2020.05.26 - I generally focus primarily on photography, but given the diverse capabilities of the Nikon D850, I have found myself occasionally delving into the videography opportunities that present themselves. In the previous post, I set up on the bank of our local lake to photograph a fantastic lightning storm, but over time the lightning strikes became so frequent that I figured it was a perfect opportunity to take a video to compliment the photographs. Since this was my first video of a storm at night, and my first time editing a video at all, it is readily apparent that I still have a lot to learn, but it's a start and I am enjoying developing a new skill.
Composite photo showing the full breadth of the storm over about 20 minutes
My previous post on this storm had most of the photos I wanted to publish, but after going through all the images for that post, I felt compelled to create a composite photo that gave a better sense of the scale and variety of lightning strikes that the storm produced. It was taking much longer to create the composite than it took to process the individual images, so I chose to include the composite in a separate post with the videos that were also taking a long time to produce as I taught myself how to use Final Cut Pro. Apparently that was a good decision because with so little free time recently, it has taken me over a year of off and on work to get back to finalizing and publishing this second blog post.
You might be wondering why I chose to use 16 images in the composite. The truth is that originally I had over 20 images, but there was a point where adding more images (especially strikes that overlapped) was blowing out the detail in the image, introducing strange colors, and/or distorting the form of the clouds too much. I painstakingly went through every photo to blend them all into a final image that I found pleasing. Unfortunately, this means that I lost some interesting lightning strikes in the process, but the tradeoff is worth it to me to create a more balanced image.
Controlling the color of the image was particularly difficult. The color of all the different types of light in the scene varied wildly from image to image. There are many factors that lead to the shift in color. Lightning bolts on the surface of the cloud were a cooler bright white, while strikes within the clouds were warmer. It created a very uncomfortable aesthetic to just layer them on top of each other. In addition, the artificial light from the neighborhood, and reflecting off the bottom of the clouds nearest to me, created a green tint that was very distracting. In the end, I chose to modify the tone of highlights, mids, and shadows in order to come to the final warmer image that keeps the internal strikes closer to how they looked in person.
Close up of the main group of strikes in the upper 2/3rds of the storm
It was a very dark night, so I already had the camera set up on a tripod to take the 10 second exposures of the storm with the Nikkor 50mm f1.4. Transitioning to video was as simple as flipping a switch on the back of the camera to bring up the video recording screen. However, modifying the settings for video was considerably more complex. As previously mentioned, I have limited experience with video, so I chose to switch the camera from Manual Mode (which I prefer when photographing) over to Program Mode so the camera could auto-meter the image and balance the difference between the lightning strikes and the time in-between. I made a few tweaks to the exposure compensation to balance the live view image on the screen, and began to record once I felt that the overall video had a balanced exposure. I recorded for about 5 minutes during which there was some fantastic lightning. I sat back in the soft, gently cool grass and enjoyed the show while swatting away mosquitos as they bit through my shirt and pants.
There is a notable difference between taking photos of a storm and taking video. When taking photos I can generally keep ISO low to reduce noise, and the aperture closed enough to provide sufficient depth of field, by stretching out the shutter speed to whatever the exposure demands. However, a video needs to maintain a minimum shutter speed of 1/60 which gives the sensor 1/600th of the time to gather light as I had with the 10 second exposures. This means that noise and depth of field had to be sacrificed to create a useable exposure, so the resulting video is notably softer and noisier than the still images. In addition, the camera's metering system made the assumption that I wanted the entire image metered, so my first 5 minute video was very overexposed. While this made the lightning less interesting (which is unfortunate because the lightning was bigger and more frequent during this video), it was interesting to see how the clouds moved. I decided to focus on the cloud's movement by speeding up the video 1000%. This definitely gives a better sense of the cloud movement, but it creates an almost strobe effect out of the lightning which is rather distracting. At the very least, the video gives a good sense of how dramatically the storm can shift over a short period of time which is hard to see with the naked eye.
WARNING: This video may trigger seizures for people with photosensitive epilepsy. Viewer discretion is advised.
After viewing how overexposed the first video was, and how much the resulting noise was distracting from the subject, I modified the exposure compensation to -3.0 to hopefully limit these issues. In hindsight, I think I might have preferred manually setting a single exposure value to limit the exposure ramping between strikes, but since I haven't tested that, it is equally likely that I would have lost detail in the strikes themselves as a single exposure without real-time modification might over or underexpose the scene. There was a considerable difference in exposure between the isolated single strikes and the larger internal ones. Overall, the automatic metering was fairly accurate and quick to respond to new conditions, so it's hard to complain.
By the time I took the second video, the frequency of strikes had noticeably slowed, so it took a considerable amount of time to create this video as the lightning strikes were spread out over about 2.5 minutes. While it is perfectly enjoyable to sit and watch a storm roll by with lightning striking every 5 - 30 seconds, it doesn't create a very compelling video. What's more, the lightning was too fast to really appreciate the path and complexity of the bolts. For this reason, I went through the video feed frame by frame to isolate the lightning strikes, slowing down the video to around 40%, and compressing the time between strikes by speeding up the video from 150- 200%. The sped up video is largely imperceptible, but a few times during the recording the gap between strikes was simply too long to just speed things up. I chose instead to splice the video and create a morphing transition which helped pacing, but it is definitely noticeable. In the end though, I was more interested in the lightning strikes so hopefully the transitions don't distract from that too much.
WARNING: This video may trigger seizures for people with photosensitive epilepsy. Viewer discretion is advised.
My first foray into taking and editing video wasn't as good as I would have hoped, and it took much longer to learn Final Cut Pro than I had expected, but in the end I still think it was worth the effort to learn and create something new. Hopefully there will be more, and presumably better, video projects in the future.
Previous Post: On Thunderstorms and Parenthood
© 2017-2021 Shaun C Tarpley
Komentar