Motion Estimation in MPEG I

Because of the importance of random access for stored video and the significant bit- rate reduction afforded by motion compensated interpolation, three types of pictures are considered in MPEG I: Intra-pictures(I), Predicted pictures (P) and interpolated pictures (B-for bidirectional prediction). Intra-pictures provide access points for random access but only with moderate compression. Predicted pictures are coded with reference to a past picture (Intra or Predicted) and will in general be used as a reference for future Predicted pictures. Bidirectional pictures provide the highest amount of compression but require both a past and future reference for prediction. Bidirectional pictures are never used as reference. In all cases when a picture is coded with respect to a reference, motion compensation is used to improve the coding efficiency. The relationship between the three picture types is illustrated in Figure 3.1. The organization of the pictures in MPEG I is quite flexible and will depend on application specific parameters such as random accessibility and coding delay.

Figure 3.1

In MPEG I, the motion compensation process is block based and is performed on the basis of blocks of 16x16 pixels. Motion compensated interpolation (also called bidirectional prediction in MPEG I terminology) is a key feature of MPEG I. In the B frame, for each input block, calculate the constructed blocks with (1)forward motion vector, (2)backward motion vector and (3)weighted average of the forward and backward prediction (1 and 2). The selection is based on the chosen error criterion. Motion compensated interpolation presents a series of advantages, not the least of them being that the compression obtained by interpolative coding is very high. It deals properly with uncovered areas, since an area just uncovered is no t predictable from the past reference, can be properly predicted from the future reference, It has better statistical properties since more information is available. Its also allows decoupling between prediction and coding (no error propagation). The trade-off associated with the frequency of bidirectional pictures is the following: increasing the number of B pictures between references decreases the correlation of B pictures with the references as well as the correlation between the references themselves. The typical sequences of pictures in coding and display order are shown in Figure 3.2.

I B B P B B P B B P B B I

0 1 2 3 4 5 6 7 8 9 10 11 12

Display Order

I P B B P B B P B B I B B

0 3 1 2 6 4 5 9 7 8 12 10 11

Coding Order

Figure 3.2: Typical Sequence of Pictures in Coding / Display Order

MPEG I provides the half-pel accuracy of the motion vector. Motion compensation with half-pel precision reduces the bit rate and improve signal-to-noise ration (SNR) at least 1 dB. For half-pel shifts, the pel values could be interpolated by several methods. Since the decoder uses a simple linear interpolation, there is little reason to use a more complex method in the encoder. The linear interpolation method given in the standard is equivalent to the following. Consider four pels having values A, B, D and E as shown in Figure 3.3:

A¢Ýh¢æB¢Ý

v¢æc¢æ¡@¡@

D¢ÝE¢Ý¡@¡@

Figure 3.3: Interpolation of Half-pel Shifts

The value of the horizontally interpolated pel is:

h=(A+B)/2

The value of the vertically interpolated pel is:

v=(A+D)/2

The value of central interpolated pel is:

c=(A+B+D+E)/4