Understanding video compression

[cooljethro]

I'm having a hard time understanding how programs such as Handbrake make file sizes so much smaller?

I compressed a video of over 1.0 GB to about 0.09 GB, going from 1080p to 720p.

I thought to myself, surely going from 1080p to 720p couldn't make the video approximately ten times smaller?

So I tried converting a video from an original 1080p resolution to the 1080p preset in Handbrake, and still, the file size reduced by around 33%. What is the technical wizardry behind this? I don't understand it.

 

[Outdated]

Essentially what you need to read about for the technical details is Jpeg Compression because that's how it works.

To simplify - the computer sees an image as a single line of code. If there's a lot of pixels all in a line that are exactly the same colour then the computer can make the line of code shorter by saying repeat this colour 76 times. The trouble is that you don't get many pixels in a row that are all exactly the same colour so the computer doesn't get to save much space by that alone.

Compression works by letting the computer say "OK, all these colours are so close to each other that no one will notice the difference, I'll just make them all the same colour.". The more compression you include the less difference there is between all the different tones that make up the picture. You'll notice this first in darker areas of images and in people's skin tones.

Also, you mentioned that you downsized the image, so this is naturally going to take up a lot less space too.

 

[cooljethro]

Yeah, I understand that downsizing from 1080p to 720p would reduce the file size, but by approximately 10 times or more, seemed extreme.

Also, going from the native 1080p in a video and converting it via Handbrake's 1080p preset option still reduced the file size which is what I couldn't get my head around?

But your explanation does make things a bit clearer.

 

[abby normal]

Video compression reduces file sizes by removing redundant or imperceptible data using two main techniques: inter-frame compression (analyzing changes between frames) and intra-frame compression (compressing individual frames). Key methods include using efficient codecs like H.264 (AVC) or H.265 (HEVC), reducing video resolution, and adjusting the bitrate.

Primary Video Compression Techniques

• Inter-frame Compression (Temporal): This method analyzes differences between consecutive frames, storing only the changes rather than every full frame. It is highly efficient for video with little movement, such as interviews.
• Intra-frame Compression (Spatial): This method treats each frame as a separate image, compressing them individually by removing unnecessary pixels within that frame. It is useful for high-motion video.
• Hybrid Coding: Most modern codecs combine both temporal and spatial techniques to achieve maximum reduction in data.

 

[Shamar]

I'm having a hard time understanding how programs such as Handbrake make file sizes so much smaller?

I compressed a video of over 1.0 GB to about 0.09 GB, going from 1080p to 720p.

I thought to myself, surely going from 1080p to 720p couldn't make the video approximately ten times smaller?

So I tried converting a video from an original 1080p resolution to the 1080p preset in Handbrake, and still, the file size reduced by around 33%. What is the technical wizardry behind this? I don't understand it.

Each image at 720p is about 44% the size of the same image 1080p. There is much of your difference right there. blending colors as Outdated described accounts for the rest. High quality compression (more accurate colors) produces files much larger than low quality compression (more colors combined into one). Try experimenting with different compressions to understand the quality differences.