Tutorial Maker Application: Create Videos and AI Context Documents

Documentatie •
In dit artikel

Over the years, at Banana.ch, we developed an internal Video Maker tool that allows us to create professional video tutorials using a developer-inspired continuous improvement approach. This method has enabled our developers to produce highly effective tutorials in a very short time and keep them up to date as the software evolves and user needs change. 

We have created and maintained hundreds of presentation and tutorial videos in multiple languages. They have been very well received by our customers and have significantly reduced support requests. We have also started experimenting with using the same video production material to create documentation for AI assistant tools, and we see great potential in this combined approach. 
We therefore decided to better understand whether there is interest from other organizations in such a tool. As a first step, we made it available in the same form we use it internally. It works well, but for now it is best suited for tech-savvy users.

If you are interested, feel free to get in touch with us and let us know. 

Video Creation with a Software Development Approach

At Banana.ch, we understood that creating a good video tutorial is not something you do just once. It’s a process that improves step by step, just like when you develop software. But we were not happy with the tools we found. Most video tools are made for artistic work, need special video skills and are cloud based.

We wanted something more in line with the way software developers work. A tool that allows you to define—using a table-based interface—which components to use, the order of elements, the texts, and their translations. It should also work locally, with existing tools like Git, and support already in use capture and image editing tools.

So we created our own tool. We started from the code of Banana Accounting and built something new. It may sound strange to use accounting software to create videos, but the core programming of Banana Accounting resembles a spreadsheet, is written in C++, and is both flexible and powerful. In fact, we discovered that creating a video using clear instructions and structured data is not very different from entering accounting transactions to create a balance sheet. So, inside Banana Accounting, we created a new application, Video Maker, that uses tables and columns to enter all the information needed to build a video. Then we added the programming logic to automatically generate the video from this data.

The application works like a mix between software programming and an animation tool. It lets you create a video by putting together a sequence of captured images. You can add titles, text for the images, subtitles, and text to be converted into speech. 
We also integrated the ability to add some basic effects, such as gray zones, colored boxes and circles, arrows, and sequence numbers. You don't need to record the mouse cursor, and more importantly, the tool lets you reuse the same captured image while still achieving the effect of an animated video. All the information is written in tables, and you can add texts in different languages at the same time. 
The Excel-like table makes it particularly fast to enter data and to copy and paste text to and from other programs, making it easy to translate, correct, and improve the content. 

As developers, in addition to allowing each property to be specified individually, we enabled users to define styles, both for text and voice. This greatly simplifies and speeds up the formatting process and ensures that all scenes have a uniform, polished, and brand-compliant look.

At any time, you can press “Build,” and the tool automatically creates a video in the language and voice you choose. 

  • First, it generates the spoken voice using AI, so you don’t need a professional speaker or have to record the audio manually. We added the ability to set the type, pitch, and tone, resulting in a clear and well-suited voice for tutorial videos.
  • Then, the tool adds titles, texts, and subtitles to the images. It creates short animated scenes and adds the voice. What makes this special is that each scene is automatically adapted to match the duration of the spoken text in each language. At the end, the tool puts all the scenes together and produces a complete video, with an index and transcript.
  • Like in software development, the first build is just the beginning. 
    You watch the video, make changes, ask others for feedback, and quickly create a new, better version. This way of working helps you improve step by step and respond to what users really need.
  • We noticed that people creating videos kept making small changes and wanted to immediately see and hear the result. So we optimized the build process, and thanks to the fact that all files are stored locally, we were able to speed it up significantly.

One of our main goals was to make it easy to create videos in multiple languages. The tool automatically generates speech in different languages, and each scene's length adjusts to match the voice. There’s no need to shorten or simplify translations just to fit the timing of the original version. This significantly speeds up the translation process, improves the quality of the explanations, and makes the video feel natural in every language.  

Capturing the images is usually the most time-consuming part. Doing this for every language is not realistic, especially if you expect to make changes later. We found that automatically adding translated texts, titles, and effects directly onto the original images helps a lot. It gives the impression that the video was fully translated into other languages and makes it easier to understand, without the need to create new visuals for each specific language.

