Category Archives: Image Formats

  • TIFF

    TIFF is the most comprehensively capable format for image data, supporting nearly any kind of components. Let’s dive in and see how it does that.

    The Tagged Image File Format (TIFF) is an all-purpose container format that supports a nearly infinite variety of image types and component combinations. The name of the format gives a clue to its power. TIFF does very little to specify image encoding, but instead provides a flexible framework for the storage of complex image objects through the robust use of tags. Although the most recent TIFF specification dates back to 1993, it’s possible to include nearly all innovations in imaging technology in a saved TIFF. 

    IFD 

    The key to TIFF’s flexibility is the Image File Directory (IFD) structure of the file. TIFF was designed to be extensible, allowing new imaging components to be defined and added as a “subfile.” As new methods of image compositing, adjusting and other manipulations were developed, it became possible to save this data into TIFF by defining a new IFD structure. This has allowed TIFF to support image features that were not even contemplated at the date of the last written specification. 

    Note that many of the components that can be saved in a TIFF may be proprietary and only “understood” by one specific application. For instance, adjustment layers that can be saved in a TIFF by Photoshop typically cannot be understood by other applications. This could create a problem for interoperability if not for the requirement of a baseline image as described below. 

    The baseline image

    There is another feature of TIFF that has led to its wide adoption as the most capable and compatible image container. The TIFF specification requires that whenever there are multiple components included in a TIFF, there is always a full-resolution, full quality composite of the finished image in the file. This composite is supposed to conform to certain standardized characteristics for pixel layout, color and compression. By including this baseline image, TIFF remains technically compatible with any other program, even if the rest of the TIFF is filled with incomprehensible IFD components.

    The gold standard for archives 

    The IFD structure along with the required baseline image make TIFF an extremely popular choice as an archival image format. TIFF has long been considered the gold standard for archiving. For images that began their digital lives as traditional RGB scans, this remains true. 

    However, the popularity of digital cameras as image capture devices has begun to change this calculation. Traditional TIFF images are not designed to support the original data in camera raw files. Therefore, converting camera raw originals requires that a destructive transformation be applied to the raw data in order to create the TIFF. In order to also save a copy of the original image data, it’s necessary to save an additional file, or to save the raw data into an IFD inside the TIFF, which is not a standardized operation. 

    We will see how DNG addresses the shortcomings of TIFF as a camera raw archive file a little later. 

    Format details 

    Interoperability

    As outlined above, a properly-created TIFF file should be widely compatible with many applications and operating systems, even when subfile IFDs are not understood by the application. The TIFF specification requires that IFDs which are not understood be ignored when opening, but should be preserved if the full TIFF is saved into a new layered file. Because TIFF supports a wide array of color models and bit depths, it’s possible that certain applications will be unable to open a TIFF file because the program does not understand that color model (e.g., a CMYK TIFF may be unreadable to a web browser). 

    Supported color models

    TIFF supports most common (and uncommon) color models, including indexed color, grayscale, CMYK, RGB, LAB and more. TIFFs support bit depth up to 32 bits per channel. 

    Compression

    The original TIFF specification had a handful of limited options for compression, primarily LZW compression. Newer imaging applications may offer ZIP or, depending on the underlying image settings, JPEG compression. ZIP and JPEG are less common and may produce a file that is unreadable to older applications. For maximum compatibility, particularly in files for archives, it’s best to avoid using any compression. Note that saving TIFF files with all their IFD layers intact will make files significantly larger than a flat file with no additional image subfiles. Although these files can become quite large, it’s often desirable to preserve all adjustment layers and subfiles in the master copy of the image file. 

    Support for additional components

    TIFF has virtually unlimited support for subfile components, and is typically limited by the capabilities of the application making the TIFF. 

    File size limitations

    Traditional TIFF files must be smaller than 4GB in total size and may not be longer than 30,000 pixels on any one side. If a larger file is required, the BigTIFF or, more commonly, PSB (Photoshop Big) file formats can be used. 

    Variants

    In addition to BigTIFF mentioned above, there are a handful of variants of the TIFF format. 

    TIFF/EP 

    TIFF/EP is a version of TIFF that is designed for electronic photography. It was an early attempt at standardizing an IFD and tag structure for digital cameras. The TIFF/EP specification has been replaced by the more structured and formalized DNG specification. 

    PSD 

    The Photoshop Document (PSD) file format is a variant of TIFF that specifies certain characteristics that are useful when working with images in Photoshop. For the vast majority of purposes, PSD and TIFF can be used interchangeably. PSD will offer some advantages in smaller file size compared with TIFF. 

    The main file size saving with PSD is the ability to forgo the composite image when saving a PSD. Since Photoshop has the ability to reconstruct the finished image from the component parts, Adobe allows PSD files to be saved without the composite final image. While this can offer significant file size savings, it does so at the expense of compatibility. PSD files without the composite image can only be viewed in Photoshop, and run a much greater risk of being incompatible with other software. When saving PSD images, I strongly suggest checking the Maximize Compatibility checkbox as shown below. 

