Skip to main content

RAW Images



  • Ian Wilson
    Top Commenter

    The human eye/brain combination is capable of all kinds of adjustments that we are not fully conscious of. For instance, if you look at a piece of white paper in different light conditions, you always perceive it as white, even though the colour of the light you are seeing is different in each circumstance. Your brain knows it is white and adjusts its perception accordingly. The camera, on the other hand, just records the light arriving at the sensor. (Of course, you may be using auto white balance, in which case the camera will adjust.)

    Also the response of the eye to different lighting levels is different from the response of a camera sensor, which is different again from the response of film. And the eye/brain combination is better at allowing for different lighting conditions than the camera sensor is - so if you look at a scene with bright parts and with deeply shaded parts, you do not necessarily see some parts as blown out highlights and other parts of blocked shadows. As you concentrate on the darker parts for instance you can quite likely see details that may not be obvious in what the camera records, but as you concentrate on the brighter parts of the scene, you are not aware of whether you are still perceiving the detail in the shadows.

    The raw data recorded by the camera sensor is not primarily intended to give you a look just like what your eyes see - it is intended to capture as much information as possible from the darkest to the brightest parts of the scene, within its technical limitations, such as the dynamic range it is capable of recording which may or may not be able to encompass the whole scene. 

    A JPG (I know you aren't asking about that specifically) aims to give you an image that as far as possible looks like what you see, straight out of the camera. A raw file doesn't aim to do that: instead it aims to record as much of the available data as possible, leaving you to decide how much to boost the shadows, how much contrast you want, how saturated you want the colours to be, etc, depending on your purpose for the image, whether that is a "true to life" rendition, an artistic interpretation, or whatever.


  • SFA
    Top Commenter


    Sensors produce data that then needs to be interpreted by a processing system. In camera process produced a jpg (usually). 

    Our eyes receive data passed on to the brain, which acts as the Processing system adding its own interpretation and enhancements. It should not be assumed that colour is perceived in the same way by all eyes and brains. Indeed we know it is not through "conditions" like colour blindness. 

    There is no certainty that what we see is the same as any people standing around us are "seeing" at the same moment.

    It gets more complicated than that if we start to consider physiological details too complex to cover here.

    It also may be a lot more complicated when "human memory" is involved in defining "what we saw at a specific moment". 

  • Marcin Mrzygłocki
    Top Commenter

    There are three basic aspects: bit depth, color space, and processing - first two are roughly how much of the original view has been preserved, third determines how this will be displayed. If you perceive RAW as "dull", then in my opinion this could mean either a very difficult scene that clips on the limits of dynamic range, for example, or the processing has been explicitly chosen to be dull - how you view the image is also a form of processing and maybe you have somehow a "flat profile" screen, for example?

  • Brian Jordan

    Stefan Weinmann. Good explanations here but I'm going even more basic.  A RAW file isn't an image.  It's bits of data.  Nothing more; nothing less.  RAW converters then interpret that raw data to show you an image.  How they interpret that data is dependent entirely on the whims of whomever wrote the conversion engine.  If the conversion engine cranked the greens and saturated the blues and crushed the blacks and cooked that in, you'd be presented with a truncated display of the original data which would negate the purpose of capturing in RAW.  (Essentially, that's a JPEG.)  Liken it to cooking:  You can always add salt once the meal is prepared but if you over-salt from the beginning, well there you are.  That's why Linear Profiles give you the greatest latitude for data manipulation - linear profiles are the unsalted dishes of the image world.  Kinda bland but much easier for each person to season to their taste.

    Remember, a camera doesn't 'see'.  It records the reception of photons.  Similarly, your eye doesn't 'see'.  It receives photons, converts them to electrical impulses, and passes them along to the brain for interpretation into the 'image' we perceive.  Your brain is the RAW converter but it has a lifetime of training and bias baked into it***.  But even that isn't "reality". You and I will look at the same scene and see it differently.  How do I know this?  I'm colorblind.  I don't see colors the way you do.  My interpreter - my RAW converter - takes those impulses and processes them differently.

    ***As an aside, many cultures didn't/don't "see" blue.  It didn't even exist in their language.  Here's an interesting article speaking to this: and another: Imagine if a developer from one of these cultures wrote an image processing engine.  How would their sky look?  How could it be our "blue" if "blue" didn't even exist in their language?

  • Stefan Weinmann

    Thanks all, for the insights and the help! Much appreciated!!!


Please sign in to leave a comment.