COMPRESSION
+
DIGITAL LANGUAGE

In an age where we increasingly spend more and more of our lives staring at screens, it’s easy to overlook the fact that these experiences use a fully digital medium to reach your device. Rather than ink or paint on paper, we see millions of microscopic lights shining through glass.

But how do our devices know what to show us?

And how does all this information traverse such great distances in practically an instant? To start, let’s ask ourselves:

What are digital files made of?

The Bit

Digital information is made up of ‘bits’ (portmanteau of ‘binary digits’). The bit can be either a 1 or a 0 (one or zero). The word binary means a system of duality, based on two states. This or that.

The bit is abbreviated as a ‘b’ (lower case).

So yes, digital information does look pretty much like The Matrix: seemingly infinite cascading seas of 1’s and 0’s.

Well, kind of.

That picture you took of your cat is actually millions of 1’s and 0’s. That video? Possibly billions.

Since bits are so small, they’re usually referred to in the context of megabits. One megabit is a million bits.

In this day and age, the most common context in which you’ll hear someone refer to bits is related to video and audio streams, in which the quality (the rate of the data stream) will be referred to in how many megabits per second (Mb/s) the stream is.

These bits make sense to a computer that knows how to decode the codec language used to write the bits, all of which we’ll look at in more depth down below.

The 1’s and 0’s are a binary system that make up any and all things that your computer can show or do. Other binary systems and ideas:

• On/off,

• True/false,

• +/- (positive/negative),

• Yes/no,

• Morse Code (dots + dashes).

Bits can be stored or transmitted as:

• Alternate states of an electronic switch (flip/flop),

• Alternate electrical currents or voltage,

• Alternate light levels (fiber optics),

• Alternate magnetic states,

• Pits and lands (on a reflective surface, such as a CD, DVD, Blu-Ray)

The Byte

A byte is a collection of bits. Technically, a byte can be just about any numbered set of bits, but 8 bits to a byte is the standard.

The byte is abbreviated as B (upper case).

The reason 8 bits typically make up a byte is because, historically, computer characters are described in 8 bits. Meaning any given character (number, letter, etc) was assigned some combination of 8 1’s and 0’s.

8 bits allow for up to 256 distinct characters. You can figure this amount out by taking the number of variables, 2 (1 or 0), and taking that to the power of the number of spaces, 8. (2^8).

2x2x2x2x2x2x2x2 = 2^8 = 256

Digital file sizes are almost always measured in bytes, not bits. Why? Perhaps because it makes the file sizes sound smaller than they actually are. An 800 megabit (Mb) file is the same as a 100 megabyte (MB) file.

More practically, it takes a byte to describe a character or value, so the total number of bits is a bit irrelevant. Measuring by bytes then becomes a more accurate representation of how much data is or can be stored.

Orders of MagnitudE

Decimal System (+ Prefixes)

Working in sets of 8 makes sense to computers, but it does cause a few inconsistencies when drawing parallels with how we normally deal with numbers, which is in a base-10 system.

Below are base-10 orders of magnitude. This is what’s called the decimal system.

1,000 (thousand) = kilo

1,000,000 (million) = mega

1,000,000,000 (billion) = giga

1,000,000,000,000 (trillion) = tera

1,000,000,000,000,000 (quadrillion) = peta

Decimal vs Binary

Computers, however, need to store whole bytes, not just bits, so the orders of magnitude need to be divisible by 8. This is called the binary system.

When approximating the binary system with the decimal system, the two systems diverge with increasing orders of magnitude.

Kilo - 1,000 (decimal) vs 1,024 (binary, 2.4% difference)

Mega - 1,000,000 (decimal) vs 1,048,576 (binary, 4.86% difference)

Giga - 1,000,000,000 (decimal) vs 1,073,741,824 (binary, 7.37% difference)

Tera - 1,000,000,000,000 (decimal) vs 1,099,511,627,776 (binary, 9.95% difference)

Peta - 1,000,000,000,000,000 (decimal) vs 1,125,899,906,842,624 (binary, 12.59% difference)

What is data compression?

Data compression is taking digital information (bits) and encoding it in a way that uses fewer bits than the original data, and is a smaller file size as a result.

Think of it like using contractions, acronyms, or more efficient wording to re-write a sentence in a more concise manner (while trying to keep the message as true to the original as possible).

The goal of data compression, from a technical viewpoint, is to use fewer bits.

