
Enter fullscreen mode Exit fullscreen mode

We can now run our script with python main.py. There will be some progress bars to indicate the progress.

After the script has completed, we can now see wol_dev_edited.mp4 and wol_dev_edited.gif in the hello-moviepy directory.

The outcome looks like so:

Annotated video

Summary

Today's post demonstrated how to use the MoviePy package to add some text captions to a video programmatically.

I will be hoping to make use of this to help annotate some of my videos that I am working through to save some time in future.

The final code looks like so:

from moviepy.editor import VideoFileClip, TextClip, CompositeVideoClip

video = VideoFileClip("wol-dev.mp4")


def text_clip(text: str, duration: int, start_time: int = 0):
    """Return a description string on the bottom-left of the video

    Args:
                text (str): Text to show
                duration (int): Duration of clip
                start_time (int, optional): Time in video to start at (in seconds). Defaults to 0.

    Returns:
                moviepy.editor.TextClip: A instance of a TextClip
    """
    return (TextClip(text, font="Arial", fontsize=24, color='black')
            .set_position((20, video.h - 44))
            .set_duration(duration)
            .set_start(start_time))


# Make the text. Many more options are available.
text_clip_one = text_clip("Create tasks and subtasks", 3)
text_clip_two = text_clip("Sort subtasks by table header", 3, 3)
text_clip_three = text_clip("View goals", 3, 6)

# Overlay text on video
result = CompositeVideoClip(
    [video, text_clip_one, text_clip_two, text_clip_three])
result.write_videofile("wol_dev_edited.mp4", fps=25)
result.write_gif("wol_dev_edited.gif", fps=8)
Enter fullscreen mode Exit fullscreen mode

Resources and further reading

  1. The ABCs of Pipenv
  2. MoviePy TextClip docs
  3. Pipenv
  4. MoviePy
  5. GitHub - project code

Photo credit: element5digital

Originally posted on my blog. To see new posts without delay, read the posts there and subscribe to my newsletter.

Top comments (0)