Difference of input dimension for simpleRNN and LSTM

Question

I have simpleRNN model like this.

input_len = 10
n_in = 44101
n_hidden = 512
model = Sequential()
model.add(SimpleRNN(n_hidden, input_shape=(inpu