Video File Formats and Compression: The Complete Guide

Video is the only file type most people handle regularly without understanding any of it. A phone produces a 400 MB clip from ninety seconds of a birthday party. A screen recording refuses to upload. A client sends a MOV that will not open on a Windows laptop. An export that looked perfect on a monitor turns into a blurry mess after a social platform gets hold of it. Every one of those problems traces back to the same small set of ideas: containers, codecs, resolution, frame rate, and bitrate.

This guide covers all of them in plain language, then turns the theory into practical decisions. By the end you should be able to look at any video file, understand why it is the size it is, know which format to convert it to, and compress it to a target size without destroying the picture. No software purchase is required, and none of the steps involve uploading your footage to a company you have never heard of.

Containers vs Codecs: What a Video File Really Is

The single most useful thing you can learn about video is that a video file is not one thing. It is a box holding several separate streams, and the name on the outside of the box tells you almost nothing about what is inside.

The container

The container is the file format you see in the extension: .mp4, .mov, .mkv, .webm, .avi. Its job is organisational, not creative. It stores the video track, one or more audio tracks, subtitle tracks, chapter markers, and metadata such as the recording date, camera model, and sometimes GPS coordinates. It also stores the index that lets a player jump to the twelve minute mark instantly instead of reading the whole file.

Containers differ in what they are allowed to hold and how widely they are supported. MKV can hold almost anything, including several audio languages and a dozen subtitle tracks. MP4 is stricter but plays everywhere. That trade between flexibility and compatibility is the entire story of container choice.

The video codec

The codec is the compression scheme applied to the picture. Raw video is absurdly large: an uncompressed second of 1080p at 30 frames per second is roughly 180 MB, which means one uncompressed minute would be over 10 GB. Codecs get that down by two or three orders of magnitude using two tricks.

The first trick is spatial compression, which works within a single frame much like JPEG does, discarding fine detail the eye is bad at noticing, especially in colour. The second and far more powerful trick is temporal compression. Most frames in a video are nearly identical to the frame before them. Rather than storing every frame, the codec stores occasional complete frames (keyframes) and then stores only the differences in between, described as blocks of pixels that moved in a particular direction. A locked off shot of a person talking compresses beautifully because only the mouth changes. Confetti falling, water, fire, and fast camera pans compress terribly because nearly every pixel changes every frame.

The audio codec

Audio gets its own separate compression, usually AAC in MP4 files, Opus in WebM, and sometimes AC3 or MP3 in older files. Audio is a small fraction of a video file, typically 128 kbps to 256 kbps against a video track that might be 8,000 kbps, so squeezing audio is rarely worth the quality loss. When you compress a video, leave the audio alone unless it is genuinely enormous.

Why the file extension lies to you

Because container and codec are independent, two files with the same extension can behave completely differently. An .mp4 holding H.264 opens on a fifteen year old laptop. An .mp4 holding AV1 may refuse to play on a device from three years ago. An .mov from an iPhone might contain HEVC that a Windows machine will not decode without an extra component, which is why the file appears broken rather than simply large. When someone says a video does not work, the container is rarely the problem. The codec inside it usually is.

The Five Numbers That Decide Quality and File Size

Every quality decision in video reduces to five numbers. Learn what they do and file size stops being mysterious.

1. Resolution

Resolution is the pixel grid: 1920 by 1080 for 1080p, 3840 by 2160 for 4K. Doubling the resolution label quadruples the pixel count, because both width and height double. 4K is four times the pixels of 1080p, and 1080p is roughly nine times the pixels of 480p. Resolution does not directly set file size, but more pixels need more bits to look sharp, so raising resolution pushes the required bitrate up steeply.

2. Frame rate

Frame rate is frames per second: 24 for a cinematic look, 30 for standard video, 60 for smooth motion and gaming, 120 or 240 for slow motion source footage. Going from 30 to 60 roughly doubles the amount of picture information per second, though good codecs claw some of that back because consecutive frames at 60 fps are more similar to each other.

