Unveiling the Process: The Mechanics Behind Converting Videos to 4K Quality

The transition from analog film to digital video occurred rapidly. Tape-based video recorders were soon replaced by digital cameras and smartphones. Today, 4K resolution has become the standard for video production, leaving 1080p as the lesser option.

Even with the growing popularity of 4K, a significant amount of video content is still being produced in 1080p, along with older content available in standard or even lower definitions. This leads many to consider the possibility of converting such videos to 4K to achieve a sharper image on modern 4K displays. However, the process of converting to 4K involves numerous steps and important considerations.

When discussing resolutions like 4K and 1080p, we refer to the pixel counts within each frame. For 4K, which is widely recognized as ultra-high definition, the resolution is 3,840 x 2,160 pixels, compared to 1,920 x 1,080 pixels for 1080p. Standard definition videos typically hold 480 or 576 vertical lines, varying by region. There’s also a cinematic 4K option known as “true 4K” or DCI 4K, which features a resolution of 4,096 x 2,160 pixels.

Upscaling any file to 4K requires producing additional pixels. To convert 1080p to 4K, three extra pixels are required for every original pixel, resulting in a total of four times the pixel count. The challenge lies in achieving this without sacrificing quality, which is nearly impossible. Simply altering the resolution won’t unlock added clarity or sharpness.

Traditional upscaling software often utilizes interpolation to scale up each frame by estimating what the pixels should look like. The accuracy of this estimation is crucial, but it cannot retrieve details that weren’t originally captured.

This often results in softer footage. The approximations used to add pixels can lead to a smoother appearance. While sharpening can enhance edges, achieving the right balance is crucial, as overdoing it can spoil the final result. Noise reduction can also produce an unnatural look if not carefully applied.

For professionals aiming to convert videos to 4K, the best approach is working with the highest quality source files available. The superior the source, the better the resulting quality.

The restoration of Star Trek: The Next Generation exemplifies the difficulties in achieving superior upgrades. In 2011, CBS undertook a task utilizing original 35mm film negatives to remaster the series for a Blu-ray release. This required 25,000 reels of original film and over three years for completion, with costs nearing $12 million.

Interestingly, the shift from analog to digital made these upgrades more manageable. When dealing with older digital files, you are limited by fixed pixel counts. The original negatives in the Star Trek restoration captured significantly more detail than what was broadcast originally.

However, when working with low-resolution source files, you’re left with only guesses for the missing pixels. While the result may appear convincing, it won’t entirely reflect the original quality.

Unlike large studios, home users usually don’t have millions to spend on video conversions. Those looking to upscale their videos to 4K will typically have to choose between traditional interpolation or newer AI methods. Both approaches create additional pixels for a larger 4K frame, but they do so in different ways.

Traditional interpolation uses mathematical calculations, assessing the colors and brightness of surrounding pixels to estimate how the missing pixels should appear. While some methods focus on preserving edges to avoid jagged lines, this can lead to a soft appearance as well, lacking genuine detail and texture.

AI upscaling, on the other hand, utilizes models trained on low and high-resolution image examples to make predictions about how the sharper image should appear, offering a potential enhancement over traditional methods.

AI can produce sharper and more convincing results, but, much like interpolation, it won’t always be an exact match for real-world images. Furthermore, AI approaches might introduce unique artifacts not present with interpolation. Nevertheless, AI models are continually evolving, suggesting that the quality of these upscalers will improve over time.

For older videos containing interlacing—where frames are divided into odd and even horizontal lines—conversion may require de-interlacing before upscaling. Achieving a decent quality upgrade from interlaced formats to 4K presents its own challenges.

Most upscaling tools offer configurations to improve your results, whether using AI or interpolation, allowing for additional adjustments if required.

Simply inserting a 1080p video into an upscaler will not guarantee a flawless 4K video as a result. While the technical aspect allows for resolution increase, maintaining quality during the upscale process poses challenges and may require multiple attempts to achieve satisfactory results.

Better source material simplifies this process, and smaller resolution jumps make it easier. For instance, upscaling from 1440p to 4K requires fewer additional pixels compared to 1080p to 4K or even lower resolutions.

Software can only create educated guesses about those additional pixels, and while AI may help, an exact match remains uncertain. Resolution isn’t the only factor; lower bitrate video files, often used to save space, don’t automatically improve after upscaling. Compression artifacts also remain unless the source material is strong enough to mitigate them, making the quality of the original footage a critical component in the upscaling process.