Standard video applies a contrast curve at capture, which looks good immediately but discards highlight and shadow information permanently. Log encoding instead distributes tonal values logarithmically, roughly matching how sensors respond to light, so far more of the captured range survives into the file.
The result looks wrong on purpose — washed out, grey, desaturated. That flatness is the point: it means highlights have not been clipped and shadows have not been crushed, leaving room to place contrast deliberately in the grade. Every manufacturer has a variant, with S-Log, C-Log, V-Log and Log-C among the common ones.
It is not free. Log needs correct exposure, usually rated slightly brighter than metered, because underexposed log lifts into noise immediately. It needs a codec with enough bit depth — 10-bit at minimum, as 8-bit log bands badly. And it adds a mandatory grading step, which is why it is the wrong choice for fast-turnaround work that has to be delivered flat out of camera.