Figure 1: Schematic setup of the XRDT. Sample is rotated around the diffraction vector h || sin0X; F is the rotation angle, ? is the Bragg angle. D is the CCD detector; AS/PC is the acquisition system for accumulating of the 2D IP frames

Figure 2: Generating the pseudorandom numbers y (red curve) vs. the normal distribution PDF (blue curve) for the string numbers N = 961 × 10n for n equal to: a) 2, b) 3

Figure 3: The XRDT 2D IPs with noise level s equal to: (a) 0, (b) 0.01 (c) 0.016. The sample rotation angle F = 0°. The total grid sizes in the ‘own’ crystal coordinate system (X, Y, Z) are equal to (61, 61, 21). Sizes of the sample voxel {?X, ?Y, ?Z}: ?X= p/30, ?Y= p/30, ?Z= p/20, respectively. The vector P(true) of the 3D Coulomb-type point defect displacement function fCtpd(r-r0) assumed to be {1.50, 0.50, 1.80}

Figure 4: Statistical quit noise-filtering of the 2D IP frames. Noise level s=0.01. Rotation angle F= 0°. K is the 2D IP-frame pixel number, K= 31 × 31. N is the 2D IP-frames number. (a) no averaged noise 2D IP-frame, the string number, ñ=K, K=961, N=1; (b) the averaged noise 2D IP-frame in terms of equation (5), the string number N = KXN, K=961, N = 961. The grid sizes in the ‘own’ crystal coordinate system (X, Y, Z) are equal to (31, 31, 21). The grid voxel sizes {?X, ?Y, ?Z}: ?X= p/30, ?Y= p/30, ?Z= p/20. Wolfram program code [19] has been applied for generating the pseudorandom numbers{y_k}