AI Short Video Generators: Streamlining Content Creation

The Rise of AI-Generated Video Shorts: Streamlining Content Creation

In the ever-evolving landscape of digital content creation, the emergence of AI-powered video generation has revolutionized the way we approach video production. As the demand for video content continues to soar, the traditional three-stage process of pre-production, production, and post-production has been significantly impacted by the integration of artificial intelligence (AI) technologies.

Traditionally, video production has been a labor-intensive and time-consuming endeavor, involving meticulous planning, filming, and editing. However, the introduction of AI-powered tools has the potential to streamline this process, making it more efficient and accessible to a wider range of creators.

The pre-production stage is the planning and design phase, where ideas are conceptualized and developed. In this stage, AI can assist creators by automating tasks such as scriptwriting, storyboarding, and even generating visual concepts (Gopoint, 2022). AI-powered tools can help streamline the ideation process, allowing creators to focus on the creative aspects of their projects.

The production stage involves the actual filming and capturing of the video content. While AI’s role in this stage may be more limited, advancements in AI-powered cameras and real-time decision-making tools can enhance the efficiency and quality of the production process (Lumira Studio, 2023).

Post-Production

The post-production stage is where the magic happens. This is where the footage is edited, color-corrected, and polished. AI-driven video editing tools have revolutionized this stage, automating repetitive tasks such as scene detection, shot selection, and color grading (Vitrina.ai, 2023). This allows editors to focus on the creative aspects of storytelling and enhancing the overall quality of the final product.

Streamlining Content Creation with AI

The integration of AI into the video production process has significantly streamlined the content creation workflow. AI-powered tools can assist creators in various ways, from generating ideas and scripts to automating post-production tasks (Wistia, 2023).

One of the most notable examples is the integration of Veo into YouTube Shorts. This AI model allows creators to generate high-quality, 1080p resolution videos that can exceed a minute in length, in a wide range of cinematic and visual styles (DeepMind, 2024). This capability, combined with the ability to quickly generate backgrounds and six-second clips, empowers creators to produce more impressive and engaging short-form content (TechCrunch, 2024).

The potential of AI-generated video shorts is evident in the examples provided below, generated with a prototype AI Video Short system I have created:

  1. Latest Research on Sulfites and Potential Health Effects (2024): This short video, generated using AI, highlights the latest research on the potential health effects of sulfites, a common food preservative. The video’s concise format and informative content make it an effective way to educate viewers on this important topic.
  2. Labellens Ingredient Focus: Azodicarbonamide (ADA) Potential Health Effects: This AI-generated short video delves into the potential health effects of azodicarbonamide (ADA), a food additive often used in processed foods. The clear and visually appealing presentation makes the information easily digestible for viewers.
  3. SafeGuardianAI – Decentralized AI-Driven Disaster Response Assistant: This 42-second educational marketing short showcases how AI can be used to create engaging and informative content for a specific audience, in this case, promoting a decentralized AI-driven disaster response assistant.
  4. Drone Report – Emerging Drone Threats: This 34-second educational channel short video demonstrates how AI can be leveraged to create concise and visually striking content that informs viewers about emerging drone threats.

These examples illustrate the versatility of AI-generated video shorts, which can be used for a wide range of purposes, from educational content to marketing and promotional materials.

Technical Breakdown of an AI Video Short Generation System

I’ll be going into more technical detail in a series of upcoming blog posts, but essentially the system works as follows:

  1. Create a prompt: add some context to act as the seed for the AI script generation process
  2. Feed the prompt to a generative AI system: receive a formatted structure broken into ~6 short 1-line descriptive scene elements, with associated image descriptions
  3. Feed the image descriptions into an AI Image generator: use some parameters to guide the overall graphic style of the video: realistic, cinematic, line art, comic book++
  4. Feed the associated 1-line descriptive scene elements into a text to speech AI system: with associated narration gender and style and retrieve the audio
  5. If you wish, animate the still images to the length of the script lines with a parallax shader with parameters to define the animate style you would like. Alternatively, you can just create still image clips.
  6. Create a complete video: with all the clips and video transition effects between them. There are many different types here to vary the visual interest. In this stage you can also overlay a logo/watermark at different screen positions, as well as add an optional branding end scene with contact details. 
  7. Add the audio narration to the previous steps video: and then feed that into a captioning system (if you want captions!) that will create animated captions with custom font, colors, and position while also highlighting the current word being spoken on the video narration timeline.
  8. Mix in the background music: not too loud to conflict with the narration, and not too soft you can’t get a sense of it for atmosphere.
  9. Output the created video: upload to the various social networks for distribution

Latest Research on Sulfites and Potential Health Effects (2024) demonstrates all these options being used, but you can combine any combination of them to suit the message you want to support.

