In a previous post I described how footage previewed in Sony Vegas would always appear different post-render. To begin shining some light on this problem I talked about the relationship between bits, pixels and the RGB colour model. This post continues the discussion. Note that any statement made in reference to an 'image' or 'images' is implicitly inclusive of 'video'.
PART 2: Colour management
Consider this; if you load this webpage on any two computers with different operating systems and compare the many colours outputted to recreate the image below, they are not going to be the exactly the same on both devices. The image may even look noticeably different.
Even though both computers most likely employ the RGB colour model to display colour, the images will still differ in some way or another. This happens for various reasons. For example, different types of LCD displays (RGB-LED, CCFL, etc.) can affect the accuracy of colour reproduction, and screen resolution (i.e. number of dots per inch) can affect just how accurately colours are represented. This issue extends not only between different computer models, but also a wide range of devices capable of displaying colour. Thank goodness, then, for colour management systems. Colour management systems ensure any given RGB value is represented as accurately as possible on devices including but not limited to image scanners, laptops, digital cameras, TV screens and printers. If you own a Windows-based computer, colour management is applied at the operating-system level through the Windows Color System (WCS) platform, whilst Mac-OS based systems use an application programming interface called ColorSync. Regardless of the device, colour management is an important consideration; after all, colour is of the utmost importance in the world of digital art. The main thing that concerns us, however, and the key to understanding my problem with levels in Sony Vegas, is the part of these colour management programs that defines the output behaviour of a particular device in relation to an absolute colour space.
An absolute colour space is a set of colours associated with absolute colorimetric quantities such that each colour within the space is completely unambiguous in its interpretation and display on a particular device. Each colour space is based upon a particular colour model such as RGB or CMYK and while many colour spaces may be founded upon the same colour model, the range of colours (called the gamut) defined within those spaces will differ. For example, the Adobe RGB and sRGB colour spaces are both based upon the RGB colour model, but the Adobe RGB gamut contains a larger range of cyan-greens than the sRGb gamut. Other examples of RGB colour spaces are CIE XYZ, ProPhoto RGB and CIE L*a*b*.
Adobe RGB and sRGB are available on many digital SLR cameras as a means of storing image colour information in the JPEG container format. Both have their advantages and disadvantages. By selecting a particular colour space preset on your DSLR, you are effectively limiting the range of colours that can possibly exist within that image to the gamut encompassed by the colour space opted for. This can be overcome, to some degree, by translating the image's existing colour space into a different colour space, but here you risk losing information. Why? Well, imagine capturing an image in Adobe RGB and then converting its colour space to sRGB. Due to the gamut limitations of sRGB, any colours in the image that exist outside the sRGB colour space will be converted to colours that exist within the sRGB gamut only. Those colours will be incorrectly represented in the new colour space. But say you capture an sRGB image and convert it into the higher-gamut Adobe RGB colour space. Surely the colour information is made more accurate? Alas, this is not the case. The algorithm which converts one colour space to another will inevitably produce so-called 'rounding errors' which fail to exactly represent the old colours within the new gamut, particularly where low bit depth is concerned. These errors are also present when converting from a higher-gamut to lower-gamut colour space. The best way to avoid the above problems is to shoot in RAW mode so that a colour space (or multiple, if desired) may be independently applied to the original footage type after that data has been captured, not during.
I have elaborated upon the finer points of colour space conversion and colour management in this post because, as a videographer and oftentimes photographer, it has the potential to occur at many points throughout my workflow, particularly if I am not dealing with RAW-format data. In the third and final post of this blog series, I will name and describe in more detail the different colour spaces used natively in the various digital devices and online mediums I employ from video capture to presentation, and therein reveal the simple solution which solved my problem with levels in Sony Vegas.
This post, like all others I author, is a work in progress; updated as needed. Was it informative? Easy to understand? Feel free to leave any comments or questions below!
Tim Szewczyk -