I'm not sure exactly how, but I ended up with Matt, our CEO of MakerX, pinging me via Slack to see if I had time to whip up a 15-second promo video (viewable at the end of this post) for our new venture QRX; a fantastic new AI-powered product that MakerX is making. Given the video I recently published on the company's TikTok, he knew I had some lousy video production skills. Still, I figure he also knows I'm a pretty resourceful kind of person. So, I looked for the best tools to help me achieve this. Because my objective is a low-effort and low-budget promo video for a trial run as an ad on YouTube, I knew I didn't need to seek help from people more skilled than what AI and I could do together. So, for this little project, I could justifiably lean heavily on AI to fill my skill gap.
Before making a promo video, I had to understand the message I wanted to convey to the audience. Luckily for me, 'What is QRX?' is pretty straightforward. Meaning, I wanted to convey that with the power of AI and a few shakes of the Makers' magic jars, anyone can now produce a very cool-looking QR Code that people can scan. Gone are the boring old black-and-white ones! The future is colour! And why stop with colour when you can embed a QR code in an AI-generated image? This is the transition from black and white TVs to big colour ones!!! Divided by about 25 or so. Ah ehm. Sorry, I was getting away from myself. You may have guessed, but I am a little excited by it.
For example, you can scan a QR code to open this blog post. Feel free to share it 😉😉...
Promo Video, time
For the script, I used Chat GPT by OpenAI, and I gave it the following prompt:
Create a dialogue for a promo video that will run for 15 seconds that highlights QR codes made by AI that pop
After some human-powered curation of the Chat GPT output, I settled with the following mostly AI-generated dialogue script:
Introducing the future of QR Codes! Watch them pop to life and capture attention like never before! Crafted by AI and designed to captivate! Experience the QR Code revolution today!
Next up, I needed to convert the script into speech. Stafford, a colleague, suggested Text to Speech by ElevenLabs, and after a little play, I was happy to use it. I found their product easy to use and could produce a voice sample for my video without fuss.
With the script and voice-over in hand, next up were the visuals. I knew this bit would be the hardest, but I remember that Runway had recently released a text-to-video feature in their toolkit with a prompt that could take an image as the input. I had doubts about what it could do with the images that QRX produces, as QR Code encoding information can appear to resemble noise. But I gave it a go, and to my surprise.... 🥁
It worked better than I ever expected.
With the audio and knowledge of how I would make the video elements out of the way, the only thing left was putting together all the bits. Thankfully, with the help of Davinci Resolve, that bit was reasonably straightforward.. shout out to the people who make the Davinci Resolve YouTube tutorials.
Without further ado, I proudly present a promo video that I would have never been able to put together before AI.