Finding patterns and redundancies in the source data is a key to compression. Compression then expresses these patterns or redundancies in a more concise, efficient way, with fewer bits.

High levels of compression manipulate the data to create patterns and redundancies where there weren’t any before. This causes compression artifacts, such as banding and macro blocking. More on that below.

Data compression is a branch of the fields of computer science and information theory, and is continually advancing, which leads to the release of newer, more efficient codecs and file types.

Files that utilize different formats (wrappers) and codecs are using different forms of, or approaches to, data compression.

Why compress data?

Many high fidelity digital files are very large in size. This makes it difficult or impossible to:

• Write files to a medium (SD cards, etc) at a sufficient rate

• Store files in significant amounts

• Edit or render with the average consumer computer

• Stream, due to bandwidth limitations

When compressing, there are three major variables to consider:

• The desired resulting quality of the file

• The desired resulting size of the file

• How easily the compressed file can be accessed or read

In other words, the point of data compression, from a user perspective, is to make a file smaller and more manageable.

Many new codecs are processor intensive, and all have to be adopted by operating systems and programs before they can be of any practical widespread use.

For example, the H.265 (HEVC) codec has been around since 2013, but has only recently reached a point where some common programs, such as Finder in MacOS, are able to decode it in real time (albeit imperfectly).

What’s a Codec?

In short, the codec is the compression language. It’s the method of how digital information, such as pixels, are described in bits.

‘Codec’ is a portmanteau of ‘coder-decoder’.

Different codecs use different ways of describing information. Generally speaking, more efficient codecs - ones that use the fewest bits to describe the same amount of information - are harder to code and decode, and vice versa.

Zip files, for example, are a well established, efficient form of compression, as it relates to file size, but they take time to zip and unzip.

Codecs each have their own specific limitations, which is a consequence of their compression efficiency.

The most common codecs, for example, ‘only’ support 8-bits of color and 4:2:0 chroma subsampling:

• The most common image codec, by far, is JPEG.

• The most common video codecs are h.264 and various versions of Apple ProRes422.

Digital Container
(Wrapper, or Format)

The container is the file type extension at the end of the filename. Sometimes these directly coincide with the codec (JPG), but some wrappers have several possible codecs (MOV).

The container is how your computer, and you, know what to do with a file, and how to open it.

The most common video containers

• .MP4

  • Codec: h.264

• .MOV

  • Common codecs: h.264, Apple ProRes422, and more

The most common audio containers

• .MP3

• .WAV

The most common still image containers

• .JPG

• .PNG

Two major types of data compression

Lossless

Lossless compression retains all of the source information while also writing the information in fewer bits.

There is no loss of information, even though the information has been compressed. In other words, the compression process does not affect the quality of the image, video, or sound.

Examples:

• TIFF (16-bit)

• PNG (16-bit)

• FLAC

• RAW

There are several different forms of RAW file types, from photo formats to video formats. In general, a RAW file is one that contains lossless information that still needs to be processed to create an image and/or sound. Thus, it’s raw.

Here are some RAW photo formats:

• DNG, Adobe Digital Negative

• RAF, Fujifilm

• RW2, Panasonic

• NEF, Nikon

• ARW, Sony

• CRW, Canon

And some RAW video formats:

• ARI, Arri

• R3D, Red

• Apple ProRes RAW*

• BRAW*, Blackmagic

*NOTE: Apple ProRes RAW and Blackmagic Raw are arguably not true raw file types - as there is some (supposedly imperceptible) loss of total information - but they are both forms of near-lossless compression. In the case of Blackmagic Raw, it’s also highly efficient. Apple ProRes RAW? Not so much.

Lossy

In lossy compression, there is a loss of total information. This might be a reduction of the number of pixels or sampling rate, or perhaps just loss of information due to imperfect, excessive and/or inefficient compression.

Ultimately, the compression process affects the theoretical quality of the image, sound, or video, and the compressed information is technically inferior to the source data.

However, lossy compression is informed with what we know about psychovisuals and psychoacoustics so as to minimize perceptual losses in quality. In other words, even though the resulting file is technically inferior, you might not be able to see or hear a difference, depending on how much the data has been compressed.

Lossy compression is by far the most common compression type. Everything we see and hear on the consumer end is a type of lossy compression. This includes the internet, streaming, and home audio/video.

