do you think you can build a machine learning model for analog amplification equipment, like saturation and such. feel in terabytes of labelled raw vs amplified audio and then you can have a digital representation of the process. these AI toys are all about images these days and not many about audio