本文利用计算机模拟对各种类型的编码方案,包括最佳线性预测器、自适应量化器、帧内自适应预测器以及帧内-帧间自适应预测器进行了比较。主观试验表明,与其它帧内编码方式比较,建立在轮廓预测基础上的自适应预测器对分辨率测试卡的预测有明显改善。采用这种预测器及Huffman码,为获得演播室图象质量,每个亮度信号抽样需要3.7bit。进一步的研究可在此基础上进行。 This paper uses computer simulations to compare various types of coding schemes, including optimal linear predictor, adaptive quantizer, intra-frame adaptive predictor, and intra-frame adaptive predictor. Subjective tests show that, compared with other intra-frame coding methods, the adaptive predictor based on contour prediction shows a significant improvement on the resolution test card prediction. Using this predictor and the Huffman code, in order to obtain studio image quality, each luminance signal sample requires 3.7 bits. Further research can be based on this.