chrislynch.link

Getting Started with Pictory.ai and Text-to-Video

Are you interested in using video to promote your writing but aren’t quite ready to “dance dance dance” like Wednesday Adams? Text-to-video tools can help, so I took pictory.ai out for a try.

What is text-to-video?

Text-to-video tools take a block of text, like a blog post, and generate a video from it.

They do this by using algorithms to pick out keywords in your text and then, using these keywords, to select appropriate stock video and images that they merge together into a video. Most text-to-video tools will also add in music and animated text. Some tools, including Pictory, also provide the option to convert your text into speech using a range of synthesized voices. If you’re confident enough to record your voice then some tools, again including Pictory, also offer the option to transcribe audio back to text.

You’re probably already watching text-to-video videos

TikTok, YouTube Shorts, and Instagram reels are full of text-to-video generated content. You’ll see everything from jokes to recipes, tips and how-tos, and even short stories in the text-to-video format. As the vast majority of video is watched with sound turned off, text-to-video makes a lot of sense as a quick way to create video content without having to record yourself.

Choosing a text-to-video tool

There is a huge range of text-to-video tools available and I’ve been experimenting with many of them over the past few weeks. In this blog post, I’m focussing particularly on my experience with pictory.ai, and I’ll be following up to let you know how successful my videos have been.

Disclosure: My links to pictory.ai include an affiliate code. If you enjoy this content and want to try Pictory for yourself I’d be delighted if you use these. This remains an honest review of the system with both positive and negative points.


Getting Started

Signing up for Pictory was easy, linking through my Google account. You can generate three videos for free, using the full range of features on the platform. Subscriptions are reasonably priced, significantly cheaper than some others on the market, and come with a good number of videos per month even at the lowest tier. However, you should note that Pictory holds back some key features at its entry level tier including 16:9 videos, which are the format you would want for TikTok, Instagram Reels, YouTube Shorts, etc. This effectively makes the lowest tier useless unless you want to reframe and rescale your videos each and every time.

Once you are signed up, you can generate videos in four ways; by entering text directly to Pictory, by pulling text across from a blog post, by uploading an existing video and editing it, or by uploading a set of images.

For this blog post, I’ve tried the first two; text entered directly into Pictory and pulling across a blog post.

Experiment 1: Text Entered Directly to Pictory

The Banana Experiment” is one of my favourite “management parables” that I use in my day job to help teams understand the dangers of herd mentality and sticking to out-of-date “that’s always how we’ve done it” dogma. I’ve actually got no idea if this experiment really happened, but that has no impact on how effective the story is in inspiring people to take a fresh look and old ideas.

I took the text from the image I use on my blog post about The Banana Experiment and free-typed it into Pictory. I picked a base theme and Pictory then generated a series of “scenes” (one at each piece of text) and assigned images or video to them.

Overall, the image and video selection was pretty good early on but became a little more erratic later. One of my scenes, for some reason, featured a man sitting on the toilet. (There definitely wasn’t any toilet-related content in the text!)

Thankfully, it’s easy to select new images and videos and Pictory offers an extensive collection of licensed content for you to choose from. You can edit each scene individually or link scenes together to spread a video across more than one scene.

You can also choose your background music, if you want some, and opt for voice synthesis with a range of voices. Some sound more realistic than others but I was ultimately quite pleased with the warmth and tone of the voice that I used. You can decide for yourself if the video sounds like it is being narrated by a toaster or not!

Overall, it took around 15 minutes to make this video before uploading it to social media. It would have taken significantly longer doing this “by hand” in something like OpenShot.

Verdict: I’m pretty happy with this video. It’s not the most creative thing I’ve ever done but for 2-3 minutes of social media content I think the time required was great, much quicker than I’ve experienced before, and hopefully this will get a good response online.

Experiment 2: Summarising a Blog Article

For my second experiment, I tried out the “blog to video” feature. This feature is designed to take a long-form blog article and convert it into text content.

This was a feature I was very interested in. In The Truth About SEO I talk about the importance of re-purposing content from your blog to social media so the idea that I could write a long-form article and then automatically convert it to content for social media sounded great!

I used my article “Mastodon Servers for Writers” to test this functionality out.

The first thing that Pictory did was pull the full text of the article across. It did this really well, automatically ignoring my blog sidebar, header, and footer. This is probably helped significantly by the fact that I use the Hello template for Elementor, which is well-structured and contains semantic HTML elements to help search engines divide my content from the navigation and promos on my site.

Pictory then proceeded to pick out what it thought were the “key messages” in my content. It automatically selected the headings through the document and a few key sentences. This part of the experience wasn’t quite as good. My headings were mostly phrased as questions and so weren’t useful without other text from the article. Where Pictory had picked out sentences, it didn’t seem to have done so with much context. This could be my writing style or the algorithm or a bit of both, but there’s definitely more work required here to make this truly automatic.

Thankfully, Pictory provides a really nice user interface for selecting sentences and automatically adding these as new scenes to your video. So, with just a few clicks, I’d made a set of scenes that made sense in terms of how I would have summarised the article.

Within a few minutes, I made my video… but was it any good?

In short… no.

Pictory was more erratic this time with its selection of keywords in each scene and the choice of content that went with it. It seemed to home in on the word “server” which sometimes produced a good result but other times was not really that useful. It also juxtaposed different styles of video; merrily jumping from corporate-style graphics to cutesy cartoons and back again. This video felt less polished than the first by far.

I also hit a very objectionable combination of the text “Does this really matter?” with a video of a Black Lives Matter protest. Clearly, a completely inappropriate and highly offensive combination. This really highlights the limitations of the system in understanding content and context. This is not a system that you can use in a “fire and forget” style.

After a lot of searching through the Pictory library, I’d managed to find a set of videos that were reasonably consistent in style and replaced all the inappropriate and bizarre choices the system had made.

Conclusions on Pictory.ai

Pictory.ai is a good piece of software and, with a well written script, could certainly create some interesting videos. However, the exclusion of 16:9 videos from the standard subscription is a big issue for me and would certainly limit how useful I find the solution.

I’d recommend you give the solution a try, but keep in mind the price point.

Leave a Reply