I know that it's recommended to turn off the de-esser when using the advanced clipper, but my mantra when building processing has always been to start at the end (the clipper) and work backwards. My aim has always been to get the audio going into the clipper to a point where the clipper does as little as possible, whilst maintaining loudness. With the need to pre-emphasise the audio, one of the key things to get right, is balancing doing something to reduce the highs before the clipper, or just letting the clipper deal with it. The clipper in StereoTool is fantastic and does a lot, but you can drive it even harder if you can play around with the HF levels.
So, I got to thinking. There are, I think 3 ways in which a de-esser that is handling the pre-emphasis could work (not including lots of very clever methods that could be implemented in DSP
)
The diagram below shows a normal pre-emphasis curve (red), the normal de-emphasis curve (blue) and the combined effect (green) which is, in this case, a flat frequency response.
The first thing that could be done to reduce the highs would be to increase the frequency at which the pre-emphasis kicks in. In the diagram below, this been done to the point where the overall response (green) is 6dB lower at the highest frequency.
This approach tends to reduce the lower frequencies as well as the higher frequencies.
A second method would be to flatten the pre-emphasis curve. This can't be done using analogue components but could be done in DSP. The diagram below shows this effect. The frequency at which pre-emphasis begins is unchanged, but the curve is flattened. Again the net effect at the highest frequency is to lower the overall response by 6dB.
Note that doing this has the odd side-effect that there is a slight bulge in the overall frequency response.
Finally, a low-pass filter could be applied to the pre-emphasis to reduce the amount of HF. The result is shown in the diagram below, again with a 6dB reduction of the highest frequencies in the overall response.
This would be my preferred method as the only effect is to lower the very high frequencies whilst leaving the mids less affected than in the other two methods.
My point in doing this is just to ask what method is employed in StereoTool. I was playing around with the de-esser and it appeared to have an odd effect on the overall frequency response if used to excess. I know Hans is a genius in these things but I wondered if it might be possible to choose which type of de-esser is employed, that is if it isn't already doing something far more sophisticated than my charts!