Sustainability of Digital Formats: Planning for Library of Congress Collections

The High Efficiency Image File Format (HEIF) is an international standard defined by MPEG-H Part 12 (ISO/IEC 23008-12), first published by ISO in 2017. It is designed as a container for photographic images in any image encoding. HEIF is a special case of the general ISO_BMFF format, in which all data is encapsulated in typed boxes, with a mandatory ftyp box that is used to indicate particular file types. The initial specification for HEIF provided usage details for three compression schemes, the widely supported JPEG encoding for still raster images and two video encodings that are also applicable to still image items, namely Advanced Video Coding (AVC, aka MPEG-4, Part 10 and H.264) and High Efficiency Video Coding (HEVC aka MPEG-H Part 2 and H.265). This format description focuses on a subset of HEIF image types using HEVC encoding, with an emphasis on what the compilers of this resource have learned about implementations in cameras and mobile devices as of November 2020.

The ISO_BMFF mechanism for distinguishing among file types is based on 4-character codes known as brands; brands are registered at mp4ra.org. An ISO_BMFF file type is defined by the major brand, which comes first in its ftyp box. Also in the ftyp box is a list of brand codes identifying brands which with the file is declared to be compatible. This format description is primarily for HEIF files with the brands heic or heix either as major brand or a compatible brand. As of November 2020, heic is the most widely used brand for HEIF files with HEVC encoding because it is the variant chosen by Apple as the default format in iPhones, starting with the iPhone 7 in 2017. Apple uses the .heic extension (and pronounces it "heek" to rhyme with "leek") for such image files, which are sometimes referred to as HEIC files. HEIF files with a subset of encoding features from HEVC are used for image capture in supported devices and versions of Apple's iOS (starting with iPhone 7 and iOS 11) when the default "High Efficiency" is selected as the storage option in the iPhone's Format/Camera Capture settings. When the alternative "Most Compatible" option is chosen, JPEG is used. See Notes below for more details on Apple's use of the format for files generated by cameras in its mobile devices, based largely on details provided at Apple's WWDC 2017 developer conference. Some high-end Android phones, for example the Samsung Galaxy S10 series, when used with a new enough version of Android, also generate .heic files. In 2020, Canon and Sony brought out cameras which generate photos with the heix brand and the extension .hif. The key distinction between the heic and heix brands is that heic should only be used for the two most basic profiles of HEVC, namely Main or Main Still Picture; these profiles are limited to 8-bit color and 4:2:0 chroma subsampling. heix is primarily applied when the profile used extends either the bit-depth to 10 bits or more per color channel, or chroma subsampling to 4:2:2 or 4:4:4. For more details on brands associated with HEVC encoding in HEIF files, and examples of how the brands are represented in some example files, see Notes below.

A paper titled Objective Performance Evaluation of the HEVC Main Still Picture Profile, presented at the May 2015 IEEE conference on Transactions on Circuits and Systems for Video Technology, analyzed and discussed the compression performance of the HEVC Main Still Picture profile in comparison to the intra-only coding mode of the H.264/MPEG-4 AVC High Profile, the proprietary video coding schemes VP8 and VP9, and coding technologies used specifically in the still image application domain, such as JPEG 2000 lossy compression (using the irreversible floating point 9-7 wavelet transform), JPEG XR, WebP, and Baseline JPEG. The objective performance measures were based on the peak signal-to-noise ratio (PSNR) and associated average bit-rate savings. The experimental results clearly show the superior performance of HEVC Main Still Picture Profile with average bit-rate savings ranging from 10% relative to VP9 up to 44% relative to Baseline JPEG. The results from a later study comparing still image coding using HEVC in the HEIF format with the JPEG 2000 standard by a team from Nokia was reported at the 2016 IEEE International Conference on Image Processing (ICIP 2016) in a paper titled HEVC still image coding and high efficiency image file format. Based on the test set of images (with 16 bits per color channel in an ASCII-based uncompressed format) provided for the ICIP 2016 image compression grand challenge, HEVC still image compression was found to be superior to JPEG 2000, providing on average about 25% bitrate reduction for the same quality (measured by peak signal-to-noise ratio, PSNR). The paper includes a useful summary of the HEVC features particularly applicable to still image encoding. Table I in the paper lists the following important picture characteristics from HEVC as defined in edition 3 of ITU-T H.265 (2015) for supporting "high dynamic range" (HDR) and "wide color gamut" (WCG). These characteristics are listed below, supplemented by information from elsewhere in the paper:

• Support for several color gamut systems, color primaries, transfer characteristics, and color conversion matrices
• 8, 10, 12, 16-bit color. Improved "intra" (within image) coding methods compared to AVC. These methods can model more accurately different directional structures and can smooth regions that have gradually changing sample values. They help avoid introduction of artificial edges with potential blocking effects. The HEVC standard includes specific profiles for still images.
• Support for several chroma subsampling patterns: 4:2:0, 4:2:2, and 4:4:4.

