Raw vs. Rendered

Raw capture

We like to think of raw capture as the ultimate “latent image”. The data can be processed as many times as one chooses from the original capture information.

The dpBestflow® recommendation is raw capture for most job types and most workflows. The post-processing flexibility and image quality usually override the additional time raw requires over JPEG capture. If the original JPEG capture does not have perfect white balance and exposure, then making batch corrections can take as long (or longer if the corrections are done at the pixel level) than with raw, and the results will be inferior.

There are two types of raw format: proprietary raw files – which we will simply refer to as “raw” – and the standardized non-proprietary raw format DNG (digital negative.) At this juncture, both DNG and raw are in use as a capture format although proprietary raw format is much more common, especially in mainstream camera models such as Canon, Nikon, Sony, Olympus, Phase, Leaf, etc. Only a few camera models, such as Casio, Leica, Pentax, Ricoh, and Sinarback eMotion cameras shoot DNG as a capture format.

Raw capture advantages

There are several advantages that this file format offers:

• Maximum image quality can be obtained.
• Image quality of legacy files can be improved over time as parametric image editors improve.
• White balance can be applied after the fact, alleviating concern for “nailing” white balance on capture.
• Output can be 16-bit depth, a major advantage for post-processing.
• There is good flexibility in adjusting exposure and brightness values.
• All adjustments are non-destructive, meaning the digital negative can be reprocessed an infinite number of times in a variety of software applications.

Raw capture disadvantages

There are also some disadvantages to consider:

• Raw format is larger than JPEG (but smaller than TIFF). Storage media fill up faster and camera buffers fill faster, meaning that fewer shots can be taken before the camera will pause.
• Proprietary raw formats may become unreadable at some point in the future.
• Processing raw files requires a computer, special software and time.
• Raw files require more storage space than JPEG files (but less space than TIFF).
• Proprietary raw files can take a while to be supported by third-party processors. Buying a new camera may break your workflow for a period of time.
• Proprietary raw files are not good candidates for storing custom metadata. In fact, most software applications refuse to embed any type of metadata in them, forcing the use of sidecar files or storage of the information in a database or folder.
• The only standardized raw file format, DNG, is not supported by all raw processing software, limiting the variety of parametric image editors that can be used with it. (The list, however, is growing).

Raw file processing improves over time

While significant improvements have been made in sensor hardware (e.g. allowing improved image quality even as pixel pitch has been reduced), an even greater amount of improvement has come from improvements in raw-processing algorithms. Image quality from older raw files continues to improve, lending weight to the concept of archiving raw files.

Another implication is that image quality will inevitably be a blend of camera capabilities and raw file processing. This factor will not be limited to improved demosaicing but will include improved control over noise and improved lens performance via software.

Although Photoshop filters and Photoshop plug-ins, such as PTLens, have been available for some time, similar corrections are now being included in parametric image editors. DxO Optics has been a leader in this arena. Lens corrections have been added to Adobe Lightroom 3.

An example of just how symbiotic hardware and software have become is found in the Panasonic LX3 camera’s dependence on lens correction algorithms to produce an undistorted picture.

JPEG capture

For many point-and-shoot cameras, JPEG capture is the only option. Some medium-format backs shoot only raw while others can shoot JPEGs, and raw + JPEG. Most DSLR cameras shoot raw, JPEG, or raw + JPEG.

The essential feature of JPEG capture is that the camera shoots a raw file and applies camera settings, such as white balance, sharpening, tone curve (contrast) and saturation (also known as picture styles — vivid, neutral, portrait, etc). The raw file is then processed out to a JPEG and the original raw data is deleted. The processed JPEG ends up in a user-determined color space, such as narrow gamut sRGB or wider gamut Adobe RGB (1998). Both of these color spaces will be smaller than the gamut potential of the sensor.

Cameras that shoot JPEG generally have size and/or quality options. Size refers to the pixel dimensions (resolution) of the captured file, and quality refers to the amount of compression applied to the saved JPEG. Most of the time you will want to shoot at the maximum image size. We think it is best to shoot at the highest JPEG quality as well since this results in the least amount of JPEG compression. Increasing JPEG compression will result in more visible artifacts.

Depending on the camera, there may be only three, poorly-defined quality levels available to choose from: fine, normal and low. If this is the case, it is always best to choose “fine.” More sophisticated cameras may offer a sliding scale (1-10) for example. Unfortunately, there is no defined standard for these sliding scales. If you wish to test JPEG files for artifacts at different quality settings, the best method is to shoot something with hard edges (like a building) and look carefully in the red channel for artifacts.

