Which aspect ratio and size should your film use?

Shape first, then size

Every video file has two properties that people often mix up. The aspect ratio is its shape: how wide the frame is compared with how tall. The resolution is its size in pixels. A 1920 x 1080 file and a 3840 x 2160 file have the same shape, 16:9, and different sizes. A 1080 x 1920 file has exactly the same number of pixels as a 1920 x 1080 file and a completely different shape.

Shape is the decision that changes what the audience sees. Size mostly changes how sharp it looks. So the order of work is: pick the shape for each destination, then pick the largest sensible size your source can fill without being enlarged.

Aspect ratios are written in two styles, and both mean the same thing:

• Width to height in whole numbers, such as 16:9, 9:16 or 4:5. This style is common for television and social video.
• Width as a multiple of the height, such as 1.78:1, 1.85:1 or 2.39:1. This style is common in cinema. 16:9 is 1.78:1, because 16 divided by 9 is about 1.78.

A ratio above 1 is landscape, below 1 is portrait, and exactly 1 is square.

The shapes and sizes each destination uses

This is the table the tool works from. These are the shapes each destination displays best as of 2026. Most platforms accept other shapes too, and then either add bars or crop for you, usually not in the way you would choose.

For YouTube and TV, the tool picks 3840 x 2160 only when your source is at least that large. A 1920 x 1080 source exported at 3840 x 2160 is simply enlarged; it gains file size and no detail.

Platform guidance changes, and Instagram in particular has changed how it previews videos in the feed and on the profile grid more than once. Check each platform's current help pages before you export, and check the DCP format your festival or distributor asks for.

Worked example 1: a 4K film cut down for TikTok

The tool opens with this illustrative case: a film finished at 3840 x 2160, which is 4K UHD at 16:9, and a vertical clip for TikTok that should fill the screen.

1. Target shape. TikTok's full-screen frame is 9:16, exported at 1080 x 1920. As a single number, 1080 divided by 1920 is 0.5625.
2. Which way to crop. The source shape is 3840 divided by 2160, which is 1.78. That is far wider than 0.5625, so the full height is kept and the sides are cut.
3. Width to keep. Keep a strip as wide as the height times the target shape: 2160 times 0.5625 is 1215 pixels. The kept area is 1215 x 2160.
4. Width removed. 3840 minus 1215 is 2625 pixels. With a centered crop, that is 1312 pixels off one side and 1313 off the other.
5. Share of the picture lost. 1215 is 31.6 percent of 3840, so 68.4 percent of every frame is gone.
6. Scale. The 1215 x 2160 strip is scaled to 1080 x 1920, a factor of 1080 divided by 1215, which is 0.89. The picture is being reduced, not enlarged, so it stays sharp.

The tool returns a warning, not a failure. The picture quality is fine. The warning is about composition: more than two-thirds of each shot disappears, and a single centered crop will cut people out of two-shots and lose action that happens at the edges of the frame.

The same cut from an HD source

Run the same crop on a 1920 x 1080 source. The kept strip is 1080 times 0.5625, which is 607.5, rounded to 608 pixels wide. That 608 x 1080 strip has to be enlarged to 1080 x 1920, a factor of 1.78. The tool marks that as a failure, because enlarging a picture by more than one and a half times makes it visibly soft, and there is no export setting that fixes it. If vertical versions matter, finish the film at 4K or keep the camera originals so each vertical shot can be cut from the larger frame.

Worked example 2: vertical phone footage inside a YouTube video

Sometimes the problem runs the other way. Behind-the-scenes footage, an author interview recorded on a phone, or a clip from a vertical teaser often ends up inside a landscape YouTube video.

Take an illustrative 1080 x 1920 vertical clip placed in a 16:9 YouTube frame, keeping the whole picture.

1. Export size. The source is smaller than 3840 x 2160 in the dimension that matters, so the tool picks 1920 x 1080.
2. Scale to fit. The picture must fit both ways. 1920 divided by 1080 is 1.78 and 1080 divided by 1920 is 0.5625. The smaller number wins, so the clip is scaled by 0.5625.
3. Picture size. 1080 times 0.5625 is 607.5, rounded to 608, and 1920 times 0.5625 is 1080. The picture is 608 x 1080.
4. Bars. 1920 minus 608 is 1312 pixels of side bars, 656 on each side.
5. Share of the frame in bars. 608 is 31.7 percent of 1920, so bars cover 68.3 percent of the frame.

That is a lot of black. Editors often fill the side areas with a blurred, enlarged copy of the same clip, or place titles or a second image beside the vertical picture. The arithmetic is the same either way: the vertical picture itself occupies a 608-pixel-wide column in the middle.

Worked example 3: an HD edit made into a festival DCP

A DCP, short for digital cinema package, is the set of files a cinema projector plays. The common 2K DCP sizes come from the Digital Cinema Initiatives specification and the SMPTE standards that build on it. The 2K container is 2048 x 1080 pixels. A flat film (1.85:1) uses 1998 x 1080 of that container, and a scope film (2.39:1) uses 2048 x 858. The 4K equivalents are 3996 x 2160 for flat and 4096 x 1716 for scope.

