CNN with LSTM input shapes

audio
9

I am extracting frames from videos. Then each frame is fed into the CNN to get the features and the output from CNN is fed into LSTM. So how can I change the seq_length?

Thanks in advance.
Regards,
ananda2020