Although lossy compression is the main feature (some say benefit) of JPEG capture, there are other factors to consider. One is bit depth. Standard JPEGs provide 8 bits per RGB channel vs. 12-bit, 14-bit, or 16-bit depth per channel for raw capture. This lower bit depth puts more constraints on image edits. Repeated edits and JPEG saves will degrade images so care must be taken to avoid saving over JPEG originals.

JPEG capture advantages

There are several advantages that this file format offers:

• JPEG files are smaller than raw files so the storage media will hold more images.
• Most cameras can shoot JPEGs more quickly than raw, and camera buffers do not fill as quickly.
• JPEG files are quicker to transmit electronically due to their smaller size.
• JPEG files are immediately available for use, not requiring a processing step on a computer.
• JPEG originals, if shot carefully, can be of very high quality and are sufficient for many applications and job types.

JPEG capture disadvantages

There are also some disadvantages to consider:

• It is much more important to get the white balance correct when shooting. Accurate white balance requires constantly creating custom white balance settings. This can be time consuming unless the lighting is controlled and consistent.
• During optimization, adjusting parameters such as white balance, exposure, contrast, saturation, etc is less easily done (compared to what’s needed for raw) and can result in a deterioration of image quality.
• The camera determines image quality. There is less chance that image quality can be improved upon via software updates and improvements as is possible with raw files. All the capture and camera data is embedded (baked) into the pixels when the exposure is made.
• JPEG camera originals can be easily overwritten if picture edits are made and then saved with the same file name.
• JPEG files that are edited at the pixel level and re-saved will degrade slightly in quality with each save.
• Although parametric image editing of JPEG files is possible in parametric editing applications, such as Lightroom/Adobe Camera Raw, Aperture, Bibble and others, those edits are not as effective as they are with raw files. In addition, the edits are only visible in the program that created them or when they are output to a new JPEG or another standard file format. When PIE work is used for JPEG, it also requires paying close attention to how each of the applications preferences are set.
• The ultimate quality that the camera can provide is discarded along with the original raw data.

Raw + Jpeg capture

Many DSLR cameras can be set to shoot both raw files and JPEG files in a single capture. Camera settings, such as white balance, tone curve, color space, picture style, sharpening, etc will be reflected in the JPEG. The raw files will be available for further processing. Shooting raw + JPEG can solve some workflow problems, but create others.

Raw + JPEG capture advantages

There are several advantages that this file format combination offers:

• The JPEG file is immediately available for use.
• A raw file is available for additional processing.
• Two files captured for each exposure creates a back-up in the rare case of file corruption.
• Transfer times are reduced during tethered shooting — if only the JPEG file is transferred.

Disadvantages

There are also some disadvantages to consider:

• Storage media fills up even faster than when shooting raw only.
• Camera buffers fill faster.
• Transfer times are longer when both files are transferred during tethered shooting.
• If a software application is set up to treat raw and JPEG as separate files, it may number them as separate files, which can be confusing.
• It can be difficult to make raw files match the “look” of camera-created JPEG files, creating color consistency problems when matching camera-created JPEG “proof” files and final files. Clients who approve a shoot based on the camera-created JPEG files may be surprised by the appearance of the processed raw files.
• Additional storage space is required to archive shoot files.

The decision whether to shoot raw + JPEG will depend entirely on workflow needs. For some, it represents the best of both worlds, for others, the disadvantages of both.

TIFF capture

Although some earlier DSLR cameras offered a TIFF capture option, TIFF capture has largely given way to JPEG or raw/DNG capture. TIFF capture doesn’t have the compression and therefore small-size advantage of JPEG, nor the non-destructive adjustability of raw files, since TIFF capture bakes in camera settings in the same way as JPEG capture does. TIFF capture is more unwieldy than raw since it has been processed (demosaiced) by the camera so it will be approximately three times larger than a raw/DNG file. As the megapixel counts of cameras have grown, the corresponding TIFF size has made TIFF capture largely impracticable from both a speed and a storage standpoint.

Exceptions, but important ones, are digital scan back cameras and multi-shot cameras. These cameras do not use Bayer pattern filters. They capture RGB channels separately, either as a scan or as three separate channel shots. While the makers of these cameras sometimes refer to their output as raw data, files are actually three channel TIFF files.