Take an illustrative film edited at 1920 x 1080 and delivered as a 2K flat DCP, filling the frame.

1. Shapes. The source is 1.78:1. Flat is 1998 divided by 1080, which is 1.85. The flat frame is slightly wider than the source, so a little height is cut.
2. Height to keep. 1920 divided by 1.85 is 1037.8, rounded to 1038 pixels.
3. Height removed. 1080 minus 1038 is 42 pixels, 21 from the top and 21 from the bottom.
4. Share lost. 42 out of 1080 is 3.9 percent of the picture.
5. Scale. The kept 1920 x 1038 area is scaled to 1998 x 1080, a factor of 1.04.

The tool returns a warning because of that 1.04 enlargement. A 4 percent enlargement is small and usually invisible, but it is worth checking fine text such as credits after the DCP is made.

The alternative is to fit the HD picture inside the flat frame without cropping. The scale is the smaller of 1998 divided by 1920 and 1080 divided by 1080, which is 1.00, so the picture stays 1920 x 1080 and gets 78 pixels of black at the sides, 39 on each side. This is a common choice for HD sources. On a cinema screen those thin side bars are rarely noticed.

The same edit as a scope DCP

Choosing scope from the same 1920 x 1080 source cuts much more. The scope shape is 2048 divided by 858, about 2.387. 1920 divided by 2.387 is 804 pixels of height to keep, so 276 pixels are removed, about a quarter of the picture. If the film was shot and framed for 16:9, that crop will cut heads and feet. Scope is a choice you make when you shoot, not a conversion you apply at the end.

Crop or bars: how to decide

The tool offers two ways to put a source into a different shape.

Fill the frame (crop). The picture covers the whole screen and the edges of the source are cut away. Choose this when the subject usually sits near the center, when you can reframe shot by shot, and for vertical apps where viewers expect full-screen video.

Show the whole picture (bars). Nothing is cut, and black bars fill the leftover space. Choose this when the full composition matters, such as wide establishing shots, or when the shape is part of the film's look. Bars at the top and bottom are called letterboxing; bars at the sides are called pillarboxing.

Two rules of thumb help:

• If the crop would remove more than about 40 percent of the picture, a single centered crop will fail on many shots. Reframe each shot individually, moving the crop window to follow the action. This is sometimes called a reframe pass.
• If bars would cover more than about 40 percent of the frame, the picture will look small on the destination screen, especially on a phone. Make a version framed for that shape instead.

The tool uses 40 percent as the point where it raises a warning, for the same reason.

Caption and title safe zones

The frame you export is not the frame the viewer sees in full. Phone apps place buttons, usernames and descriptions over the video, and some TVs still trim the edges of the picture. Text that sits in those areas is covered or cut.

Vertical apps

TikTok, Instagram Reels and YouTube Shorts all place controls on the right side and text along the bottom of the screen. The exact layout differs per app and changes when the apps update. Some of these platforms show safe-zone guides in their advertising tools, but no single set of margins holds for every app, device and update.

The tool uses working margins that many editors apply to a 1080 x 1920 frame. These are a convention, not a platform rule:

That leaves a safe area of 880 x 1280 pixels, starting 60 pixels from the left and 220 pixels from the top. Burned-in captions should sit near the bottom of that area, above the 420-pixel band where the app's description text sits. Before posting, play the export on a phone in the app itself, with the description and buttons showing, and check every caption.

Reels previews on the Instagram profile grid show a cropped, shorter version of the vertical frame, at 3:4 since Instagram changed the grid in 2025. Keep the main subject and any title text near the middle of the frame so the preview still reads.

16:9 for TV and YouTube

For 16:9 video, SMPTE ST 2046-1, the SMPTE standard on safe areas, sets title safe at 90 percent of the frame's width and height and action safe at 93 percent. On a 1920 x 1080 frame, title safe is the central 1728 x 972 pixels, which means keeping text 96 pixels in from the left and right and 54 pixels in from the top and bottom. Most modern screens show the whole frame, but titles inside title safe are never at risk.

Cinema

A DCP is projected edge to edge, and there is no app layer on top. Keep burned-in text well inside the frame anyway, because projection setups vary. Festivals often ask for subtitles as a separate file in the DCP rather than burned into the picture, so check the festival's technical requirements before you make the package.

Shooting and editing so every version works

Most of the pain in aspect ratio work comes from deciding on the vertical and social versions after the film is finished. A book-to-film project usually needs several shapes: a main film, a trailer, vertical clips for social media, and sometimes key art cut from frames. Planning for them early costs little.

1. Pick the main shape at the storyboard stage. If the main film is 16:9 or flat, frame for it. If it is scope, accept that vertical cutdowns will lose more.
2. Keep the highest-resolution source you can. A 4K source gives a vertical crop that does not need enlarging. An HD source does not.
3. Mark a vertical safe area on the monitor. Many cameras and editing programs can show a 9:16 guide inside the landscape frame. Keep key action inside it in shots that might become clips.
4. Leave room above and below in vertical shots. Captions and app buttons will cover the bottom fifth of the screen, so frame faces in the upper middle.
5. Make each version its own export. Do not let one platform crop a file made for another. Export the shape and size each destination wants.