    When working in Photoshop with motion image files, it’s often advisable to save as PSD rather than TIFF since the format is being continually optimized to effectively handle motion image editing components. 

    Best uses for TIFF files 

    Master edit files

    TIFF is an excellent choice for master rendered files like those made in Photoshop. TIFF’s support for multiple components allows it to function very much like a project file where multiple components can be saved. But TIFF has the advantage of also having a widely accessible finished version inside the file as well. TIFFs are great for both the works-in-progress phase of the workflow as well as archiving the final image. 

    Delivery of high quality images

    Due to lossless compression and wide compatibility, TIFF is a very popular choice for delivery of finished images. However, when the finished image is a flat 8-bit RGB file with no alpha channel, a JPEG can provide the same level of quality at a much smaller size. 

    Storage or delivery of color models not supported by other images – CMYK, CIELAB, Duotone or other color models not supported by JPEG or PNG should typically be saved and delivered as TIFF. 

    Image archiving

    TIFF is considered the safest format for long-term image archiving, especially for images greater than 8 bits per channel.

  • PNG

    In today’s post we examine PNG, a format that was specifically designed for the internet age. PNG also solves some problems around the inclusion of text and graphical elements in image files.

    The Portable Network Graphics (PNG) format was specifically designed as an image format for use on the web. It supports a broad range of image object characteristics, including multiple color models, the use of alpha channels, transparency, and high bit depth images. PNG uses lossless compression. In response to the uncertainty of patent status of Compuserve Gif and TIFF’s LZW compression scheme, PNG was created using features that are not covered by any patents so it could be used freely. It is supported by almost all operating systems and web browsers and is therefore a good choice for many on-screen uses. 

    PNG is most appropriate for simple images that do not include processing instructions, alternate versions or multiple image components, with the exception of alpha channels. It’s a good format for archiving screenshots, logos and other text-heavy images. 

    Format details 

    Interoperability

    PNG is widely supported by both operating systems and applications making it a good choice for interoperability. However, because PNG supports color models and bit depths that are not universally supported, it’s possible that a PNG will not open in a particular application. (This would be most common for PNG files that include 16-bit per channel color information.) 

    Supported color models

    PNG can support grayscale, indexed color and RGB color. PNG supports bit depth from one bit (black or white) up to 16 bits per channel. Indexed color makes PNG efficient for saving graphics like logos and illustrations. (Transparency is important here, too, since most logos are not simple rectangles.) 

    PNG does support embedded color profiles, but this support does not seem to be as common as profile support in TIFF and JPEG.

    Compression

    PNG compression is lossless which maximizes image quality. PNG compression does a much better job of preserving and rendering hard lines like you find in type, logos or drawings. Since PNG does a better job with this type of image, it’s an ideal choice for logos and screenshots that may contain type. 

    Support for additional components

    PNG supports alpha channels, which are probably most useful for adding transparency to an image, but have other uses as well. While PNG does not natively support EXIF metadata written by cameras, it is possible to add this metadata to PNG files. There is also an animated variant of PNG which was designed as a replacement for the animated GIF, but it is not widely supported.

    Best uses for PNG files 

    PDF, web and other electronic publication

    PNG is a good choice for electronic publication, particularly when the original is a screenshot, logo or other type-heavy image. It’s also fine for high-quality photographic images, but does not offer the same file size savings as JPEG for photos. 

    Capture logos, screenshots and converted vector graphics

    PNG is the native format for many screenshot applications. It offers high quality, transparency, and some file-size saving over TIFF. 

    Library originals for read-only image applications

    When the original image is captured as a PNG, that’s a perfectly acceptable format to use for read-only library storage (as long as the library application supports color profiles for PNG). 

    In the next post, we will look at TIFF, the most extensible image format.

  • JPEG

    The ubiquitous JPEG image format is particularly useful for delivery and distribution of images. In today’s post, we look at the details of the format. 

    The JPEG format was created by the Joint Photographic Experts Group as a standardized format for digital image files in 1986. It is the most commonly used digital image format–offering good compression and unparalleled support across all platforms. JPEG is primarily an encoding format, used to specify how to turn an image into numbers, and how to compress the bits effectively. JPEG also offers some container features–such as the ability to add metadata–but this is pretty limited. 

    Format details 

    Interoperability

    As one of the oldest image standards, the JPEG format enjoys extremely wide support. Almost every digital camera can create JPEG formatted files, and JPEGs comprise the vast majority of image files on the web. 

    Supported color models

    Most JPEG files are saved in some flavor of RGB color, typically in the sRGB color space. JPEGs can also be saved as grayscale images. JPEGs can be used for CMYK file delivery. The JPEG format does not support bit depth greater than 8-bit

    Compression

    Image compression was a primary goal of the original JPEG standard, and multiple levels of compression are supported. They may be represented as “low, medium and high quality” or as a percentage scale, or as a scale from 1 to 10 or 1 to 12. At the highest levels of compression (smallest files), the resulting images are highly artifacted. At the lowest levels of compression, JPEG files are essentially lossless. JPEG compression works most effectively on images without a lot of fine detail. When a lot of detail is present, larger files are produced for any given compression setting. A file size saving on 10:1 is typically possible without sacrificing noticeable detail. 