3. Bitrate

Bitrate is the amount of data used per second, measured in megabits per second (Mbps) or kilobits per second (kbps). This is the number that actually determines file size, and it is the one most people never touch. The relationship is simple arithmetic:

File size in megabytes equals bitrate in Mbps multiplied by duration in seconds, divided by 8.

A five minute clip at 8 Mbps is 8 times 300 divided by 8, which is 300 MB. Drop that to 3 Mbps and you get 112 MB. That single calculation explains almost every oversized file you have ever met.

4. Bit depth and colour subsampling

Bit depth is how many shades each colour channel can express, usually 8 bit for delivery and 10 bit for professional footage. Chroma subsampling describes how much colour detail is thrown away relative to brightness detail, written as 4:2:0, 4:2:2, or 4:4:4. Human eyes are far more sensitive to brightness than to colour, so nearly all delivery video uses 4:2:0 and nobody notices. These settings matter when grading footage, and matter very little when exporting a finished clip.

5. Duration

Duration is the number people forget is adjustable. It multiplies directly against bitrate, so cutting a four minute video to ninety seconds removes 62 percent of the file with zero quality loss anywhere. No compression setting can compete with that.

MP4 vs MOV vs WebM vs MKV vs AVI

Container choice is mostly a compatibility question. Here is the honest summary of when each one earns its place. If you want the long version, with the history behind each format, what it can and cannot store, and how to diagnose a file that refuses to open, read Video File Formats Explained.

MP4 is the default and should stay that way

If you take one thing from this section, take this: when in doubt, export MP4 with H.264 video and AAC audio. It is not the most efficient combination available today, and a modern codec could produce the same picture in a smaller file. It is however the only combination you can hand to a stranger with total confidence that it will play. Compatibility is worth more than a 30 percent size saving in almost every real situation.

MOV is a working format, not a sharing format

MOV is Apple's container and is technically a close cousin of MP4. Editors like it, iPhones record to it, and ProRes lives inside it. The friction comes when a MOV leaves the Apple world. A ProRes MOV can be ten to twenty times larger than the equivalent H.264 MP4, and an HEVC MOV from a recent iPhone may not decode on a colleague's older PC. Keep MOV for capture and editing, then convert to MP4 for delivery.

WebM belongs on web pages

WebM pairs VP9 or AV1 video with Opus audio and is designed for browser streaming. For a background video on a landing page or an auto playing product loop, WebM can cut the file size meaningfully against H.264 at the same visual quality, which is real money in bandwidth and real seconds in load time. Outside a browser it is awkward, so it is a publishing format rather than a working one.

MKV is for keeping things, not sending things

MKV is the most capable container in common use. It will hold multiple audio languages, a stack of subtitle tracks, chapters, and attachments, and it does not care which codecs it wraps. That flexibility makes it excellent for archiving a film collection and useless for uploading anywhere, since almost no web form accepts it. Converting MKV to MP4 is often just a repackaging job rather than a re-encode, because the H.264 stream inside can frequently be copied straight across.

AVI should be converted and retired

AVI dates from 1992. Files in it are usually large, often use codecs modern software does not include, and sometimes carry audio sync problems. If you have AVI files worth keeping, convert them to MP4 once and archive both while you confirm the conversions play correctly.

The Codecs You Will Actually Meet

Codec choice is where the efficiency gains live. Newer codecs use more computation to describe the same picture with fewer bits, which is why they save space and why they take longer to encode.

The pattern is consistent: efficiency costs compatibility and encoding time. A ten minute 1080p clip that takes two minutes to encode in H.264 might take eight minutes in AV1 for a file half the size. If you are storing hundreds of hours, that trade is obviously worth it. If you are sending one clip to a client this afternoon, it is obviously not.

One nuance worth knowing: HEVC and AV1 are most impressive at low bitrates. At generous bitrates all modern codecs look fine, and the advantage of the newer ones narrows. The savings show up precisely where quality is under pressure, which is why streaming services care about them so much.