Examples:

• JPG

• MP4 (H.264)

• MOV (H.264, ProRes)

• H.265

• MP3

• WAV

COMPRESSION ARTIFACTS
(SIDE EFFECTS)

If too much compression is applied to an image or video, the process creates compression artifacts, which are visible side effects of the compression process.

The two most prominent kinds of compression artifacts are macro blocking and banding.

MACRO BLOCKING

Macro blocking is when large blocks of the image are written as the same color, rather than the subtler gradients of the source image.

You’ll first notice macro blocking on large portions of the image with similar colors, such as a wall or a clear blue sky. As you increase compression, macro blocking occurs on more and more portions of the image that don’t have significant contrast or detail.

BANDING

Banding is a term for a specific kind of macro blocking that occurs on and around gradients. Rather than the artifacts taking a block, or square-ish, shape as they often do with a flat-lit wall or sky, banding is artifacts that form consecutive ‘bands’ where there otherwise would be a gradient. These look like concentric circles or stacks of macro blocks.

Bitrate
(Bandwidth)

Bitrate is the amount of bits written or transmitted per second, and is usually measured in Megabits/second (Mbps).

This is typically only relevant to video and audio streams, although it is used to measure the bandwidth (capacity) of a data connection as well, such as internet speed, etc.

Some relevant bitrates (as of March 2021):

• YouTube (1080p upload) - 8 Mbps

• YouTube (2160p upload) - 40 Mbps

• Hulu (1080p) - 6 Mbps

• Hulu (2160p) - 16 Mbps

• Blu-ray (Full HD) - 40 Mbps

• 4K UHD Blu-ray - 100 Mbps

• Netflix is constantly messing with the bitrates of their streaming service, which can be different from device to device.

Constant bitrate vs Variable bitrate

There are two different forms of compression when it comes to bitrate: constant bitrate and variable bitrate.

Constant bitrate is a consistent data stream, and is unaffected by the subject matter being streamed. If the constant bitrate is 40 Mbps, then there will always be (about) 40 megabits of information every second, no matter what images and sounds are being compressed or streamed.

Variable bitrate fluctuates, however, depending on the information need of the subject material. This is the more common method, as the bitrate can drop to significantly lower levels if there isn’t a need for additional data, which allows either for the total file size to be smaller, the stream to buffer further ahead, or both.

For example, if a video dips to a black screen for several seconds, there isn’t a substantial data need to refresh pixels, so the bitrate can drop well below the target bitrate without sacrificing any quality. On the other hand, if there’s a sudden need for more information to adequately describe a detailed scene, the bitrate can temporarily go higher than the target bitrate, up to the specified maximum bitrate.

In setting up a variable bitrate export, you set a ‘target bitrate’, which acts as a rough average. You also will set a ‘maximum bitrate’, which, as you’d expect, places a ceiling on how high the bitrate can ever get.

Types of image compression

Intra-frame
(All-i)

Intra-frame is compression within a single image, which is also known as spatial compression. Intra-frame compression looks at the whole image, and uses identical and like pixel values (hue, saturation, lightness) to create efficiency shortcuts in the compressed data.

Still images inherently use intra-frame compression, since there’s only the one frame to compress.

In video, however, there are two major ways to compress the data: intra-frame and inter-frame.

In intra-frame compression, each frame is compressed individually. This is referred to as ‘All-I’ compression, which means that all frames utilize intra-frame compression, as opposed to inter-frame (below).

Inter-frame
(Long gop, or ipb)

Inter-frame applies compression across frames of video, and is also known as temporal compression.

Inter-frame compression is mostly concerned with what changes from frame to frame, and is a much more efficient form of compression than intra-frame.

Inter-frame compression is referred to as either ‘Long GOP’ (GOP meaning ‘group of pictures’, NOT ‘grand old party’…) or IPB.

IPB is in reference to how compressed frames behave in relation to one another.

• I-frames use intra-frame compression to record the information of the full frame, either as a starting point or as a sort of checkpoint. An I-frame is considered an ‘anchor frame’ because the information is rooted in what the camera is actually seeing.

• P-frames are predictive frames that mostly use what information is gathered from the I-frames and/or P-frames that preceded it, and only record what has changed. A P-frame is also considered an ‘anchor frame’, since at least a portion of the image is true to what the camera is seeing.

• B-frames are bipredictive frames that act in a similar manner to P-frames, but reference both preceding frames and subsequent frames. The referenced frames must be either an I-frame or a P-frame. B-frames are not anchor points because, loosely speaking, they are the processor’s best guess at what is likely happening at that moment in time.