The video can be generated with a local application, or through a SaaS based cloud-based system using a series of microservice API’s. I’ve currently developed the desktop generation system and have completed implementing the video captioning system as a FastAPI microservice running from Google Cloud.

I’ll be going into more details into this and other components of the ai video short generation pipeline in upcoming technical blog posts.

The integration of AI in the video production process has the potential to revolutionize the way we create and consume video content. By streamlining the various stages of video production, AI-powered tools can help content creators save time, reduce costs, and increase the overall quality and effectiveness of their videos.

As the technology continues to evolve, the future of AI-generated video shorts is promising, with the potential to make video creation more accessible and democratized than ever before. By embracing these advancements, content creators can unlock new possibilities and deliver engaging, informative, and visually captivating video experiences to their audiences.

Written By

The Power of Repetition and Interleaved Learning in MSFS

Today I want to talk about how repetition and interleaved learning can help you master the sim piloting skills of take offs and landings using MSFS.

Repetition is the act of doing something over and over again until it becomes automatic. Interleaved learning is the practice of switching between different topics or skills in a random or varied order. Both of these methods have been shown to improve retention and transfer of knowledge and skills in various domains, including aviation.

Why are repetition and interleaved learning important for take offs and landings?

Well, these are two of the most critical and challenging phases of flight, and they require a lot of coordination, precision, and situational awareness. They also vary depending on the type of aircraft, the weather conditions, the airport layout, and the traffic situation. Therefore, it is not enough to just learn how to do a take off or a landing once and then forget about it. You need to practice them frequently and in different scenarios to build your confidence and competence.

MSFS is a realistic and immersive flight simulator that allows you to fly anywhere in the world with any aircraft you want. You can also customize the weather, the time of day, the traffic, and the failures to create realistic and challenging situations. MSFS addons like location manager and aircraft manager provide features that let you save your favorite locations and aircraft settings for easy access.

For example, let’s say you want to practice take offs and landings at KLAX Los Angeles International Airport in California, USA. You can use the location manager to save this airport as one of your favorites, and it will automatically show you how many runways and parking spots are available, as well as the ILS frequencies if any. You can also use the aircraft manager to save your favorite aircraft types, livery, fuel load, weight and balance, etc.

KLAX ILS Training with Location Manager

KLAX ILS Training with Location Manager

Then, you can use the location manager toolbar in fly mode to quickly switch between different runways and parking spots without having to go back to the main menu. This way, you can practice take offs and landings from different directions and distances, with different wind speeds and directions, with different traffic patterns, etc. You can also use the aircraft manager weight and balance toolbar additions to change your aircraft settings on the fly, such as changing the fuel, passenger, or cargo load.

Changing Weight and Balance presets

Changing Weight and Balance presets for quick aircraft reconfiguration

By doing this, you are applying repetition and interleaved learning principles to your simulation based training. You are repeating the same skill (take off or landing) multiple times until it becomes second nature. You are also interleaving different variables (runway, parking spot, weather, time of day/night, position, distance, bearing, height, speed ) to make your practice more varied and challenging. This will help you improve your memory, adaptability, and problem-solving skills. You can also complelty randomise all these variables to really test your skills.

If you want to see how this works in action, check out this video where I demonstrate how to use the location manager and aircraft manager features in MSFS. I also show you some examples of how I practice take offs and landings at Bora Bora Airport in French Polynesia using these features, in addition to pointing out other features using Lukla (height AGL estimation) and KLAX (ILS training).

You can use it for all sorts of scenario based training:

  • Take offs
  • Take off emergency procedures
  • Landings
  • Go around
  • Landing emergency procedures
  • Varied landing approaches
  • ILS familiarisation and training
  • Whatever you can come up with!

To get the best out of Location Manager it’s best to watch the above video, and also refer to the extensive notes in the Tips section: How to best use Location Manager 

Future improvements to this process could involve things like:

  • Improved failure triggering in take off/landing (currently set via the failure menu before flight)
  • Traffic issues impacting the pattern sequence
  • ATC instructions
  • [insert here]

Will see how things progress! I hope you enjoyed this blog post and learned something new.

Feel free to send comments and feeback via the Contact Form, I’d love to hear from you.

Until next time, happy flying!.

Location Manager for MSFS – Development Overview

Location Manager for MSFS – The Challenge of Enhancing the UI/UX

Following on from my previous post about Aircraft Manager for MSFS – Development Overview , here is a short overview of how the development of Location Manager panned out. To recap what the problem is: MSFS is like an interactive earth simulator, enabled by flying around complex mini-simulations in the form of aircraft across the whole history of flight. The two methods for currently managing locations in MSFS are show below.

