Many work hours can be lost to footage in formats that render slowly or need to be transcoded but fear not! Once you unlock the knowledge of codecs and containers, you’ll be able to continue your quest for great video.
Why Formats Matter
There’s a multitude of reasons a video format is important. If you understand the format that a camera shoots in, then you can calculate how much storage space you will need for the footage you plan to shoot. To know if editing or color correction software can handle a format natively or if it will have to be transcoded, you need to have knowledge of the format of the footage. When a film festival or broadcaster asks for a delivery format for video, the better you understand it, the easier it will be to have your project looking the best it can.
What Exactly is a Format?
By understanding a little about the codecs and containers used in popular formats, you will be able to make better choices for the work you do. When someone asks what format a video is in, they often want to know what container and codec were used to make it — and possibly what standard it is encoded to — unless, of course, the video is from the dark ages and then they usually want to know what kind of video tape it’s stored on and hope that they can find something on which to play it.
A codec is the order used to layout the data of an audio or video file in such a way that it may be utilized for playback, editing or changing to other codecs (transcoding). Codecs are used to organize media data, but that data is held within a container. There are many different types of audio and video codecs, and they each have their own advantages.
A container or a wrapper is what is used to hold audio and video data together in a single file along with additional information. Containers have file extension like .mov, .avi or .mp3. While some containers only tend to hold media in a particular codec, like the .mpg file container that’s used for MPEG files, some containers, like .mov, can hold data in a variety of audio and video codecs. The container has information to tell if there are both audio and video data held within it so things like media players will know to play them both at once.
Containers often also hold metadata on media in the file. That metadata can be as simple as the frame rate of the video up to showing what camera and lens were used to record the footage, what camera settings were used, where it was shot and information about the shot and the production. The metadata within a container can sometimes also tell you what standards the footage was produced in.
Not So Standard Standards
A car dealer may tell you that a car they are selling comes with a standard spare tire, but the standards to which that the wheel was made may only match a few makes, models and years of cars and not match any of the regular wheels on any car. Sadly, video can be very similar. If someone tells you that a video is in NTSC, they may only be referring to an NTSC standard frame rate like 29.97fps. If a video is said to be Rec. 709, this refers to a specific set of standards for HDTV that covers things like frame rate, color gamut and resolution, although one may point out that even Rec. 709 supports multiple frame rates.
To make things more confusing, you have the Rec. 2020 standard for UHDTV (4K TV) and the DCI standard for 4K film. These standards have different resolutions and different aspect ratios. Rec. 2020 has an aspect ratio of 1.78:1 (16 X 9), where DCI supports 1.85:1 and 2.39:1 (roughly 17 X 9 and 21 X 9). So if you’re working on a project like a short film that will need copies in Blu-ray, and it will also play from a DCP (Digital Cinema Package) at a theater, be aware that the standards for those formats are very different.
Some formats have been created for video using common codecs but only allowing certain variations in things like resolution and bitrate so they can more easily be used across hardware and software platforms. AVCHD and DivX both are formats that use the H.264 (MPEG-4) codec but are different standards.
Avoiding Compression Depression
Compressing a video file can cause it to lose a lot of its image and sound quality. Taking the same file and compressing it multiple times can make that loss much worse. Most online video hosts like Vimeo and YouTube will re-compress the video you upload so you want to make sure that you maintain as much quality as you can before it gets to them. Whenever possible, you’ll want to edit and master from your source footage or transcode to uncompressed or lossless codecs to maintain media quality.
An uncompressed codec stores media without compression so no quality is lost but the files are big. A lossless codec stores media with compression and without quality loss but with minimal space savings. A lossy codec stores media with compression and quality loss. With lossy compression, the higher the compression, the smaller the file and the greater the quality loss.
Most cameras and recorders use some type of compression so you want to keep as much data as you can when you’re in post. If your final renders are made from your source footage, then the only loss that you’ll have is from the compression of that render if there is any. Note that if your rendering for archiving, compression is not recommended. If you use a lossy intermediate codec like ProRes 422 or Matrox MPEG-2, then you’ll see a loss in image quality, but for the ease in workflow and depending on the delivery format, it might be worth it. It’s best to test your workflow out beforehand and decide before you edit because being half way through a project and seeing how much image quality you’ve lost to compression can be very depressing.