This is an example of what an IPB pattern might look like:

IBBPBBPBBPBBIBBPBBPBBPBBI

The IPB pattern can be altered to have different lengths between I-frames, and between anchor frames (I or P). Different forms of IPB compression are defined by two traits, dubbed the M and N numbers. The M number is the distance between the anchor frames. The N number is the distance between I-frames.

In the above example, M=3 and N=12. This is a common form of Long GOP compression.

Broadly speaking, the more I-frames you have, the more ‘true’ your footage is to what the camera is seeing, but fewer I-frames means a more efficient compression.

All-i vs Long GOP

It’s common for (wishfully) discerning videographers to default to All-I when selecting a compression type for shooting, but should it really be such an obvious choice?

Spoiler alert: No.

The argument for All-I is an idealistic one. It’s a compression type that takes exactly what is coming out of the camera’s processor and compresses it, frame-by-frame. It’s pure and exactly reflective of what you’re shooting, right? Well, sure, if you believe that what’s coming out of your camera’s processor is perfectly reflective of reality, which it isn’t.

Quick aside: lens character and characteristics (imperfections), processors, and picture profiles do a lot to separate reality from images. This is actually something that I think is a good thing, and should be embraced. It’s much of what makes photography a subjective art.

In real world use, processors are sophisticated enough that, at equally sufficient bitrates, you will likely never see the difference between All-I footage and Long GOP footage, even if you tried.

To ever maybe notice a difference, you would have to shoot a very particular kind of chaotically moving subject with both compression types, back-to-back, and you would have to know exactly what to look for. Even then, you’d only be able to tell that the footage looks slightly different, much less being able to discern which looks ‘better’ or more realistic, or even which footage is which.

That’s a whole lot of ifs, onlys, and maybes. Anyone who could actually tell the difference would be a discerning videographer, indeed.

As previously mentioned, Long GOP compression is more efficient than All-I. In practice, this means that, at any given bitrate, Long GOP footage is less likely to exhibit the side effects of compression, such as macro blocking and banding, and is generally more robust.

Most cameras won’t even allow you to shoot All-I footage, and those that do won’t allow you to shoot lower bitrate footage in All-I, because the compression type isn’t efficient enough to create quality footage.

Just how much more efficient is Long GOP? That depends on the number of I-frames recorded per second, but it’s generally several times more efficient than All-I. For example, 100Mbps Long GOP footage is often just as robust as 200-400Mbps All-I footage.

This is why the most common delivery codec, h.264, utilizes Long GOP. In fact, almost everything you have ever watched on your computer or television has had inter-frame, Long GOP compression.

The only real drawback to Long GOP is that it’s more processor intensive to play back and edit the footage, since the computer has to basically assemble the footage rather than just playing back a series of images.

The way I see it: If you can’t see the difference between All-I and Long GOP, and the latter is more efficient and less likely to exhibit compression artifacts, why wouldn’t you just save yourself the storage space?

Color

HEXADECIMAL CODE

The hexadecimal code is a way of writing out color values, and is based on the RGB system.

Hexadecimal code - or hex code, for short - is ubiquitous in the digital color realm, and you’ll find it everywhere from Photoshop and Illustrator to web browsers. You’d be correct then if you assumed that it utilizes an 8-bit color space, which as mentioned has long been the norm for how colors are displayed.

Hexadecimal code uses the numbers 0 through 9 and the letters A through F to represent color values, with 0 being the lowest and F being the highest, so the order goes:

• 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, F

If you count the number of characters, or variables, there’s 16.

When writing out Hexadecimal code, there are six characters for any given color value, with two characters representing each color channel: red, green, and blue. The first two characters represent red; the middle two, green; and the last two, blue.

Since each character has 16 variables, multiplying the two together allows 256 unique combinations, which serve to represent the color values of each color channel.

The value order goes like this: 00, 01, 02, and so on until you reach 0F, then the next values become 10, 11, 12, and so on, up to FF.

Let’s look at an example:

• 35B25E

The above is the hexadecimal code for a nice blueish-green color. Broken up, the channel values are: 35, B2, 5E. In this example, the red value is 35, which is fairly low. The green value is B2, which is in the upper half of values. The blue value is 5E, which isn’t high, but higher than the red value. So when you combine these channel values, you get a color that’s predominantly green with a notable touch of blue. The red value only serves to move the overall color closer to greyscale, which lowers the saturation.

