electro wrote:Would you consider adding a SSL harmonic distortion button to The Glue?
I want to keep The Glue a clean mastering compressor. In the modern version of the analog g-series hardware is both very low noise and very linear. I felt that this was the approach to use for The Glue as well, and keep all non-linearities, other than the gain reduction itself, away from the main signal path. This also keeps the plugin focused on clean mastering compression, and keeps the cpu ussage, and price of The Glue to reasonable levels, and removes the need to add latency that would be needed for bandlimiting.
In a little more depth: getting the non-linearities correct in the older models is quite involved. You cannot model such things with a fixed waveshaper before and/or after The Glue. You cannot simply put a sin wave through the analog unit and measure the harmonics, even if you do it at multiple amplitudes - this sort of thing just doesn't work and isn't modeling what is going on dynamically in the compressor. Each component's specific non-linearities and slew rate limiting have to be modeled and the structure of where each component is in the circuit and it's feedback loops preserved, otherwise you won't get the right sound. There are multiple variations to the build of the 4000 compressor, each of which has different components. Doing detailed analysis of each of them and picking the best one would have been very costly, which would have increased the price of The Glue. Consideration would then also have to be placed on oversampling options to band-limit any added non-linearities in the main audio path, and this would also introduce latency, so all up I felt it best to keep it pure and clean. I'll do this sort of modeling for future products where it's more critical to the character of the hardware.