这似乎是工具包中的一个缺陷，由于透视图。绘制数据时未将其裁剪到正确的限值。您始终可以将数据切片为正确的值：

import numpy as np
# define limits
ylim = (-10,15000)
zlim = (0,0.1)

x = hz 
# slicing with logical indexing
y = freq[ np.logical_and(freq >= ylim[0],freq <= ylim[1] ) ]
# slicing with logical indexing
z = z[ np.logical_and(z >= zlim[0],z <= zlim[1] ) ]
    
ax.plot3D(x, y, z)
ax.set_ylim(ylim) # this shouldn't be necessary but the limits are usually enlarged per defailt
ax.set_zlim(zlim) # this shouldn't be necessary but the limits are usually enlarged per defailt

set_ylim（）和set_zlim（）方法简单地定义轴的上下边界。他们不会为你修剪数据。要做到这一点，你必须添加一个如下的条件语句来修剪你的数据：

from mpl_toolkits.mplot3d import Axes3D

fig = plt.figure(figsize=(10, 5))
ax = fig.gca(fc='w', projection='3d')

for hz, freq, z in zip(all_hz, all_freq, all_amp):
    if freq < 15000 and z < 0.1:
        x = hz
        y = freq
        z = z

        ax.plot3D(x, y, z)
        ax.set_ylim(-10, 15000)
        ax.set_zlim(0, 0.1)

plt.show()