machine learning - How to understand the batch size in CNN? -

when using theano conv3d2d.conv3d, found input condition of signal is

signals_shape=(batch_size, n_frame, n_in_maps, height, width) 

i think understand batch size mean, update frequency, why batch size here being dimension?

the batch_size number of examples going use minibatch. example, if batch_size 50, means training/testing 50 examples @ time. consider if had 2d matrix contain data. excel spreadsheet each row separate example , each column feature of example. same thing happening here, in higher dimension.