Question about the Loudnorm to reververant speech

[daxiang1993]

Very high-quality dataset. I appreciate your great dataset and the codes for simulating the noisy and reverberant speech. Now I am trying to use your recipes to do speech enhancement and dereveberation.
I notice that you have used the normalization of loudness to the simulated reverberant speech to make its loudness the same as the loudness of clean speech. But you do not norm the noisy speech. Can you tell me the underlying reasons? Because LoudNorm is better for speech dereverberation but does not work well for denoising?
I would appreciate it if getting your kind reply.