Systems, methods, and apparatuses for compressing a video signal are disclosed. In one embodiment the method includes transforming at least a portion of the video signal to produce matrices of transform coefficients, dividing the transform coefficients by at least one quantizer value to generate matrices of quantized transform coefficients, and identifying an encoding context including the transformed portion of the video signal. The method may further include determining a cost to entropy encode a non-zero value occurring within the encoding context, determining a first range including zero and having a width that is a function of the determined cost to entropy encode a non-zero value, and setting to zero at least one transform coefficient that falls within the first range.