HEVC features highlighted in the paper from Nokia mentioned above as supporting advanced imaging use cases include:

• Tiles: Support parallelized encoding and decoding of rectangular subsets of an image partitioned according to a grid. Apple, Canon, and Sony all take advantage of this feature. Apple captures in 512 x 512 pixel tiles for high performance when panning and zooming. See demo from presentation at Apple's WWDC 2017 on "Working with HEIF and HEVC."
• "Inter" prediction: Takes advantage of similarity from one picture to the next in a set of images. Helpful in many multi-picture use cases, such as focal and exposure stacks, image bursts, "cinemagraphs" (and similar animated images, known by a variety of names, including "Live Photo" from Apple and "Motion Photos" from Samsung).
• Auxiliary pictures: HEVC has standardized flags, codes, and supplemental enhancement information (SEI) message payloads (three mechanisms sometimes referred to together as "metadata signalling") for some categories of auxiliary images, in particular for transparency (alpha) planes, depth maps, and overlays. Overlays are covered in Annexes D and F of the HEVC specification. The details for alpha planes and depth maps are in Annex B.2.4 of the HEIF specification and Annexes D, F, and G of the HEVC specification. For a demo of how a depth map can be used to edit an image, see section from presentation at Apple's WWDC 2017 on "Working with HEIF and HEVC."

Among the features of the HEIF format that support advanced imaging use cases are:

• The ability to include any number of still images in a HEIF file. They can be categorized as master images, thumbnails, auxiliary images, etc. If tiles are used, each tile is a separate image, flagged as "hidden" and associated with a derived image with type grid that assembles all the tiles into a single image for display. A HEIF file can incorporate an image sequence, that may simply specify an order, but may be associated with timing intended to guide display. The track feature of ISO_BMFF is used to carry information about an image sequence. A HEIF file may also hold a collection of images with no implied order. Descriptive metadata can be applied to individual images and to the collection or sequence as a whole. EXIF metadata can be attached to every image. HEIF also supports any XMP-based metadata schema, and can be extended to handle other metadata schemas or schema languages.
• The original coded images in a file can be left alone, while a new image is derived by applying a transformation. This is known as non-destructive editing. HEIF specifies generic structures to define derivations and specifies a basic set of simple derivations relevant to all image encodings. For example, a composite HDR image derived from a set of exposure-bracketed individual images, will have references of type base to each of the source images. Other specified types for derived images include grid, for assembling tiles and iovl, for assembling layered overlays. A slightly different mechanism is used for images derived by cropping, rotation, or mirroring.

For HEIF: MPEG-H Part 12. See ISO/IEC 23008-12:2017 from the ITTF list of publicly available standards for full text and ISO/IEC 23008-12:2017 catalog record and preview for the initial specification.

For HEVC: MPEG-H Part 2, aka ITU H.265: High efficiency video coding. See ITU Recommendation H.265 (approved 11/2019) for full text of Version 7. See ISO/IEC 23008-2:2020 catalog record and preview for the equivalent ISO/IEC specification published in August 2020.

The HEIF file format has been adopted as an in-camera format for a number of cameras and mobile devices. In June 2017, Apple announced the adoption of HEIF at its WWDC 2017 developer conference. Built-in support for the format was introduced for capture on new iOS 11 phones. Qualcomm announced its first system on a chip (SOC) supporting HEIF in late 2018. See With Snapdragon 855 chip, Android phones get iPhone's photo-packing ability. In January 2020, Qualcomm introduced the lower-end SOC, Snapdragon 662, which also supports HEIF. Samsung introduced support for HEIF capture in its Galaxy S10 phones in 2019. See Here’s how to make Galaxy S10 camera photos take less storage space. Canon's EOS-1D X Mark III camera, introduced in January 2020, has the option of using HEIF internally as do the EOS R5 and R6 cameras, introduced later in 2020. See HOW TO... shoot in HEIF format on Canon cameras from Camera Jabber. Sony supports HEIF with HEVC encoding for still images captured with its A7S III movie camera. Both Canon and Sony are using the .hif extension for their HEIF images and offering options for 8-bit and 10-bit color as well as other advanced options.

