CompView
View compression gain reduction
CompView
Features
Background
Both the CompView analyser and the Versatile Compressor plugins were originally designed and released as part of the entertaining and intentionally hyperbolic "compressor scam" YouTube video series. The goal was to demonstrate how a compressor with wider parameter ranges and a few additional shaping features was able to emulate other expensive compressors.
Source code
More Detail
CompView is a simple plugin which allows you to see the exact gain reduction applied by a compressor. This is not the same as the diagram you see on some compressors with a graphic display. This is the exact gain reduction applied. It requires a test tone to work, which is provided along side the compiled version. Note that your session must be set to 96k for useful results and the two lines in the plugin work on a dual mono basis, meaning you should pan one compressor hard left and the other hard right.
Some people have misinterpreted this as a claim that the compressor, in combination with CompView, can faithfully and completely emulate any compressor in the world. This was never the claim and it is also completely not the case. There is a big limitation of CompView in analyzing gain reduction, specifically because it needs to use a very high pitch test tone for technical reasons, and the compressor you are analyzing is very likely applying frequency dependent compression.
TECHNICAL LIMITATIONS AND POSSIBLE IMPROVEMENTS
You can experiment with reducing the frequency of the test tone but you will see that the lower the frequency, the more grainy and inaccurate the curves are. More specifically, the wavelength of the test tone must be significantly shorter than the fluctuations in gain that we are looking to inspect, otherwise the gain fluctuations of the compressor are obfuscated by the movement of the waveform of the test tone itself.
I have considered making a new version of the compressor which has a sandwich type setup, like Bertom EQ Analyser, which produces a test tone and then monitors the deviation from the original tone, and then the gain reduction curve could be plotted as the direct delta between the original and processed tones. This would not only provide superior resolution without the need to have a 96kHz or higher session sampling rate but also allow the user to send anything through the compressor without the analyser plugin freaking out and plotting fuzzy lines. It’s the correct and better way of dealing with this problem and I’m interested in coding this.
This may solve the resolution problem in one place but unfortunately overlooks the exact same problem in another place… the processing of the compressor itself. Even when considering a given compressor as a black box, its gain detection circuit cannot have a “superior resolution” reacting to fluctuations in its program material than the wavelength of the program material itself. This raises the question of using noise bursts with pre-emphasis EQ, the relevance of impulses and so on. I’m sure with this upgraded approach, many more questions could be answered and this might even be a useful tool for developers of compressors, more than it is an analysis tool to emulate them with free plugins.