Align output levels for binaural rendering

While comparing the output levels of renderers for uncoded HOA3 -> BINAURAL, it was observed that the output levels of the Fastconv and Parametric renderers were lower than the time domain rendering using the python scripts/external renderer (Crend) by approx. 1 and 2.5 dB respectively.

Probably a compensation factor should be applied to align the levels of the renderers. Something seems to already be applied for the Parametric binaural renderer here: https://forge.3gpp.org/rep/ivas-codec-pc/ivas-codec/-/blob/main/lib_dec/ivas_dec.c#L321

It would be good to verify the following

• Output levels of all renderers (Crend, Fastconv, Parametric, TD Object) are in line with each other
• Transformation of the HRTF should ensure consistent output levels (possibly via a compensation factor), i.e.
  • Whether transformed to SHD from SD (i.e. HOA3 HRTF representation)
  • Whether transformed from SD or SHD to CLDFB (for Fastconv and Parametric renderers)