Matplotlib : Live animation works fine but display

I made a little forest fire animation. My code is at the end of the question.

Here is some information before I ask my question :

No tree : forest[i,j] = 0
A tree : forest[i,j] = 1
A tree on fire : forest[i,j] = 2

Basically what happens is that constructforest creates a 2 dimensional array called forest of size n by m with a probability of tree occupancy called p. After that setonfire sets on fire the forest and while the forest can burn spreadfire spread the fire.

When I run forestfire with the Python prompt or the IPython prompt I get a nice animation but when I go check the video file that I saved I only see a blank plot.

I did some research, I found many questions about this issue but none of the advice I read was helpful:

matplotlib animation produces a blank
Matplotlib animation not working in IPython Notebook (blank plot)
Animation from matplotlib not working in spyder
Spyder Python Animation not working

Can someone tell me what is going on please ?

forestfire.py

from random import random

import numpy as np

import matplotlib.pylab as plt
import matplotlib.colors as mcolors
import matplotlib.animation as animation


def hazard(p):
    r=random()
    assert p>=0 and p<=1
    return r <= p


def constructforest(n,m,p):
    forest = np.zeros((n,n))
    for i in xrange(n):
        for j in xrange(m):
            if hazard(p):
                forest[i,j] = 1
    return forest


def setfire(forest,i,j):
    forest[i,j] = 2
    return forest


def spreadfire(forest):    

    n,m=forest.shape
    c = np.copy(forest)

    for i in xrange(n):
        for j in xrange(m):

            if c[i,j] == 1:

                Y, X = xrange(max(0,i-1),min(n,i+2)), xrange(max(0,j-1),min(m,j+2))

                for y in Y:
                    for x in X:

                        if c[y,x] == 2:
                            forest[i,j] = 2                        
    return forest


def canburn(forest):    

    n,m=forest.shape
    c = np.copy(forest)

    for i in xrange(n):
        for j in xrange(m):

            if c[i,j] == 1:

                Y, X = xrange(max(0,i-1),min(n,i+2)), xrange(max(0,j-1),min(m,j+2))

                for y in Y:
                    for x in X:

                        if c[y,x] == 2:
                            return True                      
    return False


def forestfire(forest):

    fig, ax = plt.subplots()

    movie = []    

    # Colormap
    red, green, blue = [(1,0,0,1)], [(0,1,0,1)], [(0,0,1,1)]   

    colors = np.vstack((blue, green, red))
    mycmap = mcolors.LinearSegmentedColormap.from_list('my_colormap', colors)

    # Initialization
    k = 0

    forest = spreadfire(forest)

    im = plt.imshow(forest, animated=True, cmap = mycmap, interpolation="none", origin='lower')
    movie.append([im])

    # Fire propagation
    while canburn(forest):
        k += 1
        print k

        forest = spreadfire(forest)

        im = plt.imshow(forest, animated=True, cmap = mycmap, interpolation="none", origin='lower')
        movie.append([im])

    return animation.ArtistAnimation(fig, movie, blit=True, repeat_delay=100)



ani = forestfire(setfire(constructforest(101,101,0.4),50,50))

ani.save("forestfire_test.mp4", writer = 'ffmpeg', fps=5, dpi=500)

EDIT

As requested by @Y.Luo by @ImportanceOfBeingErnest in the comments I downgraded matplotlib to 2.0.0 and I changed the framerate of the animation but forestfire_test.mp4 still displays a blank plot.

Here are my settings: