right let's look at representing images in binary specifically the type of image we're looking at is a bitmap which is stored as an array of individual pixels a 2d array in practice and columns and rows of pixels and another type of image is a vector image but virtual process is quite different to this so we are going to in this video so as a pixel pixel is an individual picture elements where Peaks will come from picture element and there's no definitive definition really but it's basically the smallest distinguishable feature of an image so monitors also and VD use like a visual display units is a technical term though divides into rows and columns of pixels so we a color will kind of take up a single pixel and we don't divide it any further and as you would have seen the size of an image or a monitor can be expressed with pixels in form width and height so width times height and this it can also be referred to as a resolution of an image so resolution really is all about detail and image hold so kind of a density of pixels so resolution again a few measures of what this is it can be written like this for if the x height in pixels or it can be in terms of things like PPI pixels per inch or dots per inch so a few measures but really it's what about for detail an image holds here you can see some settings from my computer with a resolution in terms of the pixels of width times height and also color depth which will occur now second and also you can what resolutions in terms of pixel entering this is from Photoshop so it varies the measure of resolution does vary so to actually get an image in the first place in the camera too huge a simplified process effectively a grid is placed over an image of each box representing a pixel so an average color isn't found for each pixel and a bount and a binary value is assigned to this color so a color will have a unique binary value a shade of color so tell you for black and white this is very simple it can just be the two options so it's either gonna be 0 or 1 so it could be 0 for Y it could be 1 for black but it doesn't matter matter which way round it is but for color it's more complicated each pixel is gonna be represented by multiple bits to represent more than two colors so it's one combination per shade and the number of bits each pixel has to represent the color is called for color depth or bit depth so the black and white this is just one color therefore one is only one bit used but if we had a color depth of three we can represent after eight colors because 2 to power 3 is 8 so a larger color depth allows us to represent more shades and different colors if I computer one over 6 and binaries blue one is black zero is white and two is yellow so just know for that's the value associated with that color so it can recreate the image look at this visually we've got three different images of three different color depths first has got four second eight third twenty four so the first image here which is very very grainy it's pretty much black and white this has only got 16 colors in this whole image because they're only 16 colors we can represent with 4 bits to 2 power 4 so it's not very good quality with 8 is actually quite a lot better so you have 256 colors here this represents 4 main colors that most images have I suppose but if you went all the way over up to 24 which due to the nature of exponential growth is about 17 million colors so a huge increase this is pretty close to the actual image there's all the subtle greens and so on can be represented clearly ris has a drastic increase and file size so there's always a downside to increase quality automat FEMEN if you increase the resolution and/or Carla debt if you do get a better quality image but this increase to the file size and so work this out you can it was a very intuitive formula for width in pixels x for height in pixels x for color depth in bits which makes sense obviously for resolution times for color depth because for each pixel we've got a assign however many bits for color depth is to it in reality is mostly going to be in well kilobytes and megabytes perhaps but invites you can just divide by 10 because there's 8 bits in a byte and divide this 4,000 4 kilobyte and a thousand 4 megabytes so on almost subjects about file sizes we're mentioning metadata which is really important in loads of aspects of computer science but this is data about data so we need to store this alongside a raw image data basically the array of pixels and we need to display the image property the computer needs some other details in order to actually show you the image on your screen and process it so a few visa kind of required most data attributes are things like the file name file format that's true for any files but in images specifically we have caller debt for an Internet and the resolution because it needs to know first of all what kind of size of the image is going to be and in - no kind of when a pix was finished so if a colleague that furs - it knows for every second pixel it needs to change color and so on so it needs to basically use information to format the image to display to you and there are other kind of normal required metadata attributes and get new things like camera details so apertures and the making for model and so on which can be used by applications but will increase the file size marginally it's not going to be a huge difference but it's not required you don't need this religious if you if size was completely a premium you wouldn't add all this extra mesh data