What Impact Does Color Have on Video Content?

According to something called the trichromatic theory, our perception of color comes from the different cones, or receptors, in our eyes responding to different wavelengths of light—short (blue), medium (green), and long (red). What this means is that if the surface of an object reflects a particular wavelength, our brain will interpret it as a certain color. Or to put it simply, the human eye and brain, together, translate what we see into color.

But what impact does color have on video content? We’ll get to the bottom of it. 

This article will focus on the invisible, brilliant, and deceiving nature of video-on-demand applications. Not so much about the content itself, but more so about the way the content is represented. 

The truth behind your perceptions

When interacting with video-on-demand applications our perception is being manipulated from numerous sources at any point of time. We can notice things in a split second or see hidden messages, sometimes even when there are none.

We need to realize that we have less of a say in things than we would like to, or are aware of. The advertising and entertainment industry tries its best to manipulate our minds into thinking we’re in control, when in truth, we’re not. Everything from button placement, color usage, and format, to trends, songs, and drama in the content that is pushed to the front page of your favorite application—absolutely anything goes to promote impression generation.

However, I want to focus on the attributes of color—brightness, contrast, saturation, and the perceived color itself. Things we can measure. What inspired me to look into the impact colors have on video is a finding I came across quite randomly and unexpectedly. Specifically, after spending some time on TikTok, I discovered that the videos on their platform are darker than the videos on the platforms of their closest competitors. But not only that, the entire video color scale is shifted. It seems that the blue color is more dominant than the other two primary colors—red and green. 

This got me wondering, what if other apps also change how the video is displayed and make some color attributes more prevalent than that of competitor apps. I continue to wonder, could that minimal difference in color further impact the effectiveness of blue light that is emitted from our device screens. With a late night session, is there a particular app that influences people more than its alternatives? Would that mean that I will watch an extra video because the underlying scheme will send more signals to my brain that it is not yet time to be tired.

I do not know and I do not have the scientific background to claim it as fact, and without sufficient evidence from a human study I can not say that these things really do impact us, but that does not stop me from speculating a little and conducting an experiment to see how color can affect video.

Here is an example of the same video on different platforms, specifically TikTok, YouTube, and Instagram Reels.

I will share my findings below.

First findings

The data comes from a 20 second test video that was uploaded and then downloaded from the target applications. In the case of TikTok, the video was downloaded in a way that does not involve the classic watermark with the logo in the bottom right corner. The histograms come in the same order. What we see here in the histograms, as well as in the images above is that YouTube, surprisingly, makes the video much, much brighter. The contrast is increased, which of course correlates to the second histogram, which shows the bars being skewed off to the right side, which means that the image is brighter.

The other two images show another interesting thing. Instagram Reels have a visible slight pink tint to them while TikTok have those darker, more blue shades. We can also see this in TikTok's histogram with the blue bar reaching higher and spiking stronger than the red and green colors, while Instagram Reels' histogram shows almost the exact opposite thing going on, with blue being much lower and red and green taking over. 

Now, these changes are very subtle, but interesting nonetheless. How much does this affect the viewer? Let’s imagine a nightly scenario and let’s say you want to watch a couple of videos before you go to bed. Would it mean that you are likely to spend more time watching videos on YouTube due to the fact that the videos are brighter? 

Secondary findings

Although the used codec is the same for the videos across platforms, the H.264, the upload time varies by a significant amount. Yes, VP9 and now AV1 have been here for a while, yet these codecs seem to be applied to the more popular videos on the platforms. The way the platforms approach video processing is different, as is the post-upload timing which I am not quite sure how to fairly represent.

TikTok is not only the fastest to upload, but also the fastest to show the final quality product to viewers. Instagram Reels come in second, taking a couple minutes to arrive at its destination. Most likely, the processing time of the high definition end product on YouTube is much shorter if you are a more popular creator, but otherwise it takes a lot of time.

The video processing and upload time seems to directly correlate to the final video resolution. Higher resolution means longer upload. As mentioned in an earlier blog, TikTok offers the lowest resolution to the user, and YouTube—the highest. Back to color, my idea was that the highest unique color count would be found in the highest resolution images, but that was not the case here.

The only time the original got close to our competitors in unique color count was when the image featured a bright red fire extinguisher, with other bright colors taking up space on most of the screen at the 14th second. Instagram Reels was the closest in most cases to the original, color-wise. Although it only streams the video at 720p, the color matching to the original is great. From looking at the images, TikTok versus YouTube seem to be on opposite sides of the color characterization spectrum. They take the original file and spin it their way. The increased average color count in both applications is likely due to a need to mix and blend other existing colors to align to the desired color attributes.

Tertiary findings

Globally, the average watch time on mobile devices is higher for TikTok than for YouTube, but the difference is not massive, and YouTube also has considerable desktop and TV viewership. Given that 63% of YouTube viewing time comes from mobile, we can not say that the quality of video is underappreciated. Format differences might come into play, but possibly, the reason why TikTok is dominant is because it is not offensive and very light on the viewer's eyes. YouTube might hold the viewer's attention for longer due to higher contrast and video brightness. 

By training a BRISQUE model on a public-domain subjective image quality database, we evaluated the videos from different applications. The end result surprised us. Without additional training, this algorithm does not discern between the videos. We can assume a similar lack of awareness for a regular viewer, one that can not be attributed to ignorance or unsophistication. Perhaps, processing time and load can be saved by sacrificing a bit of quality. Even though technological advancements are being made every day, it just isn’t worth doing yet because the trickle-down effect needs time. We need to focus on maintaining colors and adjusting contrast to attain the desired effect. More contrast and brightness—more attentiveness and drive. Darker, subtle colors—less offensive and easier on the person and a reason for them to come back for more.

So what can say about the impact of color on video content?

By looking at how the same video is displayed on various platforms—in our case TikTok, YouTube, and Instagram Reels—we can say that different attributes of color such as brightness, contrast, saturation, and the perceived color itself, all influence how we feel about the video. Brighter colors can impact how long we spend on a particular app, while some color schemes may send signals to our brain, telling it that it is not tired and can handle, for example, watching one more video before going to bed. 

All in all, we are not just number-crunching machines, it’s about feeling as well. So some things need to be looked at differently, not just the pure value that a BRISQUE or VMAF algorithm gives out. There is still a lot to explore to continue the development in this area and explore the impact color has on video content. 

Want to find out how we manage and work with these solutions or perhaps you have a video solution that you want to test or compare against competitors? Just shoot us a message. We’d be happy to help.