Color Bit-Depth

In short, color bit-depth is how many colors are possible in a color space. Above all else, a higher bit-depth means there are more shades of luminance and saturation possible.

Before we get into the digital conventions, let’s look at the human science. While it’s impossible to have exact numbers, various sources figure that the human eye (and brain) can perceive anywhere from 7 to 12 million unique colors, averaging about 10 million. Naturally, there are individual variations to consider in these figures as well.

With that in mind:

The standard for practically the entire history of digital color has been 8-bit color, which means each color channel - red, green, and blue - can be represented in 2^8 different shades, or 256 shades per channel. Since colors are created by blending the channels, there are 16.78 million possible colors in an 8-bit color space; 256 times 256 times 256.

Most camera sensors ‘see’ in 12-, 14-, or even 16-bit color (even if most cameras aren’t capable of the dynamic range necessary to fill these color spaces). Raw photo files retain the full color space that the sensor reads.

While these spaces are exponentially larger than what is perceivable to the human eye, there are advantages to having excessive color information, especially as it relates to recovering information in the highlights and shadows, or when applying sophisticated color grades.

After editing, however, you’ll want to post your beautiful photo somewhere, at which point you’ll likely export a jpeg file (JPG), which operates in the standard 8-bit color space.

Video is another story. Most consumer cameras still ‘only’ record 8-bit video files. While 8-bit is certainly sufficient for delivery, starting with an 8-bit image can be limiting, especially as it relates to accurately rendering high dynamic range, recovering information from highlights and shadows, and sufficiently rendering gradations in an altered image. The biggest mistake to be made here is capturing 8-bit Log footage. Just because your camera can, doesn’t mean you should.

That being said, you can still create stunning images from 8-bit video files, you just have to keep these limitations in mind. When recording 8-bit video, the best approach is to try to achieve a look in-camera that is as close as reasonably possible to what the final image will be.

With video, the most common export, or ‘delivery’ type is a .MP4 file with h.264 codec, which works in an 8-bit color space with 4:2:0 chroma subsampling.

10-bit?

The consumer market is pushing 10-bit color, from the capture (camera) side to the display side.

The 10-bit color space contains 1,024 shades in each color channel, totaling almost 1.074 BILLION colors. Keep in mind that humans see about 10 million colors, which is just 1% of one billion. This begs the question: Is there really a need for higher bit-depth?

There are advantages to starting with an image with a higher bit-depth, especially if you want to start with a more flexible image for editing and coloring, such as a Log profile. So yes, higher bit-depths (like 10-bit) can certainly help on the acquisition and editing side.

On the display (delivery) side, however, 10-bit and higher doesn’t make much logical sense, considering science all-but-confirms that we literally can’t see the difference in color.

There are plenty of people out there that would disagree with me, however, with the best evidence being the visual differences between standard Blu-rays and 4K UHD Blu-rays, which can support 10-bit display color. The visual difference, however, is because many movies have distinctly different finishing grades for 8-bit and 10-bit displays in mind, with 10-bit versions usually having subtler contrast levels as part of the marketing to sell a visual difference.

This leads to the reality of the 10-bit display push: Companies are trying to sell you TVs.

Resolution bumps alone just aren’t enough to make the sale anymore, so they’ve invented a new marketing tactic: HDR.

One area where these HDR displays actually are beneficial is that many of them have very high contrast pixels, or diodes. This allows them to be very bright or very dark, with minimal bleed between pixels. That said, this isn’t a trait exclusive to 10-bit HDR displays, but rather a separate technological advancement that frequently goes hand-in-hand with the 10-bit displays, so as to perceptually maximize the imperceptible ‘benefit’ of all those extra colors.

All that said, there’s no real drawback to a 10-bit display over an 8-bit display. Even if you can’t perceive the difference, they are still technically ‘better’, so why not, right?

DOES 8-BIT CAUSE BANDING?

The problem: Many images we see on social media or stream to our TVs or phones exhibit noticeable, sometimes distracting, levels of banding and macro blocking. Remember those?

There’s a very common misconception that 8-bit color is the culprit. This doesn’t make a lot of sense when you realize that probably everything you’ve ever seen on a display is in 8-bit color, and not everything you’ve seen has banding and macro blocking.

What, then, causes these artifacts? A very relevant thing indeed: Compression.