The first versions of Apple's operating systems to support viewing, editing, and conversion of HEIC images were iOS 11 and Mac OS 10.13 (High Sierra), released in 2017. See Using HEIF or HEVC media on Apple devices. Support is available in Windows 10, but not without an explicit choice to add codecs. See Windows 10: Add Support for HEIC to Photos App. In Frequently Asked Questions about IrfanView, the answer to "How to load HEIC files in IrfanView? " offers equivalent advice. HEIF (.HEIC) Image Support for Linux/Ubuntu [quick guide] describes a mechanism for handling HEIF/HEIC files in Linux/Ubuntu, using the heif-gdk-pixbuf package.

Mac OS applications that have built-in support for .heic files, include Photos, Preview and Quick Look. These applications can render/display images that have characteristics beyond those found in Apple-generated images (see Notes below), for example, images with 10-bit color from Canon and Sony cameras. For support in third party image editing software, see Wikipedia entry for HEIF, bearing in mind that use of such software in Mac OS later than High Sierra (10.13) will typically work as it comes, but for Windows applications it is typically necessary to install the codec extensions mentioned in the previous paragraph. For example, Adobe Photoshop can open .heic files on recent Mac OS versions and hardware, but not on Windows 10 without the codec extensions.

As of writing, in November 2020, there is no browser-based support for HEIF images; see https://caniuse.com/heif. Hence, many photographs captured in HEIF format are converted into JPEG encoding in JFIF files for use on the web or as email attachments, often without the user who took the photo realizing the conversion has happened. For example, the option to email pictures using the Photos app on an iPhone sends JPEG/JFIF files with EXIF metadata in the size chosen by the user.

For software libraries for the HEIF/HEVC combination of file format and encoding, see HEIF. Most software that says it supports the HEIF format for reading/decoding or writing/encoding supports some profile(s) of the HEVC (H.265) encoding standard, typically including the profiles used in Apple-generated .heic files.

Canon and Sony provide support for editing their own variants of HEIF/HEVC combination. See Working with files saved in HEVC format from Canon with information on use of its Digital Photo Professional (DPP) software with .hif images generated by Canon cameras. Sony provides a HEIF Converter tool for its HEIF files; conversion to TIFF retains the full bit depth and BT.2100 HLG color profile indicator from the HEIF source; conversion to JPEG reduces the image to 8 bits.

Image conversion tools that claim to support HEIC, primarily to convert to JPEG or PNG, include: iMazing HEIC Converter; CopyTrans HEIC for Windows; Convertio. See also Top & Free HEIC File Converters for Windows/Mac. A useful tool for looking at the structure of a HEIF file with HEVC encoding (or indeed for looking at any ISO_BMFF file) is ISOBMFF Explorer, available as source code on github or as a ready-to-use app for the Macintosh [at the App Store]. The app reveals the nested box structure of the file and much of the structural metadata.

The underlying HEVC encoding is heavily encumbered with patents. According to HEVC Advance Passes 10,000 Patent Milestone - announces Toshiba Corp. Joins as a Licensor, a March 2020 press release, there are around 17,000. Patents have proved a general problem for video coding schemes. See, for example, Patents, Standards Brew Licensing Woes (June 2019) from EE Times. Rather than there being a single patent pool, there are three, as shown in a blog post from January 2018 by Leonardo Chiariglione.

In January 2019, a study of HEVC / H.265 standard-essential patent royalty rates was released by Unified, an international membership organization that seeks to improve patent quality and deter unsubstantiated or invalid patent assertions in defined technology sectors (Zones). Video Codecs form one of the zones for Unified's activities. The report, based on a methodology Unified describes as objective economic evaluation (OVAL), claims that it "demonstrates that HEVC rates should be comparable to or less than the cost of licensing the main patent pool for AVC (MPEGLA), HEVC’s predecessor codec." The report found that "each of the three major licensors are asking for licensing rates that likely exceed what the total HEVC rate should be, both individually and when taken together."

