VideoMAC: Video Masked Autoencoders Meet ConvNets