Because software and processes change often, tutorial videos can quickly become outdated, contain errors, and lose their usefulness. Even just a few outdated videos can give the impression that the others might be outdated too. With this tool, the people who know the process can directly create updated versions of the videos, so the content stays accurate and helpful.

Videos are very useful, but some users told us it would be helpful to have a PDF document to consult, making it easier to study certain steps. So we quickly figured out that we could fully reuse the elements used to generate the video to automatically create what we call PDF Tutorials, with an image of each scene and the accompanying text. This way, both the video and the manuals in different languages are generated automatically and can easily be updated after any change. PDF Tutorials can also be given to ChatGPT, allowing quick answers to specific problems. 

Video has become very important. This solution completely lacks the friendly look and features of modern cloud-based video creation tools, but it does extremely well—and very fast—what it is designed to do, especially when changes are needed. It might really be worth taking a look and setting aside some time to give it a try.

How It Works

Creating a Video Tutorial for a Desktop Application.

  1. Capture Screenshots.
    Begin by capturing screenshots of the desktop application and saving them to a directory (recommended size: 1920x1080).
  2. Fill in the Clip Table, 
    Enter all necessary information to create a video:
    • Select the image to be used.
    • Apply effects to the image, such as highlighting a section, adding arrows, or circles.
    • Choose the text and voice style for the speech.
    • Input the text for titles and/or text-to-speech.
    • Translate the text into other languages as needed.
  3. Build the Video.
    On executing the build video command, the program performs the following tasks based on the specified language:
    • Adds effects, titles, and subtitles to the images.
    • Generates speech files using internal or Microsoft or Google voices, as selected.
    • Creates a video sequence for each image. The duration of each sequence matches the length of the speech or a duration you specify.
  4. Final Video Compilation. 
    The final video is created by combining all the video sequences.

How it works

Assuming you are creating a video tutorial for a desktop application.

  1. You capture screen shots and save to a directory (size 1920x1080).
  2. In the Clip table you enter all the information needed to create a video:
    • The image to be used.
    • Create effects on the image. Like highlighting a part of image, adding an arrow or a circle.
    • The text and voice style to use.
    • The text to be used for titles and/or for the text to speech.
    • The text translated to other languages.
  3. You give the build video command and the program, based on the language you specify:
    • It add effects, titles and subtitles to the images.
    • It generates the speech files using the Microsoft or Google voices you choose.
    • For each images it create a video sequence.
      The duration of the sequence is the duration of the speech, or one you specify. .
    • A final video is crated by putting together all video sequences.

Table Clips

Video tutorial of the Tutorial Maker application

Create video from images, with speech, subtitles and text decorations.

Here is the full project and files on github.com.

For tech-savvy users

The Tutorial Maker application offers advanced features specifically tailored for tech-savvy users.

The Tutorial Maker application is currently available in the Dev-Channel version and has been extended to a select group of partners. You're welcome to try it out; however, please note that it is not a fully supported product, and it requires the installation of additional components. Its features are also subject to change at any time.

Currently in its alpha phase, the application has been fine-tuned for internal use. It requires a specific setup and is still in the early stages of implementing error handling. Moreover, the interface is quite raw, and all messages are in English only. As such, it is currently best suited for tech-savvy and technically skilled users.

Please note that the structure and format of the data may change over time, and we may not always provide tools for migrating to new versions. Terms of use and pricing are also subject to change. We reserve the right to modify technical content, availability, conditions, and pricing at any time without prior notice.

We invite you to share your experience using the application and greatly value any feedback or suggestions you may have.

 

 

 

 

 

help_id
fvideo

Help ons de documentatie te verbeteren

We horen graag uw feedback over hoe we deze pagina kunnen verbeteren.

Vertel ons welk thema een betere uitleg nodig heeft of hoe we een onderwerp kunnen verduidelijken.

Share this article: Twitter | Facebook | LinkedIn | Email