How to Compress Video Without Wrecking It

Compression is not one lever. It is four, and pulling the right one for your situation is the difference between a clean small file and a smeary one. For the full treatment, including CRF reference tables, per platform size targets, and how to work backward from a hard file size limit, see How to Compress a Video Without Losing Quality.

Lever 1: cut the duration

Always first, always free. Remove the setup at the start, the fumbling at the end, and any dead air in the middle. A video that is 40 percent shorter is 40 percent smaller with no quality cost at all, and it is also a better video.

Lever 2: lower the bitrate

This is the main compression control. Halving the bitrate halves the file. The question is how far you can go before artefacts appear, and the honest answer is that it depends entirely on the content. Talking heads, slide presentations, screen recordings, and static product shots survive aggressive bitrate cuts. Anything with heavy motion, grain, water, foliage, or confetti falls apart quickly, showing blocky patches in shadows and a mushy look during movement.

A reliable method is to compress to a target and then check the worst moment in the clip rather than the easiest. If the busiest second still looks acceptable, the rest certainly does.

Lever 3: reduce the resolution

Going from 4K to 1080p removes three quarters of the pixels, and for most viewing it is invisible, because the video is being watched in a small player on a phone. Ask where the video will actually be seen. Footage destined for an Instagram story does not benefit from 4K in any way a human can detect on that screen, and 1080p is usually the sensible ceiling for anything not going to a television.

Lever 4: reduce the frame rate

Use this one sparingly. Dropping 60 fps to 30 fps saves real space and is perfectly fine for interviews, tutorials, and screen recordings. It is a poor idea for sport, gaming footage, or anything with fast panning, where the loss of smoothness is immediately obvious. Never drop below 24 fps for normal video, because motion starts to look like a slideshow.

Putting it together

A practical sequence for a large clip that needs to get small: trim to the essential portion, set resolution to the maximum the destination will actually display, keep the frame rate unless motion allows otherwise, then lower the bitrate until you hit the size target and inspect the busiest section.

Doing that in a browser avoids the two usual annoyances of video work, which are installing a heavy application for one file and waiting on an upload to somebody else's server. A browser based Video Compressor processes the file locally, which means your footage never leaves the machine and the speed depends on your processor rather than your upload bandwidth.

Resolution and Frame Rate for Each Destination

Exporting for the destination rather than exporting once for everything prevents most quality complaints. Platforms re-encode whatever you give them, so your job is to hand them a file that survives that second compression pass gracefully.

Two rules apply across all of them. First, upload at a higher quality than the platform delivers, because the platform will compress your file again and a weak source gives it nothing to work with. Second, never upscale. Exporting a 720p source at 4K adds file size and adds no detail, since the missing pixels were never recorded.

Trimming, Converting, and Repurposing Clips

Most video work is not editing in the cinematic sense. It is small mechanical jobs: cut the first eight seconds, rotate a clip somebody filmed sideways, convert a MOV so a colleague can open it, or pull one good minute out of a forty minute recording.

Trim before you do anything else

Trimming first makes every later step faster, because everything downstream operates on less data. It also forces a useful editorial decision: what is this clip actually for? A recorded webinar becomes far more useful as four separate three minute answers than as one ninety minute file nobody opens.

Trimming is also the least destructive operation available, since a clean cut can often copy the existing video stream rather than re-encoding it. A browser based Video Trimmer lets you set the in and out points and export only that section, which is the fastest possible route from a long recording to a usable clip.

Convert only when there is a reason

Every conversion is a chance to lose quality, so convert with a purpose: to make a file playable for a recipient, to meet an upload requirement, or to reduce size. Converting for tidiness is not a reason. When you do convert, work from the original file rather than from an already compressed version, since compressing a compressed file produces noticeably worse results than compressing the master once.