Compatibility with different codecs and formats for hardware and software can still be a huge challenge. You want to ensure that any codecs you plan on using from acquisition in production through post and onto distribution and archiving are compatible with your hardware and software before you start on your project. You can often see things like color shifts in images when moving from one piece of software to another since they are probably not processing the image data the same way. On a large project or one where consistency is critical, testing your workflow before you begin can you help avoid or learn how to compensate for problems like this.
The Academy of Motion Picture Arts and Sciences (they give out the Oscars) recently launched a standard image file format and workflow called ACES. It’s designed to eliminate shifts in color and other file compatibility issues between software from production through post and archiving. It’s already seeing support in software like Premiere Pro and DaVinci Resolve and use on films like “Chappie”. They are also working on a hardware certification program for ACES. ACES is free and easy to use; it’s the first cross platform (film and video) standard for digital motion picture archiving.
Intermediate codecs or formats are used for post-production work in many situations; however, for the fastest results in post, you’ll want to avoid using intermediate formats and work from the source footage. There are times when using an intermediate is preferable or a must. If the hardware and/or software you’ll be using for your workflow doesn’t support the source footage, you’ll need to use an intermediate format. If you’re not keeping the source footage and have limited storage space for files, then an intermediate format may work better for you.
WAV files (.wav) could be called the default industry audio standard (there is no global audio standard; ACES only supports image files). They are compatible with almost all production and post software and are always uncompressed because the files are small compared to video files. Some cameras record in a WAV compatible format, but it will list as PCM or Linear PCM (LPCM) because this is the codec used, and the audio must be placed in a container like .avi or .mov in order for it to remain synced with the video that is also being recorded. PCM audio encoding is also used in AIFF files. While there are a large number of audio codecs around, for most workflows WAV is going to be the best option for video work.
While there are hundreds of audio and video codecs around made for different uses, this is a list of common codecs and their typical applications.
Commonly referred to as MPEG-4, H.264 uses lossy compression and is one of the most common video codecs in use today. The codec is widely supported and used in production, post and distribution of video. Many camcorders and DSLRs record in H.264. It’s the standard for Blu-ray disks as well as many web video hosts. H.264 is more efficient for compression than MPEG-2 and it typically delivers better video quality at the same bitrate.
It’s the standard for DVDs and was originally used for cable TV. It was used for HDV videotape and was popular for web video. MPEG-2 is still used by some current cameras, and it’s commonly used by editing software to render video previews. MPEG-2 is a lossy compression but when used at lower levels of compression it can deliver a high image quality.
H.265 (MPEG-H, HEVC)
A lossy codec and the follow up to H.264, H.265 offers better compression than its predecessor. While there isn’t much support for H.265 now, it won’t be long before it’s more widely used.
Flash was once the most popular option for encoding online videos, but now, the Adobe developed, lossy codec is mostly used for animations and games.
MJPEG (Motion JPEG)
MJPEP was used in the past for web video and some post work. The lossy codec isn’t as efficient as MPEG-2 or H.264 and is seldom still used. MJPEG was based on JPEG compression used for still images.
A lossy compression, JPEG 2000 is the follow up to the JPEG format for stills. The JPEG 2000 format allows for very high quality image sequences, and it’s the compression used for Digital Cinema (DCP).
Red Digital Cinema developed its own variation of JPEG 2000 for their cinema cameras called REDCODE. It’s a low loss, high image quality compression that is supported natively by most professional post software. REDCODE uses the .r3d file container.
Apple ProRes is a series of codecs that offer both lossless and lossy compression. Although the ProRes codecs were designed for intermediate work in post, they’re being used as acquisition formats by makers of cameras and recorders because of the popularity of the codecs with users as well as the wide spread support of the codecs by software companies. The ProRes codecs were the replacement for the older Apple Intermediate codec.
Avid’s lossy intermediate codec, DNxHD, was designed for work with their software. Much like ProRes, hardware manufacturers are now using DNxHD in their products.
Windows Media Video is typically used as a lossy codec and has never been widely supported except by Microsoft products. WMV is the preferred video codec for PowerPoint.
VP9 is the lossy codec designed by Google that’s used for YouTube and supported by many web browsers for HTML5 video. There’s talk of adding uncompressed options to the codec.
HuffyYUV and Lagarith
HuffyYUV and Lagarith are both free lossless video codecs that are often paired with an .avi wrapper. The two codecs are not as popular as they once were possibly because at a compression rate of around 3:1 they don’t save a lot of space compared to uncompressed video files.