Inbuilt MSFS method 1 for location management

Inbuilt MSFS method 1 for location management

Inbuilt MSFS method 2 for location management

Inbuilt MSFS method 2 for location management

But there’s no way to save and/or favorite locations in the sim itself. While you can save locations outside MSFS in all sorts of ways (OneNote, LittleNavMap, Excel, etc) it’s not the most convenient of methods if you want to build up a log of world exploration discoveries for easy access within the simulator. This isn’t the same as loading a saved flight plan either. There are many use cases where you don’t actually want to use a flight plan, but just go to a location and start flying from there. So the intent here was to derive a solution to solve this problem.

Step 1 – Validate the Technical Approach

In development, it’s more efficient if you can leverage past knowledge and skills so you don’t have to reinvent the wheel every time. Having just completed Aircraft Manager for MSFS, I’d climbed the learning curve on both the MSFS UI framework as well as some Javascript solutions to integrate into that. So it seemed logical to ask, how could I apply that to the unique problem of location management? First off, just as in the previous post, I needed to validate some technical issues before diving into the prototyping stage. This was a very different project from Aircraft Manager, as that relied mainly on the inbuilt aircraft database supplied by the sim. In the case of Location Manager for MSFS, one of the big issues here was we needed to develop a robust way to save locations with associated user data via the inbuilt MSFS UI storage system. There were also unanswered questions as to how robust this storage system is in terms of how much data can be safely stored, on both PC and XBox. I have received some advice from one of the Asobo devs on this (thanks!) but still need a little more clarity, so need to keep an eye on this. In summary, some of the biggest technical issues to be validated were:

  • How to capture and save the location information from the MSFS UI
  • Where would be the best place to insert the UI to manage saved locations
  • How to replicate the map zoom functionality in MSFS for saved locations (this was tricky!)
  • How to efficiently store locations for easy retrieval, display, updates, and deletion

So I got to work, solving each of these unknowns in turn, or at least getting me into the ballpark of “this looks doable in a performant manner so the projects not dead and I can keep going”. Like the previous project, there were a number of dead ends and frustration moments which necessitated some creative thinking in order to spark the idea for another way to solve the problem. Some of these solutions come from gaining a deeper understanding of the MSFS UI framework, some from creatively applying workarounds. A workaround is a little different from a hack btw. A hack is typically a short term fix, designed as a purely quick and dirty temporary measure while you search for a better more reliable long term solution to a problem. A workaround is a solution you have confidence in being robust enough for long term use, even though it’s not how you originally envisioned it working. Workarounds are always preferable to hacks, giving you better maintainability over the long term.

Step 2 – Validate the Prototype

After gaining sufficient confidence I had working solutions to the core technical issues, I could start constructing a full prototype incorporating these into a working demo. This lets you test & tweak the full workflow, and weed out other issues/bugs.

Early Prototype of the Map interface

Early Prototype of the Map interface

Early Prototype of the location management interface

Early Prototype of the location management interface

Location Manager for MSFS Free Demo

The next stage from this is to get a demo into peoples hands to get wider feedback, while continuing to tweak workflows and solve issues/bugs as they appear. Similar to AM, I fired up a thread at the Flight Simulator forums to see if anyone is interested in trying it out. Have at it, all feedback is welcome: The Good, The Bad, and The Ugly. You can download the free demo lite version from the Location Manager for MSFS project page. Like the AM demo, it’s fully functional for casual use cases, with no obligation to upgrade. Enjoy!

Location Manager for MSFS Pro Release

Version 1 of LM Pro

Version 1 of LM Pro

Location Manager Pro V1.0 was released on 2023-02-14. It includes a toolbar widget for accessing LM from fly mode, so you can teleport to saved locations as well as add new one. The release announcement on the MSFS Forums is here @ Aircraft Manager Pro + Location Manager Pro now available.

Start Anywhere in MSFS

It was quickly followed up with another solution to a long standing request from users of MSFS: the ability to start cold and dark from anywhere. You are currently limited to being able to start only from airports or in the air.

With V1.0.3 of Location Manager we introduce Start Anywhere, where you can teleport to saved locations in Fly mode and start on land, water, or air to begin your flight experience.

More details in the video below:

Happy Flying!

Aircraft Manager for MSFS – Development Overview

Enhancing the MSFS UI & UX to solve the Aircraft Management Problem.

