Error Sensitivity of CCITT Standard Facsimile Coding Techniques

Abstract

The CCITT has defined Group 3 facsimile apparatus as that which digitally transmits an ISO A4 document over a switched telephone circuit in approximately one minute. Data compression is employed to achieve the reduced transmission time. Study Group XIV of the CCITT has drafted Recommendation T.4 to achieve compatibility between Group 3 facsimile devices. The standard data compression technique specified by T.4 is a one-dimensional coding scheme in which runlengths are encoded using a modified Huffman code (MHC). The recommendation also includes an optional twodimensional compression technique known as the modified READ code (MRC). It is recognized that the switched telephone network is prone to error when transmitting digital data at the standard T.4 data rate of 4800 bits/s. This paper evaluates the error sensitivity of the MHC and MRC when operating over a typical telephone circuit. The error sensitivity analysis is accomplished by means of computer simulation. The error performance of the two coding techniques is analyzed both quantitatively and qualitatively. The quantitative analysis is accomplished using the error sensitivity factor which represents the average number of incorrect pels in the output document caused by a transmission error. The qualitative analysis is based upon viewing actual error-contaminated images generated in the simulation process. Two separate analyses have been performed. First, error sensitivity data for both the MHC and MRC are developed under identical operational conditions, and their relative performance is discussed. In the second part four different techniques for processing the received facsimile signal (MRC), to minimize the subjective effect of transmission errors, are analyzed.