David83
Advanced Member level 1
Hi,
To apply DFT transformation in OFDM systems, the convolution between the transmitted sequence and the channel has to be made circular. There are more than one way of doing so.
Suppose that the length of the data sequence is M, and the channel length is L. Then if we pad L zeros in the data sequence such that the data sequence length is N=M+L, and M zeros to the channel sequence such that the channel sequence is N=M+L, then transmitting the data sequence over the channel is equivalent to circular convolution. In this case we need N-point DFT. Another option is to add the last L symbols of each block of data sequence to its beginning, this is called cyclic prefix (CP). In this case, we need M-point DFT.
Now the question is: why most of the papers prefer to use CP over zero padding? Is the DFT size is the reason.
Regards
To apply DFT transformation in OFDM systems, the convolution between the transmitted sequence and the channel has to be made circular. There are more than one way of doing so.
Suppose that the length of the data sequence is M, and the channel length is L. Then if we pad L zeros in the data sequence such that the data sequence length is N=M+L, and M zeros to the channel sequence such that the channel sequence is N=M+L, then transmitting the data sequence over the channel is equivalent to circular convolution. In this case we need N-point DFT. Another option is to add the last L symbols of each block of data sequence to its beginning, this is called cyclic prefix (CP). In this case, we need M-point DFT.
Now the question is: why most of the papers prefer to use CP over zero padding? Is the DFT size is the reason.
Regards
