Pseudo-Analog

D_d

R(D_d) =

(1)
¹ log ^λ , 0 ≤ D_d ≤ λ,

real pixel values.
Contributions:

1. 
   A novel design of SDR-based pseudo-analog wireless video transmission system is proposed in which the graphical interface can greatly reduce the operation dif- ficulty. It provides a convenient and effective experi- mental platform for the research of pseudo-analog video transmission technologies, providing a seamless switch
   

0, D_d > λ,
where D denotes the distortion caused in the transmission process. R(D_d) denotes the minimum source coding rate required to achieve the distortion D_d. According to the capac- ity of additive white Gaussian noise (AWGN) channel, i.e., C = log(1 + γ)/2 and the rate-distortion formula given in Eqn.(1), the minimum distortion of digital communication can be denoted as follows.

between simulations and experiments.

2. 
   From the perspective of Shannon theory, we reveal the
   

λ
D_d =
1 + γ
, (2)

reason of the inherent performance loss caused by the separate source and channel coding which is adopted in the conventional digital transmission system. In addition, we explain that the analog method can also obtain the same optimal performance as digital method from the perspective of rate-distortion theory.
The rest of the paper is organized as follows. Section II
where γ represents the channel signal-to-noise ratio (SNR). Note that Shannon theory is actually idealized because it assumes that the channel condition and the source distribution remain unchanged, and the length of the source code is infinite. However, the physical layer can only process signals with limited length in practice. Therefore, in some specific communication scenarios (e.g., wireless video broadcast), such

a n factor, i.e., ^√ , for the source . Therefore,
separate source and channel coding system may result in inherent performance loss. Fortunately, Goblick et al. have proved that linear analog coding can also achieve the optimal performance [13]. Consider the case of transmitting the source X in AWGN channel using analog coding, that is, transmitting the source X directly after power scaling. Assuming that the total transmission power is P and the channel noise power is σ² . In order to minimize the distortion, the transmitter assigns
scaling G = P/λ X
the received signals can be represented as follows.
Y = GX + W, (3)
where W denotes the Gaussian noise. According to the min- imum mean squared error (MMSE) estimation criterion, the estimation of X can be denoted as
Gλ

n
X_MMSE = _{P + σ2} Y. (4)
Therefore, the distortion using the analog method can be calculated as follows
D_a = E(X_MMSE − X)²

blocks with uniform size. The transmitter allocates different transmission power for each divided block according to its pixel variance. Then, Hadamard transform is performed for each block to reduce the peak-to-average power ratio (PAPR). Finally, the transmitter sends these transformed coefficients in high-density modulation mode. At each receiver, a series of operations are performed in order including the signal de- modulation, inverse Hadamard transform, MMSE estimation, inverse 3D-DCT transform.

A large number of variants of SoftCast have also been pro- posed. In [23], a pseudo-analog transmission system called D- Cast is proposed where the correlations between video frames are fully utilized to enhance the video quality. Specifically, the received video frames are considered as the side information to assist with the reconstruction of the current frame. A data- assisted cloud radio access network for visual communications named DAC-RAN is proposed in [24]. DAC-RAN separates the control plane and the data plane in the conventional digital transmission infrastructure, and integrates a new data plane (that is specifically designed for video communications) into the virtual base station. The correlated information retrieved