    Support for additional components

    Official support for additional components is very limited and officially includes metadata and color profile. JPEGs don’t even support alpha channels. Taken together, these limitations mean that JPEG is unsuitable as a rich media format and is really only suited for basic digital  photographic images. 

    Warning : Resaving lossy JPEGs throws away information 

    Every time you resave a JPEG file with lossy compression, you throw away image data. If the quality level you are using is very high, then it might not be noticeable. But if you are using a medium quality or lower setting, you can produce significant degradation in some images after just a few saves. This continual loss of quality upon lossy resaves makes JPEG a poor choice for an image editing format when working in Photoshop or other program that requires resaving the original file. 

    When working in a read-only parametric image editor like Lightroom, a high-quality JPEG file can be a perfectly acceptable source image file, especially if it is the original JPEG produced by your camera. 

    Best uses for JPEG files

    PDF, web and other electronic publication

    Due to high levels of support, JPEG files are excellent choices for on-screen display. 

    Delivery of 8-bit images

    When sending simple image files to another person, JPEG is usually a highly appropriate delivery format. Unless one of the more exotic flavors is used, like CMYK, or perhaps the JPEG 2000 format, JPEGs are the most dependably accessible format for images. When saved with lossless (or nearly lossless) compression, they provide extremely high quality. 

    Camera-created originals

    Most cameras allow capture in JPEG format. To create the JPEG, information from the raw capture must be discarded, so it is inherently lossy as a capture format. However, if exposure and color balance are properly set, the resulting JPEG may provide a perfectly acceptable file for nearly every purpose, compared to a raw file. Capture should be done at the largest size and quality setting that is practical. JPEG cannot contain depth information or multi-frame captures, so it is not a suitable format for capture when that information is desirable. 

    Library originals for read-only image applications

    Camera-created JPEGs as described above make an acceptable format for archiving when used in a strict read-only workflow.

    JPEG 2000 

    In the year 2000, the JPEG committee created an update to the JPEG specification which offered several improvements to the original standard. JPEG 2000 increased support for bit depth of 16 and 32 bits per channel. JPEG 2000 also modernized the compression algorithm, increasing efficiency and making a lossless version a universal part of the standard along with several other modernizations. 

    Adoption of the JPEG 2000 standard was originally hampered by uncertainties in the patent rights to certain features. While those patents are probably not at issue any longer, they prevented the use of the format at the time when it was most opportune. No cameras capture in JPEG 2000, and software support for the format is limited. Now that image objects are growing more complex, and as storage efficiency becomes less important, other formats offer some clear advantages over JPEG 2000. (TIFF is more capable as a full-service container, DNG is a better container for camera raw data and for some archived data, JPEG and PNG are better formats for electronic publication due to wider support, HEIF supports complex image objects, etc.) JPEG 2000 does have some significant adoption in the library and cultural heritage sectors. The high bit-depth and lossless compression made it a good choice in the early 2000s, particularly for institutions where adequate storage for TIFF files presented a financial challenge.

    In the next post, we take a look at PNG, a more modern format designed for use on the internet.

  • GIF

    This week, we’re going to dive into some specifics of image formats. In each post, I’ll give an overview of a format structure and its capabilities. This will be followed by examples of good uses of the format. We will start with the most basic image format, GIF, and move to more complex ones.

    The Compuserve GIF is one of the oldest image formats. As a still image format, it’s largely obsolete. It uses an indexed color model that does not reproduce photographic color accurately. And it does not support robust metadata embedding. For many years, it was the subject of patent dispute which made developers wary of including it. But it does do one thing well and continues to be wildly popular for that function. 

    GIF supports the inclusion of multiple images in the file which can be used to create animations or short videos. Due to the small file size, variable frame rate, and widespread support, GIFs have become an extremely popular format for moving images. In particular, GIFs are often used in social media memes. A quick web search will find many sites offering pirated snippets from movies and video that can be used to communicate some comment or emotion. 

    If you have still image GIF files you want to preserve, you will probably want to convert them to a more modern format. Full color photos are better saved as JPEGs; screenshots, drawings, or images with a lot of text are better saved as PNG. 

    Best use for GIF

    There really is only one good use of GIF and that is as a low-quality, low-bandwidth way to send a short moving image. That said, the use of GIFs in visual communication is one of the most widely adopted forms of visual speech in the mobile era. GIFs, particularly those with popular culture references, are an excellent example of connotative communication. By referencing a television or movie plot, a GIF comment can draw a parallel to a complex story line. It can communicate character, plot, situation, emotion and more. 

    Pronunciation

    GIF – Graphic Image Format – It’s a hard g.