Repurpose one recording into several assets

A single recording usually contains at least four deliverables: the full version, a short highlight in vertical format, a silent loop for a web page, and an audio only version. Producing all four from one shoot costs very little time and multiplies the value of the footage. The workflow is always the same: trim to the segment, adjust the frame for the target aspect ratio, export at the destination resolution, then compress to a sensible size.

Turning Video Into GIFs

GIF is a strange survivor. It is technically an image format from 1987, limited to 256 colours per frame, with no audio and no modern compression. It is also the only moving format that plays automatically and silently in email clients, chat apps, documentation, and issue trackers, which is why it refuses to die.

The trap is size. A five second GIF at full resolution and 30 frames per second can easily exceed the MP4 it came from, because there is no temporal compression worth the name. Keeping GIFs small comes down to four decisions:

• Keep it short. Two to four seconds is the sweet spot. A GIF is a demonstration, not a video.
• Reduce the frame rate. 10 to 15 frames per second looks perfectly fine in a loop and halves the frame count.
• Reduce the dimensions. 480 to 640 pixels wide is plenty for a screen recording embedded in a document.
• Prefer flat content. Interface recordings, simple animations, and screen captures suit the 256 colour palette. Camera footage with gradients and skin tones does not, and will show banding.

When those constraints fit, converting a clip with an MP4 to GIF converter produces something you can paste into a pull request, a support reply, or a newsletter where an MP4 would simply appear as an attachment nobody clicks.

Pulling Audio Out of Video

A surprising share of video files are really audio files that happen to have a picture attached. Recorded meetings, interviews, lectures, sermons, and podcast video feeds are all consumed by ear far more often than by eye.

Extracting the audio track has three practical benefits. It cuts the file to a fraction of its size, because a 90 minute recording that runs to 2 GB as video is around 80 MB as audio. It makes the content portable, since an MP3 plays in a car, on a walk, and at double speed in a podcast app. And it makes the material far easier to transcribe, because transcription tools handle audio faster and more reliably than video.

Once the audio is separated, everything you might want to do to it, trimming the dead air, raising a level that was recorded too quietly, joining several takes together, is covered in How to Edit Audio Online.

The extraction itself is not a conversion in the damaging sense when done properly, since the audio stream can frequently be copied out intact. Using a Video to MP3 converter turns an unwieldy recording into something you can actually listen to, share, and archive at a sensible size.

One caution: audio extracted from video inherits every problem of the original recording. If the room was echoing or the microphone was across the table, the MP3 will be echoing and distant too. Fixing the recording setup is always cheaper than fixing the recording.

Aspect Ratios and Safe Areas

Aspect ratio is the shape of the frame, written as width to height. The three that matter now are 16:9 for landscape video, 9:16 for vertical, and 1:1 for square. Getting this wrong produces either black bars or an automatic crop that removes somebody's head.

Two habits prevent most aspect ratio pain. The first is to shoot with the widest destination in mind and leave room around the subject, so a landscape shot can survive a vertical crop without cutting into the action. The second is to respect the safe area: on vertical platforms, roughly the top 12 percent and bottom 20 percent of the frame are covered by interface elements, profile names, captions, and buttons. Any text placed there will be partly hidden, which is the most common reason a carefully designed title becomes unreadable in a feed.

When you must convert between ratios, you have three options and all of them cost something. Cropping keeps the picture sharp and loses the edges. Letterboxing keeps everything and adds bars, which look dated in a vertical feed. A blurred background fill duplicates and blurs the footage behind a centred version, which is the compromise most social editors use because it fills the frame without lying about the composition.

Where the File Size Actually Goes: A Worked Example

Abstract rules are easy to forget, so it helps to follow one real file through the whole process. Take a phone recording of a workshop: 12 minutes long, shot in 4K at 60 frames per second, arriving as a 3.1 GB MOV in HEVC. The task is to send it to twelve people, some of whom are on older Windows machines, and to post a short version on social media.

