The nonlinear inverse synthesis (NIS) signal coding and processing technique exploits that the spatial evolution of a signal in lossless optical fiber becomes trivial in the so-called nonlinear Fourier domain in order to combat the nonlinear transmission impairments that limit performance of modern fiber-optic communication systems [1]. It has been shown [2] that the NIS method can be effectively integrated with any modulation format, offering a Q2-factor improvement up to 4.5 dB, which is comparable with those provided by multi-step per span digital back propagation. However, the relatively high implementation complexity is a serious challenge in NIS-based transmission schemes because the nonlinear Fourier transform (NFT) usually requires O(N2) floating-point operations [2], where N is the number of samples. The linear FFT, in comparison, requires only O(N log(N)) operations. The practical development of NFT-based digital signal processing (DSP) techniques and, in particular, the NIS method all requires computationally efficient NFT algorithms. Recently, a fast NFT algorithm was introduced in [3], which reduces the required floating-point operations to only O(N log2(N)). Herein, we investigate the practical suitability of the fast NFT algorithm for NIS-based transmission schemes, and compare them with the well-developed piecewise-constant approximation (PCA) method [2]. We consider 56 Gbaud orthogonal frequency division multiplexing (OFDM) NIS-based transmission systems in burst-mode with QPSK, 16QAM, and 64QAM modulation formats (the channel model includes noise, see [2] for details). The block diagram of the simulation set-up is given in Fig. 1(a). The transmitter encoded the data directly onto the continuous part of the signal’s nonlinear spectrum using the inverse NFT. At the receiver side, after coherent detection, the NFT was performed to recover the signal’s nonlinear spectrum and then a dispersion removal step (which is trivial in the nonlinear Fourier domain) was applied to compensate for all deterministic nonlinear impairments.