Square pixels and anamorphic footage

The tool assumes square pixels, which is true of nearly all modern camera and edit files. Footage shot with an anamorphic lens is recorded squeezed and must be stretched back to its true shape (desqueezed) before any of these calculations apply. Use the desqueezed size as your source width and height.

Odd numbers of pixels

Intermediate crop sizes in the tool, such as a 1215-pixel-wide strip, can be odd numbers. That is fine as a working area. The export sizes in the table are all even numbers, which most video encoders require for common color formats. If you set a custom export size, keep both dimensions even.

Letterboxing a wide film for streaming

A film finished in scope still has to play on 16:9 televisions and in 16:9 players. It is letterboxed: the full width of the scope picture fills the width of the screen, and the leftover height is black.

The arithmetic is the reverse of the scope crop. On a 1920 x 1080 frame, a 2.39:1 picture is 1920 divided by about 2.387, which is 804 pixels tall. The remaining 276 pixels become bars, 138 at the top and 138 at the bottom. A flat 1.85:1 film on the same frame is 1920 divided by 1.85, which is 1038 pixels tall, with 21 pixels of bar at the top and bottom. Streaming services accept letterboxed delivery as normal, but each publishes its own rules on how the active picture is marked, so read the delivery specification you are given.

Flat or scope for an adaptation

If the film will screen in cinemas or festivals, you will choose between the two standard theatrical shapes. Neither is more correct. They suit different material.

Flat (1.85:1) is close to the shape of a television, so it moves to streaming and broadcast with only thin bars. It gives more height, which suits stories built on faces, rooms and conversations: most domestic drama, romance and character pieces. It also loses less when cut down to vertical clips.

Scope (2.39:1) is much wider. It suits landscapes, crowds, action and any story where the space around the characters matters: westerns, war stories, large fantasy settings. It costs more height on every other screen. On a 16:9 TV a scope film sits inside bars totalling about a quarter of the frame, and vertical clips cut from it keep only a narrow slice.

For a book adaptation, read the source for clues. A novel told through close interior moments and dialogue usually plays well flat. A novel whose chapters open on wide settings and journeys may justify scope. Make the decision before shooting, because it changes how every shot is framed.

Posters, thumbnails and key art from frames

Still frames from the film often become thumbnails, posters and social images. Each of those has its own shape too. A YouTube thumbnail is 16:9. A vertical poster or a story image is closer to 2:3 or 9:16. A square social post is 1:1.

The same crop arithmetic applies. A 2:3 poster cut from a 3840 x 2160 frame keeps a strip 2160 times 2 divided by 3 wide, which is 1440 pixels, and loses 62.5 percent of the frame. So when you plan key art, shoot or render a few frames composed for the poster's shape, rather than hoping a landscape frame will crop well. The movie poster size guide covers print sizes for the poster itself.

Frame size is not frame rate

This tool deals only with the shape and pixel size of the picture. The number of frames per second is a separate decision with its own delivery rules: cinema commonly runs at 24 frames per second, European television at 25, and North American television at 29.97. Changing the shape of a file does not change its frame rate, and a festival or platform may ask for a specific rate. The timecode calculator explains how those rates affect running time and timecode.

Common mistakes

Uploading a 16:9 file to a vertical app and letting the app decide. The app will either add bars, which makes the picture tiny on a phone, or crop it with no regard for the subject. Make the vertical version yourself.

Exporting larger than the source. A 1920 x 1080 film exported at 3840 x 2160 is not 4K in any useful sense. The tool only picks the larger size when your source fills it.

Cropping once for the whole film. A centered crop works for some shots and fails for many. Reframe shot by shot for any crop that loses more than about 40 percent.

Captions in the bottom band of vertical video. The lowest part of the screen on vertical apps is covered by descriptions and usernames. Keep captions above it.

Converting to scope at the end. Scope from a 16:9 source loses about a quarter of the picture. Decide on scope before you shoot.

Sources and rules this page relies on

• Digital Cinema System Specification, published by Digital Cinema Initiatives (DCI), for the 2K and 4K DCP container sizes and the flat and scope image areas.
• SMPTE standards for digital cinema packaging, published by the Society of Motion Picture and Television Engineers, which formalize DCP image structure.
• SMPTE ST 2046-1, the SMPTE standard on safe action and safe title areas for 16:9 television, published by SMPTE, for the 90 percent title-safe and 93 percent action-safe areas.
• YouTube Help Center, published by YouTube, for the 16:9 player and recommended upload resolutions, and for Shorts.
• TikTok Help Center, published by TikTok, and Instagram Help Center and the Meta Business Help Center, published by Meta, for vertical and feed video shapes.

Platform layouts and recommended sizes change without notice. The vertical safe margins on this page are a working convention, not a published platform rule. Confirm the current guidance with each platform, and the DCP format your festival or distributor names, before you export.