Start by working out why it is 3.1 GB. Twelve minutes is 720 seconds. 3.1 GB is roughly 25,000 megabits, which divided by 720 gives an average of about 34 Mbps. That is a normal phone recording bitrate for 4K60, chosen so the footage holds up under editing. It is roughly eight times what the finished video needs.

Now apply the levers in order. Trimming removes three minutes of setup and two minutes of packing up at the end, leaving seven minutes. That alone takes the theoretical size to 1.8 GB before touching quality. Dropping from 4K to 1080p removes three quarters of the pixels, and since nobody is watching a workshop recording on a cinema screen, nothing of value is lost. Dropping 60 fps to 30 fps is acceptable here because the camera is static and the subject is a person talking. Encoding to H.264 at 5 Mbps then produces a file of about 260 MB, which is a 92 percent reduction from the original, plays everywhere, and looks perfectly good in a normal player window.

For the social cut, the same source produces a very different file: 45 seconds of the best explanation, cropped to 9:16, exported at 1080x1920 and 6 Mbps, which lands around 34 MB. Notice that both outputs came from the original master rather than from each other. Exporting the social clip from the already compressed 260 MB version would have compressed compressed footage, and the difference is visible in the fast movement.

The lesson generalises. Most oversized files are oversized because they are still carrying capture settings into a delivery role. Recording settings are meant to be generous. Delivery settings are meant to be exact.

Subtitles, Captions, and Accessibility

Captions have quietly become the most important accessibility and engagement feature in video. A large share of social video is watched with the sound off, which means an uncaptioned clip is simply silent to a lot of its audience, and captions also make content usable for deaf and hard of hearing viewers.

There are two ways to attach text to video, and the difference matters. Open captions are burned into the picture during export. They cannot be turned off, cannot be translated, and cannot be read by a search engine, but they always appear, which is why social platforms favour them. Closed captions live in a separate track or file, typically SRT or VTT. They can be toggled, styled, translated, and indexed, and platforms such as YouTube use them to understand what the video is about.

Where you have the choice, upload a closed caption file. It gives viewers control, gives you a text asset you can reuse, and gives search engines something readable about a file that is otherwise opaque to them. Where the destination does not support caption files, burn them in and keep the source text separately so you are not retyping it later.

Practical caption rules: keep each line short enough to read in one glance, split at natural phrase boundaries rather than mid clause, keep them on screen long enough to read comfortably, and place them clear of the interface safe areas discussed earlier. Automatic transcription gets most of the way there now, but it consistently mangles names, jargon, and acronyms, so a quick manual pass is always worth the few minutes it takes.

Common Video Problems and How to Fix Them

The file will not play on someone else's computer

Almost always a codec problem rather than a container problem, usually HEVC from a recent phone. Convert to MP4 with H.264 and the complaint disappears. Ask what device they are using before assuming the file is corrupt.

The video is sideways or upside down

Phones record in one orientation and write a rotation flag into the metadata. Some players honour the flag and some ignore it, which is why the same file looks correct in one app and wrong in another. Rotating the video properly bakes the correct orientation into the picture so every player agrees.

Audio drifts out of sync as the video goes on

Progressive drift usually means a variable frame rate source, common in screen recordings and phone footage, that was re-encoded to a constant frame rate incorrectly. Re-exporting from the original with a fixed frame rate normally solves it. A constant offset rather than a drift is a different problem and simply needs the audio track shifted.

The upload keeps failing

Check the platform limit against your actual file size and duration. If the file is close to the ceiling, trim first and compress second. Uploading a 2 GB file over a slow connection also invites timeouts, so a smaller file often succeeds where a large one repeatedly fails for reasons that have nothing to do with the format.

The exported video looks worse than the preview

Usually the export bitrate was set far below what the content needs, or the export upscaled a smaller source. Look at the export settings rather than the editing timeline. If the source was 720p, exporting at 1080p cannot recover detail and simply wastes space.