It is clear to the compilers of this resource that the patent situation has been an impediment to adoption. In particular, the lack of built-in support for HEIF formats in browsers and in operating systems other than Apple's iOS and Mac OS seems likely to be related to the challenges and risks associated with the patent landscape. Comments welcome. Although related to the earlier AVC (H2.64) coding scheme, the patent-related battle between Nokia and Lenovo in Germany in 2020 illustrates the risks. See, for example, Setback for Nokia in German patent battle with Lenovo, a November 2, 2020 post by Reuters. Nokia has not joined any of the HEVC patent pools.

Excellent support for online image functionality, panning, zooming, etc., especially when the tiling feature is employed. Requires hardware coding support for real-time image coding at capture time; decoding/rendering benefits from hardware implementation.

An overarching objective for this format was reduction in file size, while maintaining image quality and decoding performance; found superior to image formats JPEG, JPEG 2000, JPEG XR, and application of video codecs, AVC (H.264), VP8, and VP9 to single images. See Objective Performance Evaluation of the HEVC Main Still Picture Profile (2015) by Tung Nguyen and Detlev Marpe.

Excellent support for images with large picture sizes and high spatial resolution. See Demo of zooming into a huge panorama that has 2.9 gigapixels (91,000 pixels by about 32,000 pixels). In comparison, JPEG is limited to 64,000 pixels in both directions.

Bit depth depends on HEVC encoding profile used. HEVC has profiles that support up to 16 bits per color channel, but as of 2020, implementation in cameras and mobile devices does not appear to exceed 10 bits. Color spaces are indicated in the ColorInformationBox (with box type colr ), which permits three different 4-character values for color type, allowing three mechanisms for specifying the color space. The first mechanism is indicated by the colour type nclx as defined in Annex A.7.21 of the JPEG-XR specification for PTM_COLOR_INFO. This provides a compact way to specify color primaries, transfer characteristics, and matrix coefficients via integer codes as listed and interpreted in tables A.8, A.9, and A.10 in the same annex. Sample images from Canon and Sony showed use of this mechanism, with different choices for the coded values. The compilers of this resource have not attempted to explore the differences or to discover whether the cameras use different values based on user choices or camera configuration. The second method, indicated by prof, declares an "unrestricted" ICC profile. For example, the color space found in photos taken with an iPhone 7 is an Apple profile called Display P3, which uses DCI P3 primaries, a D65 white point, and the sRGB transfer function. The compilers of this description have not identified any use of the third mechanism, indicated by rICC which appears to be for use in specialized workflows. Comments welcome.

Brands that identify variants of HEIF files with HEVC encoding: Types for files complying with the ISO_BMFF format are distinguished by means of brands. Each file has a single major brand and may have a list of brands with which the file is declared to be compatible. Brands are registered at https://mp4ra.org/#/brands.

Brands associated with HEIF image files with HEVC encoding are listed below. The file extension and MIME type (aka Internet media type) of a file in deriving from the ISO base media file format usually reflect the major brand in the filetype box. See the IANA registrations at https://www.iana.org/assignments/media-types/image/heic and https://www.iana.org/assignments/media-types/image/heic-sequence for the brands associated with the two MIME types registered for HEIF files with HEVC encoding. HEVC-specific brands are defined in Annex B of the HEIF specification. The brand depends on which of the many HEVC profiles defined in Annex A of the HEVC specification was used for encoding the primary image in the file. The brands associated with HEVC encoding fall into three groups. The first group of brands applies to variants intended for image capture, editing, and display:

• heic: for HEIF files with individual image items and image collections where the image encoding of the items conforms to the Main or Main Still Picture profile of HEVC. The Main profile is defined in clause A.3.2 of the HEVC specification. The Main Still Picture profile is defined in clause A.3.4 of the HEVC specification. These profiles are limited to 8 bits per color channel with chroma subsampling of 4:2:0. Associated primarily with .heic file extension (although .hif seems permissible according to the IANA registration) and MIME type image/heic. See IANA registration.
• heix: for HEIF files with individual image items and image collections where the image encoding of the items conforms to the Main 10 profiles or any of the "format range extensions profiles" of HEVC. The Main 10 and Main 10 Still Picture profiles are defined in clause A.3.3 of the HEVC specification. The "10" in these profile names indicates the use of 10 bits per color channel. Format range extension profiles are explained in clause A.3.5 of the HEVC specification. They are a collection of profiles that extend either the bit-depth to 10 bits or more per color channel, or chroma subsampling to 4:2:2 or 4:4:4, or both. Also included are profiles for monochrome images. Associated with .heic or .hif file extension and MIME type image/heic. See IANA registration.
• hevc: for HEIF files with image sequences where the image encoding of the items conforms to the Main or Main Still Picture profile of HEVC. Associated with .heics file extension and MIME type image/heic-sequence. See IANA registration.
• hevx: for HEIF files with image sequences where the image encoding of the items conforms to the Main 10 profiles or any of the format range extension profiles of HEVC. Associated with .heics file extension and MIME type image/heic-sequence. See IANA registration.

