Abstract: Amplification of a pure state by any linear deterministic amplifier always introduces noise in the signal and results in a mixed output state. However, it has recently been shown that noiseless amplification becomes possible if the requirement of a deterministic operation is relaxed. Usually, nondeterministic amplifiers rely on using single photon sources. In contrast, we propose a noiseless amplification scheme where the energy required to amplify the signal originates from the stochastic fluctuations in the initial field itself. The proposed scheme consists of devices that are generally used in quantum optical experiments i.e. beam splitters, photodetectors, and a quantum nondemolition (QND) measurement apparatus. A QND measurement utilizes the energy fluctuations of the initial field to replace the single photon source that would otherwise be needed as in the scheme suggested by Zavatta et al. [Nature Photonics, vol. 5, pp. 52-56, 2011]. The operation of our proposed amplifier is shortly described as follows: first a single photon is subtracted from the initial field by a beam splitter and it is verified using a QND measurement apparatus. Second, the subtracted photon is added back to the field by a beam splitter and a photodetector. Finally, another photon is subtracted from the field using a beam splitter and a conventional photodetector. The resulting output field is an amplified coherent field with high fidelity. We apply the Wigner function formalism to analyze our noiseless amplification scheme and to investigate the relation between the amplification and its success rate as well as the statistics of the output states after successful and failed amplification processes. Furthermore, we also optimize the setup to find the maximum success rates in terms of the reflectivities of the beam splitters used in the setup and discuss the relation of our setup with the previously reported setups. In addition, the states after failed amplification are examined and the possibility of repeated amplification process is discussed.