MSFS (Microsoft Flight simulator) has been a going concern for 40 years, with some ups and downs along the way, and the latest iteration launched in 2020 is revolutionizing the flight simulation space. It’s a complex simulation for sure, not without issues or growing pains, but it’s undeniable that when it works well it is an amazing experience in either flatscreen or VR. I’ve been using it extensively on and off since the launch, and while I like the UI there’s always been a few niggling issues with the main interface that I’ve been itching to scratch. One of the pain points is with the Aircraft Management aspect via the Aircraft Selection Screen, and Aircraft Manager for MSFS is my solution to that. This blog post is a short story of how that came about as a new addon for MSFS called Aircraft Manager.

MSFS Aircraft Selection

MSFS Aircraft Selection Screen

A couple of times in the past I’d taken a look through the User Interface code along with exploring the official SDK, but as there was no documentation on it I steered clear of doing anything with it because it looked too complex. Diving into a complex codebase with no documentation to guide you can be a bit of a scary prospect, and you need to be prepared to persist with it in order to get that series of lightbulb moments where you being to understand it.

Anyway, towards the end of 2022 I was casting around for new side project ideas that could solve a real problem and make an impact. Two areas in MSFS stood out for me as candidates to develop something for:

  • Aircraft Selection and management, made more difficult with the growing number of aircraft now available both paid and free.
  • Managing locations inside the sim. MSFS is like an interactive earth simulator, enabled by flying around complex mini-simulations in the form of aircraft across the whole history of flight. But there’s no way to save and/or favorite locations in the sim itself.

So I decided to look at enhancing the Aircraft Management problem first. More about Location Management in a follow-up post.

The Problem To Solve

Some of the issues with it are:

  • A growing aircraft collection means too many planes, and it’s easy to get overwhelmed deciding what to fly using the current UI
  • Frustration you can’t tag your aircraft the way you like
  • Can’t sort or filter them in meaningful ways to use them in various contexts eg: racing, stol, vfr etc
  • Can’t take easily accessible short notes ingame as reminders before flying
  • [insert your itch here]

First Steps to Developing a Solution to Aircraft Management for MSFS

Getting to grips with the MSFS UI (User Interface) was the first big rock to move, as we say in project management. You can’t cut into a codebase unless you know what you’re doing.

Unfortunately, there is no documenation on the UI at all (some is finally coming in SU12 release, though) so it meant doing it the hard core way: reading and debugging, piecing it together line by line.

Fortunately, this is one of my core skills as a developer. I’m a regular bug magnet, so my debugging and analysis skills are fairly well developed. That said, I knew it wouldn’t be easy, which is why I’d avoided it so far!

So I blocked some time out and I went for that long overdue deep dive into the MSFS UI codebase. Three days later, after some hair pulling, screaming, wtf’s, and aha’s, I surfaced with *some* idea of how it was put together. Shoutout to both Microsoft and Asobo for providing the source code in the first place, and also to the dev(s) that architected and implemented it. Impressive framework effort, and not an easy project to develop even leveraging a UI library.

My pathway to developing this was through iterating rapid prototypes to learn more about how the MSFS UI framework worked, while testing out ideas for how I might enhance it, before launching into the actual development solution. This next phase was even more educational, as I ran across the “gotcha’s” inherent in an undocumented system built on top of  Game UI library Coherent GT.

Not having the ability to implement anything at the Coherent GT level meant understanding fully what might be possible to do, or not, and finding the best solution. I had a short list of technology solutions I wanted to use, based on my previous experience, and prioritised them into a risk reduction approach to prove various technical issues could be solved in order to develop a working enhancement solution.

I worked my way through various ideas I had, and was getting a little dejected when they all ended up on the cutting room floor. This project involved lots of creative pauses, while I had to go for a walk and let it stew for a while in order to spark an idea to find another solution to try. Eventually, I hit paydirt with a robust solution I could build upon that seemed reliable enough to surgically insert into the MSFS UI system and enhance it without disrupting regular use, a pretty critical requirement!

Developing the Prototype

Prototyping a more complete solution was the next stage, where you take the bits you now know work reliably and start to craft a proof of concept that covers the core use cases. The goal of the product development wasn’t to develop an overly complex solution, but a lean mean aircraft management machine, that solves a couple of core use cases very well, as opposed to being a swiss army knife solution.

Once I had a working prototype that was functional enough to prove the concept I reached out to other MSFS users at the MSFS forums to see if anyone was interested to test it and give me some feedback.

