Lossy and Lossless Compression: The Difference

What is video compression?

Video compression is the elimination of data to reduce file size. The word ‘elimination’ is used with full intent. When the data is eliminated it is gone forever.

True Lossless Compression

When a file that has been compressed can be returned back into its original form with zero loss of information, the compression is a True Lossless Compression.

If you’ve found my content useful support me and watch my film A Count Of Ten today.

CLICK FOR STREAMING OPTIONS

Is this just a pipe dream, or can true lossless compression really be achieved?

Here are two scenarios where this is possible:

• Vector images are scaled based on mathematical principles. Simple shapes can be made quite small in file size. However, complex shapes can be larger than raster images.
• An algorithm that stores duplicate pixel data can replace them when prompted. Just like vectors though, once the image gets complicated, the level of compression will become negligible.

Since films and video are made of raster images, let’s understand compression with a simple example.

The Idea behind Compression

Let’s take a teeny-tiny image with a resolution of 256 x 256.

We’ll assume this image is in 8-bit color. The file size can be calculated to be 256 x 256 x8 x 3 = approx. 197 KB (KiloBytes).

What if this image is totally black, end to end? Each pixel has a value of 00000000. A smart algorithm can:

• Study the file and decide that each pixel value can be changed to just 0, and
• Add metadata that explains the changes it made. E.g., it can add 00000000 in the metadata to show what each pixel value should be.

A decoding algorithm already expects instructions in the metadata. It will know that:

• 00000000 represents black that has to be used for all pixels, therefore
• Each 0 is to be replaced by 00000000

Theoretically, the decoding algorithm, can perfectly recreate the original image without any kind of loss.

How much can we reduce the file size by taking out all the extra zeroes and adding a simple instruction?

The file is now (256x256x1)+8 = 65,544 bits, which is about 8.2 KB.

That’s a 24x reduction!

This example is, of course, a highly simplistic view of the compression process. Computer engineers and mathematicians have found much more complex and brilliant ways to compress data.

It’s purely a mathematical idea, even with LLMs.

Most compression algorithms don’t understand the subject matter being presented. They apply generalized formulas and systems to reduce file size.

Is there a True Lossless Compression in Film?

Actually, yes. Arri and Codex have developed a system called HDE:

HDE stands for High Density Encoding. It is an encoding technique that is optimised for Bayer pattern images. ARRIRAW images encoded with HDE are approximately 60% of the original size. HDE encoding is completely lossless – when an HDE file is decoded, it is a bit-for-bit perfect match to the original file.

Codex HDE

The above image shows how much data you can save with HDE. At the moment you have to use software to convert ARRIRAW files to HDE (*.arx) format. This format is recognizable by software like Baselight, Resolve, etc.

Lossy Compression

If, after compression, the original file cannot be brought back again (like humpty dumpty), then the compression is said to be Lossy.

As far as video codecs are concerned, I have yet to see lossless compression. Name your codec – it is lossy.

Compression in video is always lossy.

What about “Visually Lossless” Compression?

The aim of every compression algorithm is to achieve a visually lossless look. I

Visually lossless just means you can’t tell if it’s compressed or not just by looking at it. Whatever you stream on YouTube, Netflix, etc., are all compressed to the lowest possible data rate current technology allows.

However, most people who watch the same movie on Netflix or in cinemas won’t be able to tell the difference. Even professionals will have a hard time if it isn’t the film they shot themselves.

To help you understand it, let’s start with a test image.

I created a test TIFF file that was compressed to JPEGs under five settings – Maximum (12), High (8), Medium (5), Low (3) and Zero (0).

The original TIFF file has a resolution of 1920 x 1080, is 8-bit, with a file size of 5.95 MB.

Here are the compressed images compared (The numbers in orange show you the size in KB (5.95 MB = 6096 KB). The numbers in blue show the size relative to the original size):

What do you see? Here’s what I gather:

• After a point, compression makes the image worse, but does not reduce file size proportionally.
• Large expanses of color get worse faster than fine detail. The first and last boxes show degradation at level 8, while the fine detail holds up till level 5. One of the fundamental tenets of compression is the elimination of duplicate data. Fine detail makes that difficult.
• Once fine detail has been compressed to the limit, compression artifacts give birth to patterns of their own, giving the illusion of detail and sharpness.
• The maximum quality setting for JPEG is visually lossless, while at the same time being only 10% the size of a full raster image.
• If the quality of the original image is poor, even the best compression level will contain artifacts. Maybe the source footage will have artifacts, too!
• Level 12 has a size of 575 KB – visually lossless JPEG at maximum quality. If you had a JPEG image sequence at 25 fps, it would result in a data rate of 112 Mbps.

You can perform similar tests on any algorithm that claims to reduce file size. Some algorithms are designed to stop after a certain point – giving the illusion that they are ‘better’ somehow. Really cool codecs like JPEG gets bad publicity from misinformed souls who don’t know when to stop!

Why are the data rate requirements of inter-frame codecs smaller than intra-frame codecs?

If you don’t know what inter-frame and intra-frame codecs are, check out this article:

If you study broadcast quality specifications, you’ll see:

• The data rate requirement for inter-frame codecs is 50 Mbps, while
• The data rate requirement for intra-frame codecs is 100 Mbps.

Aren’t inter-frame codecs worse? After all, in cameras we’re encouraged to use intra-frame codecs (called ALL-I sometimes) for better image quality.

Shouldn’t they need more space if this were true?

Inter-frame codecs like H.265, H.284 and AVCHD are better at compression than intra-frame codecs!

In other words, Inter-frame codecs can provide similar visually lossless results in a lesser file size. Why it’s not recommended to film with them, is because any further processing (like color grading), will make it a lot worse. Intra-frame codecs hold up much better to grading and manipulations.

What about Common Camera Compressed Formats?

Here’s a table comparing multiple formats with a simplified “intended for” column:

How to choose the Right Compression

Keep it simple:

• If you want to color grade, pick RAW if possible. Otherwise, pick the highest data rate intra-frame codec you can afford to film in.
• If you are making a video straight for YouTube or social media, you can film in inter-frame codecs. Some software like Davinci Resolve allows you to bypass further encoding if the source and destination codecs are similar.
• If you’re using a camera that has no option but an inter-frame codec, try to get the look down in camera, so you don’t have to color grade later. With care, you can do some level of color correction work on inter-frame codecs.
• If your delivery spec mandates an intra-frame codec, then pick that strictly!

I hope this helps. You needn’t be concerned by all the myriad options available in your camera.

When in doubt, pick the one with the highest data rate!