Abstract: DCT coefficient data obtained by transforming inputted picture data at a DCT (Discrete Cosine Transform) circuit 2 is quantized by selected one of plural quantizers at a quantizing circuit 7 to allow the quantized data to undergo variable length encoding at a variable length encoding circuit 8 to output it. A class designation circuit 30 compares DCT coefficient data from the DCT circuit 2 with threshold values Th1, Th2, Th3 in macro block units to carry out class designation to output activity code AT. A code quantity estimation section 10 estimates total encoded data quantity when data corresponding to 1 video segment which is 5 macro blocks of DCT coefficient data which have been read out from a zigzag read-out circuit 3 is quantized by respective quantizers to allow the quantized data thus obtained to undergo variable length encoding to send, to the quantizing circuit 7, quantizer number for determining an optimum one of the quantizers in which the estimated data quantity is reference value or less.