Measured Compressor Decay (release) Times
For a peak compressor block, I'm measuring with a scope the actual release times for different values of decay entered into the Decay (dB/S) parameter.
Compressor attenuation is set for -20dB. Up to values of about 400 dB/sec decay, total release times measured on the scope match the expected 20dB x sec/dB setting (e.g. about 50mSec at 400dB/sec). However, for Decay setting greater than about 400dB/sec I see no further decrease in total release time.
The capture window does show that different values are being written for each decay setting.
Any thoughts on why this is occurring?
I didn't have time to go through the calculations in detail, but I think you're running up against the theoretical limits of speed that the release algorithm can handle, due to the number system in the DSP.
Internal to the algorithm is a multiplication factor. When the release first happens, it is very fast (constant dB), so the first jump is quite large. Afterwards, the multiplication factor decreases gradually each frame, causing the level to change more slowly as it reaches the target.
Since the maximum number we can store in the DSP is 15.9999999, I'm guessing that perhaps the setting of 400 dB/s is causing some internal parameter used in the calculations to max out at 15.9999999. In other words, I suppose that first jump on the release has a limit to how big it can be. Any setting above that limit would cause the parameter to simply saturate, meaning all settings above that limit would behave the same.
I'll check with the software team about this, but that's my original gut feeling on the matter.
If the capability of the algorithm does not match what Sigma Studio indicates is a valid range for the control I'd like to know.
OK, I'll confirm.
This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.
Thank you,
EZ Admin
Please log in to reply.