quant values

I was wondering what people mean when they are talking about quant 2 quant 3 and stuff like that. I really did search but I didn't know what terms to use. Sorry if this has already been discussed. Here is a link for an example.

As I understand it, a video frame is divided into individual 8 x 8 macroblocks which are fed through an FDCT. The coefficients from the FDCT are rounded to integers before being stored. The quantizer is a scalar divisor which is applied prior to the rounding. If the quantizer is 1, the coefficients are simply rounded to the nearest integer and stored. If the quantizer is 2, then they are first divided by 2, then rounded and stored. If the quantizer is 3, they are divided by 3, and so on. (The decoder reconstructs the coefficients by multiplying the stored values with the current quantizer value for each macroblock.) In short: the higher the quantizer, the less accurate the stored coefficients are. Among other things, this can lead to visible edges appearing where one macroblock meets an adjacent macroblock (a.k.a., blockiness).

So then you put those values under the quantizer settings fo I and P frames under the quantization tab. Alright , thanks guys.