The second group of brands applies to variants of ISO_BMFF that address requirements for streaming services, using an approach described as Layered HEVC (L-HEVC), which employs a concept known as Network Abstraction Layers (NAL). This approach is specified in clauses 9 and 10 of ISO/IEC 14496-15:2019 Information technology — Coding of audio-visual objects — Part 15: Carriage of network abstraction layer (NAL) unit structured video in the ISO base media file format. The compilers of this resource have not encountered or investigated use of these L-HEVC variants beyond assembling the following list of associated brands. Comments welcome.

• heim: for HEIF files with individual image items and image collections where the image encoding of each layer of the item conforms to the Main profile or the Multiview Main profile of HEVC. Associated with .heic file extension and MIME type image/heic. See IANA registration.
• heis: for HEIF files with individual image items and image collections where the image encoding of each layer conforms to Main profile, Main 10 profile, the Scalable Main profile, or the Scalable Main 10 profile of HEVC. Associated with .heic file extension and MIME type image/heic. See IANA registration.
• hevm: for HEIF files with image sequences where each layer in the reconstructed access units of the image sequence track conforms to the Main profile or the Multiview Main profile. Associated with .heics file extension and MIME type image/heic-sequence. See IANA registration.
• hevs: for HEIF files with image sequences where each layer in the reconstructed access units of the image sequence track conforms to the Main profile, Main 10 profile, Scalable Main profile, or the Scalable Main 10 profile. Associated with .heics file extension and MIME type image/heic-sequence. See IANA registration.

The third group of brands are structural rather than codec-specific. These structural brands serve as a lowest common denominator for HEIF files, applicable to all codecs.

• mif1: structural (not codec-specific) brand applicable to all HEIF files for images and image collections. Defined in clause 10.2 of the HEIF specification. Typically appears as the first compatible brand in all HEIF files for individual images or image collections with HEVC encoding.
• msf1: structural (not codec-specific) brand applicable to all HEIF files for image sequences. Defined in clause 10.2 of the HEIF specification. Typically appears as the first compatible brand. Has been observed in use as the major brand for early demonstrations of image sequence files with the .heic extension. However, HEIF: all you need to know about High Efficiency Image File Format, a January 2020 post from IONOS, states that "newer Apple devices (A9 processor required) automatically use the .heics extension when using the container format for image sequences." The compilers of this resource have not been able to confirm this statement. Comments welcome.

Shown below are a few examples of the ftyp boxes and their major and compatible brands found in image files with .heic or .hif extensions. In all these cases the ftyp box is the first box in the file and these strings show up in a text editor. Since the ftyp box starts with its length (as a 32-bit unsigned binary integer, occupying 4 bytes), the "ftyp" string usually begins in the 5th byte of the file. The four periods in the middle of these strings represent 4 bytes of binary zeroes; these bytes are reserved to support future versioning.

• ftypheic....mif1miafMiHBheic
  taken with an Apple iPhone 7 using iOS 14, with extension .heic. Note: the miaf and MiHB brands are defined in ISO/IEC 23000-22:2019 and are described in the MP4RA list of brands as "Multi-Image Application format brand for general MIAF requirements" and "Multi-Image Application format brand for MIAF HEVC Basic Profile." See also Multi-Image Application Format (MIAF) from Nokia.
• ftypmsf1....msf1hevcheicmif1iso8
  early example of an animated image sequence ("Starburst), with extension .heic. This example can be found at Examples from Nokia under the heading "Image sequences." The iso8 brand is defined in the underlying ISO_BMFF specification and identifies the version of that format.
• ftypmif1....mif1heichevc
  from an example available from the github repository for the libheif software library
• ftypheix....mif1heix
  from a Canon camera, distributed with .hif extension. The photo uses 10-bit color, and indicates this with a major brand of heix. See Inside Canon High Efficiency Image File format for more Canon examples.
• ftypheix....mif1heixmiafMiHAjpegSHIF
  from a Sony AS7 III camera, distributed from New From Sony: HEIF to TIFF Converter with .hif extension. Image has 10-bit color and declares itself compatible with the miaf and MiHA brands described in the MP4RA list of brands as "Multi-Image Application format brand for general MIAF requirements" and "Multi-Image Application format brand for MIAF HEVC Advanced Profile." Additionally, it declares itself compatible with the use of JPEG encoding within the HEIF structure and the brand SHIF, which does not appear in the list of registered brands as of November 2020, but may be a new brand in use by Sony. Comments welcome.