Colours look washed out after export

This is often a colour space mismatch, particularly with footage recorded in a flat or log profile that was never graded, or HDR footage exported to a standard dynamic range target without proper conversion. Exporting standard 8 bit footage in a conventional colour space avoids the issue for everyday work.

A Repeatable Video Workflow

Every video task fits the same seven steps. Following them in order prevents the two classic mistakes, which are compressing before trimming and losing the original.

1. Keep the original untouched. Copy it into an archive folder and never export over it.
2. Trim to the content that matters. The single largest saving available.
3. Fix orientation and framing. Rotate, crop, and set the aspect ratio for the destination.
4. Choose the destination resolution. Match where it will be watched, and never upscale.
5. Export MP4 with H.264 and AAC. Unless you have a specific reason to use something else.
6. Compress to the target size. Then check the busiest few seconds for artefacts.
7. Name the file properly. Something like project-topic-1080p-2026-08.mp4 beats final_v3_REAL.mp4 in six months.

Run that sequence a few times and it becomes automatic. The tools change over the years, but the order of operations does not, because it follows the physics of what compression can and cannot recover.

Frequently Asked Questions

What is the best video format for general use?

MP4 with H.264 video and AAC audio. It plays on essentially every phone, browser, television, and editor made in the last fifteen years, and every social platform accepts it cleanly. Choose something else only when you have a specific requirement such as browser bandwidth (WebM) or editing quality (ProRes in MOV).

Does compressing a video always reduce quality?

Practically speaking yes, because the common codecs are lossy. The realistic goal is invisible loss rather than no loss. Keep the resolution and frame rate intact, lower the bitrate to a sensible level for the content, and most viewers will never see a difference on the screens they actually use.

Why is my video file so large?

Because file size is bitrate multiplied by duration, and cameras record at deliberately high bitrates so footage survives editing. 4K at 60 frames per second, or a ProRes recording, can use ten to fifty times the data a finished delivery clip needs. The recording settings, not the length alone, are usually the cause.

What is the difference between a container and a codec?

The container is the wrapper (MP4, MOV, MKV) holding the video, audio, subtitles, and metadata. The codec is the compression method inside it (H.264, HEVC, AV1). The extension tells you the container only, which is why two files ending in .mp4 can behave completely differently.

Is MOV higher quality than MP4?

No, not inherently. They are closely related containers that can hold identical video. MOV files often look better only because they were written at much higher bitrates or contain ProRes, which is an editing codec rather than a delivery one. The same H.264 stream in either container looks identical.

How do I make a video small enough to email?

Trim it, drop to 720p, and compress toward a target size. A one minute 720p clip at roughly 2 Mbps lands near 15 MB, comfortably inside typical 20 MB and 25 MB limits. For anything longer, a link is a better answer than an attachment.

Should I use HEVC or AV1 instead of H.264?

Use them when you control playback and want smaller files, such as archiving your own library or streaming at scale. Use H.264 whenever the file is going to another person, because it is the only codec you can assume everyone can open.

Does converting a video multiple times hurt quality?

Yes, and the damage accumulates. Each lossy encode discards detail permanently, so a file compressed three times looks meaningfully worse than the same file compressed once at the final setting. Always export from the original master.

Is it safe to use online video tools?

It depends entirely on whether the file leaves your device. Tools that upload your footage to a server hold a copy of it, with all the privacy and retention questions that implies. Tools that process video in the browser keep the file on your own machine, which is the safer default for anything personal, confidential, or commercially sensitive.

Bringing It Together

Video stops being intimidating once you separate the box from the compression. The container decides who can open the file. The codec decides how efficiently the picture is stored. Resolution, frame rate, and bitrate decide how it looks and how big it gets, and duration quietly multiplies against all of it.

The practical version is even shorter. Keep the original. Trim first. Export MP4 with H.264 at the resolution the destination actually displays. Compress toward a target and check the busiest moment. Convert only when there is a reason, and always from the master. Those five habits solve the overwhelming majority of video problems people run into, and they will keep working long after today's codecs have been replaced by better ones.