Let’s see a picture of just exactly how the use of a dynamic range reduction compressor helps, when we have more severe impairment.
Here is the graph from before, showing how variations around the correct value for vTuning affect the perceived loudness:
The green curves show our perception when vTuning ranges from -3 dB to +3 dB, in 1 dB steps around the correct value of 60 dB. The black line is the ideal, and is matched when vTuning = 60 dB.
We want to focus on the region in the vicinity of 30-40 dB. There you see quite a wide range of perception when we have an incorrect vTuning level by only a few dB.
The next graph shows what happens if we pre-compress the music ahead of Crescendo, using a compression ratio of 1.5 to narrow the whole dynamic range to 2/3 of its original range, pivoting around the 0 dBVU level of 77 dB:
You can see that sound levels around 40 dB are now much less affected by errors in the vTuning level. We still have some difficulty for sounds near 30 dB, but truly, that is whisper level. So that may not much matter for us.
The minor “kink” in these curves, at 53 dB, is caused by our compressor threshold 24 dB below the nominal 0 dBVU level at 77 dB. The compressor gives up below that, and just supplies a constant boost of 8 dB.
Dynamic range reduction pre-compression sacrifices some of the original dynamic range, most noticeable at the loud levels. But for such a mild compression ratio as shown here, it seems a reasonable compromise in order to gain more control over the faint details in the music.
But, you be the judge…