How can I filter images together in 3 dimensions

I have a sequence of images and I want to filter them in 3 dimension. The spatial filter is gaussian filter and for temporal filtering I use sinc filter so as spatial an temporal kernels are separable, first I implement gaussian filter for each image and then I save all images in a 3D array and convolve the array with a sinc filter in only t axis. But when I'm looking at the result all of my images become exactly simillar to each other like all of them are one image. Here is my code please tell me where I'm doing wrong and how can I fix it?

t = np.linspace(0,32,32)
h2= np.array(np.sinc(t-2.5))[:,None,None]
for filename in glob.glob('02291G0AR/*.bmp'):
    img = np.array(Image.open(filename).convert('I;16'))
    src = cv2.GaussianBlur(img, (3, 3), 5,cv2.BORDER_DEFAULT)
    img_array.append(src)

v = np.array(img_array)
k1 = signal.fftconvolve(v,h2,mode='same')
plt.figure(3)
plt.imshow(k1[0,:,:],cmap='gray')
plt.figure(4)
plt.imshow(k1[2,:,:],cmap='gray')