MIP mapping — From the Latin phrase "Multum in parvo", meaning "a multitude in a small place," - is a method of texturing that uses multiple copies of the same texture with different levels of detail. It means that the level of details of textures will change depending on the distance from the camera. The technique allows to get rid of "noise" in remote objects and significantly improves rendering performance.
MIP mapping improves the quality of textures by using textures with different resolutions for the different objects of the same picture, depending on its size and depth. Thus, the memory contains multiple copies of the textured image in different resolutions. As a result, the image quality remains high whether it is close to the camera or far away from it. With this method, you will see the image in high resolution when standing close to the object, and a low-resolution one when looking at it from the distance. MIP Mapping reduces noise effect that appears during texture mapping. It uses some clever techniques to compress the data of the texture images in memory. In order to use MIP mapping it is required to take all the texture sizes and multiply that number by two and then create one large map. All the other maps are usually filtered and become averaged reduced versions of the largest map.
Image looks best when detailing of the texture is close to the resolution of the screen. If the resolution is higher (the texture is too small / object is too close), the image turns blurry. If the texture resolution is too high (the texture is too big / object is too far away), we get random pixels which means the loss of small details, flickering and problems with cache. It turns out that it is better to have several different textures with different level of detail and use the one which resolution is as close to the resolution of the screen as possible.
Depending on the performance of your PC you can enable and disable this function in Settings.