I finally managed to reproduce almost exactly jit.scope's waveform mode by reverse engineering it with little matrices.
The aim was to be able to output this waveform as a matrix and use for something else than just analysing.
I first made a lazy java prototype as I had no idea how to implement this in jitter directly without OpenGL (I still don't).
Then I made my first jitter object in C.
From my perspective the [jit.wavescope] is extremely useful and I'm still using this external in some projects. Leopold, thanks for creating it.
I just updated original code a little bit (hope it's not prohibited). Changes are rather "cosmetic", but important for my workflow, and maybe for others... So, I updated the source for latest SDK and 64bit (what was the main reason to dig into it) and tweaked routine to avoid "noisy" artefacts in produced image (especially when "accum" is set to relatively high values) - now [jit.wavescope] act even more like [jit.scope] (but you can use "legacy_mode 1" message to restore previous behavior).
Source, compiled external and help file are in attached ZIP. OSX only, but it should be simple port it for Windows.