I have a really simple setup: Cubase v5.5 32-bit float, no effects on any input/output tracks. On one location in the song the master bus peaks at +0,1 dBFS. Then I unmute two acoustic guitar tracks panned hard L and hard R without touching any volume faders and now the master bus peaks at -0,2 dBFS. I'm able to repeat this unlimited nr of times/every time. By adding more signal to the setup I've mysteriously been able to remove the clipping. This in turn of course allows me to boost all tracks in the setup by +0,2 dB, since I now have 0,2 dB of additional headroom by adding the acoustic guitars to the mix. Are the acoustic guitar frequencies somehow creating some resonance on the mix as a whole smoothing out the summed transients? I've tried different meter configurations to try forcing a different output, but the same data remains. This applies when I have the pan law setting set to -4,5 dB or -6 dB. However, when I set the pan law to -3 dB or -0 dB the peak difference increases from 0,3 dB to 0,7 dB. This puzzles me. Am I able to get 0,7 dB of relative added headroom simply by having a particular pan law setting in combination with adding two tracks hard panned L and R? Also, lowering the project resolution from 32-bit float to 24-bit has no impact. But if I adjust the pan settings on the acoustic guitar from L and R to L50 and R50 then the margin is gone. At L75 and R75 the headroom margin is 0,1 dB, so with the tracks more panned I get more headroom. Do hard panned instruments in combination with -4,5 dB pan law result in mix headroom improvements of at least 0,7 dB? If I offset one or both of the acoustic guitar tracks by 1 frame when unmuted, I get a jump from -0,2 dBFS to +0,5 dBFS, that's a headroom difference of 0,7 dB! But at least then I don't get an improvement in headroom by adding tracks to the mix.. Another very strange thing here is that if I further attenuate the signal on the ac guitar tracks by -10 dB (hard panned, -4,5 dB pan law), then the mix peak is further reduced from -0,2 dBFS to -0,7 dBFS. That's all fine, except when I then further attenuate the tracks -10 dB more, (total of -20 dB attenuation), then the mix peak registers -0,2 dBFS again. This data confuses me, it's as if the summing engine is very random and non-linear, the data jumps unexpectedly... cellfone
↧