We’ve all seen video mashups – where many people are signing the same song, and their video is edited to make it look as if they are all signing along at the same time. There are hundreds of examples on YouTube, and they very often go viral.
This one is on the front page of YouTube today:
- Now, you might be wondering how you might make one of these! Maybe you don’t have any video editing software Actually, once you’ve created the recording, it is not terribly hard to create the sing along effect.
You’ll need a number of videos singing along to the same clip. Ideally, they will all start at the same moment (this will make stitching it all together a lot easier). In my examples, I’ve used a version of this Glitch to create celebrities ‘reciting’ a line from Dr. Seuss’ “Green Eggs and Ham.” It’s as terrifying as it sounds:
[Taylor Swift “reciting” Green Eggs and Ham]
I created a bunch of these, and I want to stitch them all together, so each line jumps to another celebrity. We can do this with Cloudinary, using start offset and end offset parameters to add and remove clips.
Establishing timings for each Clip
How do we know when to switch between clips? We can watch the video carefully and apply timings based on the counter on the video player. In addition, I used the transcript add-on – when I uploaded one of the videos, I asked for the Google transcription service to be run (i describe this process in my post on video captions). The output from this service is a caption (which we can add to the final video), but also a transcript with timings for each word:
From the text above, I know the first sentence ends at 4.0s, and the second sentence begins at 4.2s. Let’s add Elon Musk into the game:
I establish the main video to be 1200 pixels wide (this turns out to be 1480 pixels high). I then add a video overlay (l_video) of the “greeneggs_elon” video, with a start offset of 4.1 seconds. I then apply the layer to start at 4.1 and set the end offset to 6.6s). The last parameter is ac_none. This mutes the overlay – because it is the same audio on the original video.
When I apply this filter – it works…almost: the Elon Musk video is a bit smaller, so we see the Taylor Swift video peeking out of the top and bottom. Adding a cropping parameter fixes this issue:
Awesome, right? Now I can keep going, using the timings from the transcripts and adding new videos at the appropriate points.
So, I added one speaker per ‘scene’ for most of the video, copying the format above. For the end lines of Green Eggs and Ham, I wanted all the celebrities on the screen. This leads to a pretty long url, but I resize each video to 600×600, and place them in different parts of the video. For example, I place Snoop Dogg’s video with gravity south west – so he’s on the bottom left.
Doing this for each celebrity allows me to easily create a layout with each person appearing on the screen:
Now, putting it all together, we can create a complete video mashup of celebrities reciting Dr. Seuss:
The url, as you might expect, is pretty insane, and the pink is just the last 9 seconds with all 7 celebrities on the screen at once.
Video editing with Cloudinary using just the url to edit, crop and splice videos together is possible, and not terribly difficult. You can splice together your own videos – either song mashups, conference talks (slides next to speaker video), all with Cloudinary.
Have you always wanted to mash up a video with different sources and splice them together, but you don’t have the tools to do it? You can use Cloudinary and do all of the editing online. This is the 4th post in LivinLaVideoLoca series of posts, so please enter your video in our contest and tweet your video with the hastag #LivinLaVideoLoca.