A video transmission system includes an encoder to receive video images, and encode them for transfer to a decoder. The video images include macroblocks having data that are encoded according to a prediction mode, such as inter-prediction or intra-prediction. A bit estimator for the encoded macroblock within the encoder estimates a size for the encoded macroblock and predicts whether it will be an illegal size. The bit estimator applies a bit estimation function using a number of header bits and a number of texture bits based on the prediction mode and transformed data of the macroblock.