Here are a few codecs that were once common that have been replaced or are no longer widely supported: Apple Intermediate, Apple Animation, MPEG-1, RealVideo (Real Player), Indeo and Cinepac.
Common Containers and Formats
There are dozens of digital video formats and containers out there; this is a list of common containers and formats and their usual uses.
Developed by Adobe and sometimes confused with AdobeDNG (for still cameras), CinemaDNG was designed to be a standardized image sequence format for feature film work. CinemaDNG supports uncompressed and compressed image files. Up until the past few years, there was little support for CinemaDNG even from Adobe. Now, there are both hardware and software products that support CinemaDNG but the files for HD, 2K and 4K tend to be very large.
ACES (Academy Color Encoding System)
ACES is a color management and image file interchange system that is free, and it is growing in software support. ACES utilizes the OpenEXR file format that was developed by Industrial Light and Magic (ILM). The ACES OpenEXR files are uncompressed image sequences, but you don’t need to render those to work in ACES. You just chose the ACES profile that matches your work in your post software, import your footage and go. ACES is the only global archiving standard for digital video, so if you want your grandkids to be able to watch your work years from now, using the same archiving standard as The Academy (Oscars) is a good start to ensuring they can.
The AVI (.avi) container is supported by almost all post software. It can be used with possibly the largest number of audio and video codecs of any container which makes it very helpful.
Apple developed the MOV (.mov) container, but it’s not limited to Apple codecs or hardware. You can work with MOV files in Linux and Windows as well as Apple platforms with a large number of codec options.
Jointly developed by Sony and Panasonic as an HD recording format, AVCHD is typically utilized by consumer camcorders. The standard uses the H.264 video codec combined with support for compressed or uncompressed audio. Since the format is used by many cameras, software support is widespread. AVCHD typically uses the .mts and .m2ts file extensions.
Panasonic developed the AVC-Intra format for its professional camcorders. AVC-Intra uses intra-frame compression meaning the image is compressed one frame at a time as opposed to compression across multiple frames like AVCHD. AVC-Intra is supported by most professional post software and uses the MXF container.
MXF is a format designed to be a standard for file exchange of compressed video. While there is software support for MXF, it’s never had the broad reaching use of AVI or MOV.
XAVC and XAVC-S
These file formats were developed by Sony and utilize H.264 compression for recording HD and 4K video to cameras. Software support is growing for these formats that also use the MXF container.
Putting it All Together
The value of a codec or container is how it fits into a workflow. The format doesn’t have to be a popular one to work well for you, but there’ll be more information available about working in the more common formats should you need help. Remember to check the specs of your hardware and software when putting together your workflow plans. By looking at the projects you want to work on, you can figure out wich codecs and containers will be best for you.
Sidebar: Why Use an Image Sequence?
Image sequences can be used for rendering video; additionally, some animation software only export to image sequences. Saving a video clip as a series of still images instead of frame based video in a single file can have some advantages. If you have a video sequence that has a lot of heavy effects work that you’re rendering to an image sequence, and your render fails midway through, you can pick up where you left off because the render will be good up until the last frame you completed. If you’re rendering to a video file format like AVI, and the render is interrupted, the whole file is usually unusable.
Image sequences are supported as import and export formats by many popular post-production software packages including some used for still images like Photoshop. You may even see better performance using an image sequence as opposed to a video file in some software because the encoding of image sequences is more simplistic.
One of the biggest advantages to image sequences is in archiving. If a file in an image sequence that has been archived is corrupt and can’t be repaired or opened by any available software, then all that’s lost is a single frame of video. By analyzing the frame that came before the lost frame and the one that followed it, tools found in many post programs can create a frame to go in between the two; if done properly, it is often very convincing that it always belonged. While these tools were designed to create frames for slow motion, they can be used for restoration as well. If a file containing frame-based video (as opposed to an image sequence) is corrupt than it can be very difficult and sometimes impossible to recover any of the footage.
There are also a few drawbacks to using images sequences, however. They don’t have audio so you’ll have to deal with audio in its own file. Additionally, some software doesn’t perform well when using compressed image sequences. Despite this, depending on the types of projects you have, image sequences may still fit in well with your workflows.
Odin Lindblom is an award-winning editor and cinematographer whose work includes film, commercials and corporate video.