Can AI really take a soft 720p video to usable 4K? A practical test

Older videos do not lose their value just because recording standards have changed. Once, 720p resolution was standard for video quality. Now it looks thin next to everything else in a feed shot in 4K. 

However, many creators and businesses still have older clips worth keeping — product videos, early social content, archived interviews — even if resolution quality no longer matches current expectations.

The challenge is bringing those clips up to current standards without losing their original look. 

AI upscaling claims to solve this. But can it actually turn 720p footage into something that holds up at 4K? This test looks at what AI video enhancement can really deliver, and where it still falls short.

What upscaling actually does

Traditional upscaling is more focused on increasing the frame size rather than adding new details. A 720p resolution video is made up of fewer pixels than a 4K resolution video. Upon enlargement, the lack of such pixels may lead to:

● Soft edges around objects

● Blurry details

● Less sharp text

● Reduced texture quality

● A less professional appearance

AI video enhancement takes a different approach. Instead of only stretching the existing image, AI models analyze the footage and predict missing details. The goal is to create a cleaner result while keeping the original content looking natural.

However, the quality of the final output depends on the original footage. AI can improve a soft video, but it cannot fully recover details that were never captured.

Testing the process with Vmake Labs

To see how much AI could improve older videos, a batch of 720p clips was processed using the video enhancer from Vmake Labs （Formerly Vmake AI). The test included talking-head videos, product close-ups, and indoor footage recorded under low lighting.

Vmake Labs is an AI-powered content creation platform that offers tools for improving both videos and product images.  Along with the video enhancement feature, there are other tools available, like background remover, caption creator, video translator, and more, to help people create content for digital platforms.

The video enhancer tool operates in the browser without the need to download any software.

Vmake keeps uploading simple. 

The upload process supports three methods. 

Users can upload files directly by clicking, dragging and dropping, or pasting them into the editor. The tool also accepts YouTube, TikTok, and Facebook links through URL upload. In case there are several files, you may use batch upload and work with up to 30 files at once.

Once the video is uploaded, you will see that instead of one generic upscale button, the tool splits processing into resolution modes and scene specific modes. Resolution options include Fast Mode for quicker processing and Quality Mode for higher-resolution output, while scene modes are tailored for different types of footage, including Portrait, Portrait Pro, Product, Lowlight, De-AI, Game, and Anime.

That 8K ceiling under Quality mode is a recent addition to the tool, and it puts Vmake Labs video enhancer ahead of most consumer upscalers that stop at 4K. For anyone working with product photography or footage intended for large screens and print, that extra headroom matters.

What changed after enhancement

The first thing to understand is that the results were not the same for every type of footage. A soft 720p video does not have one single problem. Some clips lose detail because of poor lighting. Others struggle because of compression, motion blur, or weak camera sensors.

After processing different types of footage, the biggest improvements appeared in areas where AI could rebuild visual information more effectively.

Product footage

Product videos showed one of the clearest improvements.

In the original 720p clips, small details were blended together. Product edges looked softer, and textures were harder to notice. After enhancement, the footage looked cleaner and more defined.

The difference was especially noticeable when products had:

● Fine textures

● Small labels

● Sharp edges

● Reflective or shiny surfaces

The footage still had that phone-camera look to it. It didn’t magically turn into studio video, but it was now in a much better state to use for online listings, ads, and social media posts.

Talking-head videos

Talking-head footage is often one of the hardest types of video to improve because people naturally notice faces first.

Poor upscaling can easily make faces look unnatural. Skin may become too smooth, and facial features can end up looking “plastic-like”.

Vmake Labs includes Portrait and Portrait Pro modes to handle this type of footage. These modes focus on improving facial clarity while keeping the overall appearance natural.

In tests conducted, the portrait modes generated fairly reasonable outputs. There was an improvement in the clarity of facial details especially around the eyes, eyebrows, and hair areas while at the same time maintaining the natural skin texture. The enhancement made presenters appear sharper without creating the plastic-looking effect that some AI upscalers introduce.

Low-light footage

Dark videos are usually where older 720p footage struggles the most.

When lighting is poor, cameras capture less detail. Shadows become noisy, and the image loses clarity.

The Lowlight mode helped improve these clips by making darker areas easier to see while keeping the video balanced. It did not completely remove the limitations of the original recording, but it made the footage easier to watch.

Conclusion: Can AI turn 720p into usable 4K?

Yes, but with realistic expectations. 

After the test, the difference was easy to see. AI upscaling improved clarity and gave older videos a much sharper look on modern displays. However, it is still unable to create details that were never captured in original footage. It can make a 720p video look noticeably better — it cannot match the depth and natural detail of footage recorded in 4K. 

For older videos that are too valuable to discard but too soft to use as they are, AI enhancement offers a “middle ground”. Tools like Vmake can help refresh this type of footage without requiring a complete reshoot. It may not turn every old clip into true 4K, but it can make outdated videos much more usable for current platforms.