An illustration is given in FIG. 18 of a PRACH preamble timing estimation with an interlaced frequency domain mapping. Here, the sub-carriers which are not used by the PRACH preamble are zero when inserted into the IFFT. This leads to a very wide IFFT spanning up to the whole system bandwidth.

Timing estimate errors are illustrated in FIG. 19, both with narrow IFFTs, as described in FIG. 12, and with a wide IFFT, as described in FIG. 18. Here, the accuracy of the timing estimation is increasing with increased oversampling. However, a wide IFFT results in the most accurate timing estimations.

Computational complexity analysis is used as a decision basis for receiver algorithms together with performance evaluations. A coarse evaluation of computational complexity is done below in which computational complexity of FFTs are approximated by a radix-2 FFT. The number of real valued multiplications by IFFT and absolute square equals:
2N_FFT log₂(N_FFT)+2N_FFT
where N_FFT is the size of the IFFT. The number of real valued multiplications can be measured in terms of Multiplications and Accumulations (MACs).

A receiver with narrow IFFT and oversampling has the following number of MACs:
N_RB(2N_IFFT log₂(N_IFFT)+2N_IFFT)
where

• • N_IFFT=12·D_OS is the size of the IFFT,
  • the oversampling is parameterized with D_OS=1, 2, 4, 8 or 16, and