Specifics for Apple-generated HEIC files : The first Apple iPhones to support image capture in the HEIC format were the iPhone 7 and the iPhone 7 Plus, released in 2017 and incorporating the A10 Fusion processor chip, the first Apple chip to support hardware encoding of HEIF and HEVC. The HEIC subtype of HEIF is used to capture images on supported mobile devices using Apple's iOS 11 or later, when the default "High Efficiency" is selected as the storage option for photographs in the Camera app. See Using HEIF or HEVC media on Apple devices for information on the iPad models able to capture images in the HEIC format. Note: Apple typically uses a new processor chip with the release of each new family of products. See Wikipedia entry for Apple-designed processors. For example, in October 2020, Apple released the IPhone 12 family of phones with the A14 Bionic chip.

Apple provided detailed information about the characteristics of the HEIC images generated by its devices in some presentations at its WWDC 2017 conference for developers. See, in particular:

• HEIF Configuration with Camera App, the final content from a voice with slides presentation on the High Efficiency Image File Format. This is a concise summary of the characteristics of the HEIC image files generated by Apple mobile devices.
• High Efficiency Image File Format, starting from the slide on file extensions from https://developer.apple.com/videos/play/wwdc2017/513/?time=565. This slide is followed by detailed sections on HEIF anatomy and the HEVC codec as applied by Apple to its images. The sections on Tiles and HEVC are particularly helpful. The section on HEVC compares this codec with JPEG compression.
• Presentation on Introducing HEIF and HEVC starting at https://developer.apple.com/videos/play/wwdc2017/503/?time=1008, a link to another summary of the characteristics of the Apple-generated HEIC images, followed by material aimed at application developers, including discussion of the software tools Apple provides for encoding and decoding and information on which Mac processors offer hardware support for encoding and decoding.
• Presentation on Working with HEIF and HEVC starting at the beginning of the section on HEIF at https://developer.apple.com/videos/play/wwdc2017/511/?time=1208. This presentation covers, not only matters mainly of interest to developers, but also demonstrations of what the HEIF/HEIC combination format as implemented by Apple makes possible to enhance the experience for the ultimate end users of HEIC images in the Apple Photos application and features useful for developers of image editing applications. [Aside: There is also a sprinkling of humor and poetry.]

A brief summary of the technical characteristics of a still image captured by an Apple device, drawn from the video resources above, is:

• .heic extension.
• 512 x 512 pixel tiles positioned in a grid.
• 312 x 240 pixel thumbnail (larger than for JPEG, because of better compression using HEVC).
• use of HEVC Main Still Picture profile for encoding tiles. Since the Main Still Picture profile is a subset of the Main Profile, this implies 8 bits per color channel. In the experiment reported in Objective Performance Evaluation of the HEVC Main Still Picture Profile, the Main Still Picture profile outperformed JPEG by 44% based on PSNR and associated bit-rate savings. Based on Apple's own experiments, the company claims twice the compression for the same visual quality.
• use of HEVC Monochrome profile for depth map.
• EXIF metadata (equivalent to that generated for JPEG by earlier iPhones).
• Technical metadata for depth map in XMP.
• After the ftyp box, the order of boxes/items in the file (not constrained by ISO_BMFF or HEIF specifications) is structural metadata (to allow decoders to prepare for processing the coded data), thumbnail, followed by coded data for the main image.

AVCapturePhotoApple frameworks for developers to work with HEIF files with HEVC encoding include: ImageIO, the lowest level of Apple image framework access, which supports HEIF as a destination format, as well as a source for decode, incremental loading, metadata, and thumbnail abstraction. Higher level access for editing or managing images in Mac OS is available using Photokit APIs. On a suitable Apple iPhone or iPad, developers can employ the AVFoundation framework, using the new AVCapturePhoto and AVCapturePhotoOutout classes, to control capture and import HEIF content from the camera and write files in the HEIC format or in another format, such as JPEG.