Remember that compression can create patterns and redundancies where there aren’t any, so when a detailed image or an image with subtle gradations goes through the process of compression, the encoder saves space by lumping similar colors together, which results in what we see as banding.

Many social media platforms, with various blends of images and videos, are data hogs.

In a world where many consumers still pay for data by the gigabyte, this is a problem. If scrolling social media starts costing customers $50 a month in data, customers would eventually wise up and ditch the social media. Alright, probably not, but I’d like to think so anyways.

The companies’ solution? Compression. Every social media company has its own file size limits and compression methods. It’s these aggressive and often inefficient compression methods that make every photo and video look awful, not an 8-bit color space.

The demand for higher resolutions exacerbates these issues.

Most Blu-rays - which, by the way, have 8-bit color spaces - don’t show much banding, if any, and the reason is because they have a sufficiently high 40 Mbps bitrate for their 1920x1080 video. By contrast, streaming services use only 5-8 Mbps to represent a 1920x1080 video. In other words, there is significantly more compression applied to the source file, and fewer bits trying to describe the same image.

4K UHD Blu-rays help perpetuate this misconception, due to their subtler gradations and 10-bit color space, which don’t show much banding, if any. The difference, however, is that they have a bitrate around 100 Mbps; much higher than other UHD sources like YouTube (40 Mbps), Hulu (16 Mbps), and Netflix (25 Mbps). The lack of banding, then, is really a result of there being less compression applied compared to streaming services, not the bit-depth of the color space.

Chroma Subsampling

Using the science of psychovisuals, people long ago figured out that humans are much more perceptive of changes of luminance (brightness) than they are changes of chrominance (color).

This information was applied to the compression process by reducing the color detail within an image to some fraction of the luminance detail. This process is called chroma subsampling.

Chroma subsampling is written and defined as 4:x:y. In any two rows of four pixels (4 by 2), x represents the number of potentially unique colors in the top row, and y represents the number of potentially unique colors in the bottom row, relative to the top row.

Almost every image we ever see is compressed with 4:2:0 subsampling, which means that in the top row of four pixels, there are only two unique colors (x). These are always equally spread out, so the first two pixels on the left share a hue, and the two on the right share a hue. Each pixel has its own unique luminance, however. There are no unique colors in the bottom row (y), so they share the same hue as the pairs of pixels above them. Ultimately, this means each unique hue occupies a 2x2 space, while luminance is unique for every pixel, so there’s half as much color detail in each direction.

While chroma subsampling can make a visible difference at lower resolutions, increasing the overall resolution makes it more difficult, if not impossible, to visually perceive a difference between 4:2:0, 4:2:2, or even 4:4:4 (the latter of which has the potential for unique color information per pixel). For example, it’s relatively easy to perceive the difference in 4:2:0 and 4:4:4 at resolutions like 480p, but at UHD (2160p), the difference is arguably imperceptible.

From a realistic perspective, most color patterns aren’t nearly fine enough for chroma subsampling to make a visual difference, especially in the present high-resolution landscape. And as people continue to push for higher resolutions, chroma subsampling will continue to matter less and less. This is especially true considering many lenses aren’t capable of resolving enough detail for 4K, much less 6K or 8K, so cutting hairs at the pixel level is a rather pointless argument.

As people continue to push for higher resolutions that greatly exceed our visual perception, we will likely see chroma subsampling that becomes even more aggressive than 4:2:0.

Perhaps 6:2:0:0? If each hue occupies a 2x2 space in 4:2:0, how large would the hue area be in 6:2:0:0?

APPLICATIONS?

If I’m being honest, I was taking photos and making videos long before I understood the vast majority of this information, with little to no consequence. Much of this is technical minutiae that can help you create a better quality product, if you know what to leverage in your favor.

Many apps and software, however, have curated presets that help the laymen among us navigate which technical traits are appropriate for various applications.

Adobe Premiere has presets for YouTube, Vimeo, Twitter, etc, which have the proper codecs and approximate bitrates, among other things, that are optimized to play nice with these platforms. This helps to upload your content in a timely manner, reduce processing on the other end, and maintain a certain threshold of quality for your works.

Much of this information, however, is important at earlier stages of production, such as when you are deciding which format to shoot in, or which codec you should use as an intermediary when sharing files with others that you are working with.

Above all else, however, it’s something to be learned, and will help you to stand out as someone who genuinely knows what they are talking about, which is much more uncommon than you might think.