When integrating video stabilization into a camera product, tuning is a key part of the product development process. As opposed to calibration, when you optimize video stabilization by ensuring that your hardware and software are in sync, tuning is about customizing video stabilization to your specific use case. This involves making compromises and setting priorities.
If you’re an engineer tasked with deploying video stabilization software for a camera in motion, you should be aware of key considerations for tuning so you can make the right choices for your needs. We are putting the spotlight on five key tuning considerations from our guide “Mastering video stabilization in product development”to show you what your options are and provide some examples as inspiration.
1. Battery life considerations
You don’t want a drone that will be used to film during critical operations for a long period of time to run out of battery power too soon. If you opted for real-time processing, then this could already be taking a toll on your processing capacity and battery life. One way to compensate for this is to streamline video stabilization performance to consume slightly less processing power and thus also battery power.
On the other hand, if your battery life is already more than satisfactory for your use case, you can probably afford to tune up your video stabilization performance.
2. Field of view considerations
An example of cropping with independent video enhancement features from different providers, each additional feature means additional cropping.
An example of Vidhance image cropping uses the same input parameters which results in less cropping
For bodycams, you’ll probably want to sacrifice more of your field of view to boost video stabilization. Try running like a cop chasing a criminal and record a video on your regular smartphone while doing so. Standard video stabilization is not designed for moving so quickly so the video will probably be too shaky for comfort.
Therefore, you can imagine that real-life law enforcement officers need more stabilization than the average use case. In other contexts, like a surveillance camera, you may want a broader field of view so you can monitor a wider area.
3. Motion blur considerations
Video artefacts such as motion blur already exist in video feeds initially but applying video stabilization can increase their visibility. Motion blur reduction can be applied to compensate for this, but this in turn will increase the noise level. Therefore, you’ll need to find the right balance between an acceptable motion blur and noise level.
Bear in mind that how much motion blur you will experience will depend on lighting conditions and on how well your motion sensors handle light. Consider whether your camera product will be used frequently in low-light conditions, such as indoors and at night, when you will need as much additional lighting as possible. The less external light is available, such as sunlight, the more motion blur will be an issue.
4. Video quality considerations
Electronic Image Stabilization (EIS) applies a crop to each frame. The part of each frame that the algorithm selects for cropping is what makes the video stable. As a result, if you take full-HD 1920×1080 resolution video and apply video stabilization, the resolution will decrease slightly. Depending on your needs, you can tune video stabilization performance down and resolution up, or vice versa, by configuring how much of the frame is cropped.
A sensor with higher resolution to begin with will not suffer as noticeable a loss in quality when cropped as a sensor starting from a lower resolution. However, it’s worth considering that higher resolutions will take up more battery power to process due to the higher number of pixels.
5. Horizon correction considerations
If you want a drone to be able to shoot video with good quality in windy weather, you’ll need to compensate for the wind. The drone will tilt to move in the right direction, and this can make the video extra shaky. This may require correcting for a wider angle in horizon correction, which comes at the cost of field of view.
However, in other operating environments, field of view may be more important, and you may have satisfactory results with a narrower angle of horizon correction. Simulate real-world conditions for your use case and analyze the video output to determine your horizon correction priorities.
Tune like a pro with Vidhance SDK
Now that you know what your options are, analyze your needs and tuning priorities carefully. When you’re ready to dive in and get tuning, your work will be much easier if you have flexible tools and documentation on how far you can go when adjusting values up or down.
We give you that and more in our software development kit – Vidhance SDK to empower you to tune like a pro. Don’t hesitate to get in touchwith ourvideo quality experts to learn more about how to get the most out of tuning with our SDK.For inspiration, insights and best practices for the next generation of video enhancement, enter your email address below and subscribe to our newsletter.