Early Prototype of Aircraft Manager for MSFS

Early Prototype of Aircraft Manager for MSFS

Shoutout to the following users for taking a chance on testing and providing feedback for Aircraft Manager for MSFS:

It’s always vital to touch base with other users, as even though I’m scratching my own real itch with the UI and solution other users have different itches, and sometimes they overlap in interesting ways. Stress testing for performance is also critical, as is squashing bugs which are easy to miss even with a QA checklist.

The end result (but not quite the final one, it’s changed already!) you can see in the video walkthrough below:

Update: Does it work to solve real users’ problems?

Very well! Possibly the most useful add-on ever for MSFS, in this great review from one of the top MSFS streamers.

AM Free Demo

You can download the free demo lite version from the Aircraft Manager for MSFS project page. It’s fully functional for casual use cases, with no obligation to upgrade. Enjoy!

AM Pro Release

Version 1 of Aircraft Manager Pro for MSFS

Version 1 of Aircraft Manager Pro for MSFS

I’m currently in the final stages of release for the pro version of Aircraft Manager for MSFS. Making a product is only part of the challenge, there’s a lot of other elements that go into a successful rollout, open source, free, or paid.
This stage always reminds me of the 90/90 rule:

The first 90 percent of the code accounts for the first 90 percent of the development time. The remaining 10 percent of the code accounts for the other 90 percent of the development time.[1][2]

Tom CargillBell Labs

It’s not just code. Website, payments, fulfillment, support, maintenance and the rest all need to be done. That’s a lot, and I found I had to refresh a few of these areas as well, which hung up the deployment of the pro version. The release announcement on the MSFS Forums is here @ Aircraft Manager Pro + Location Manager Pro now available.

Auto Control Preset Switching

The ability to automatically switch device control presets on a per aircraft basis has been a long standing request from the users in MSFS.

Item #17 at 749 votes in the most recent wishlist is “Aircraft Specfic Control Profiles”.

Well, now we have it! With V1.0.2 of Aircraft Manager you are able to save per aircraft control presets and have them switch automatically as a bank of device preset changes whenever you switch aircraft. No more manual switching needed. See the video below for more details.

The Story Continues: Location Manager for MSFS

As I mentioned at the start, the other problem I came up with to solve was:

  • Managing locations inside the sim. MSFS is like an interactive earth simulator, enabled by flying around complex mini-simulations in the form of aircraft across the whole history of flight.

After getting hung up on some of the product deployment issues, I decided to take a look at this problem while waiting on people to get back to me.

Location Manager for MSFS free version is now available from the project page,

I’ll be writing more about this shortly in a follow up post. Stay tuned, and thanks for reading if you manged to stick it out this far!

Update: Part 2 is now available at Location Manager for MSFS – Development Overview.

Immersive Audio and Musical AI

The sound and music are 50% of the entertainment in a movie. George Lucas

As it is in games, I would also say the same for emerging mediums such as VR, AR, and MX.

I’ve been working recently on a couple of projects related to a few different areas, such as:

As part of this, last week I presented an introductory tutorial at the VRHubTokyo Meetup on Spatial Audio in Unity3D.

This is a really interesting area, with a lot of challenges still to be overcome.

This panel video of immersive audio experts is an excellent introduction to a lot of the issues:

We also had another Canvas Tokyo Musicians Meetup where we discussed a few of these topics, among others.

You can read the event report here: Musicians Meetup #2 Event Report – Mon 11th Sept @ Tacchi Studios

Check out the Spatial Music Visualizer which I also used to discuss a few of the issues around spatial music.

Stay tuned for more news soon!

Cognitive Technology – a Quickstart Guide Infographic

“I’m sorry Dave, I’m afraid I can’t do that” HAL 9000

Cognitive Technology, Smart Apps, Data Science, Artificial Intelligence…the list of new technology buzz goes on and on.

But what do they actually mean?

Cognitive Technology Infographic

Here’s an infographic to help explain what Cognitive Technology is and how it can be applied for business use.

The use cases are many and varied, depending on your business needs.

  • Expert systems
  • Autonomous vehicle control on Land. Sea, and Air
  • Robotics
  • Sales and inventory forecasting
  • Price modeling
  • Image, video, and handwriting recognition
  • Text and document analysis for classification, concept, and sentiment extraction
  • Customer segmentation and market analysis
  • Social network analysis for brand and product affinity
  • Anomaly detection for fraud and other behavior patterns

There are also many vendors in this space ranging from the big 4

  • IBM
  • Microsoft
  • Google
  • Amazon

to many smaller players focused on providing different levels of service.

Enjoy!
Image Credit: Evolution by Jakob Vogel from the Noun Project

Cognitive Technology Quickstart Infographic

Cognitive Technology Quickstart Infographic

Pin It on Pinterest