I’m currently playing around with the well-known Calf-Plugins, but I can’t find out how to master the Calf Stereo Tool.
I want to encode a stereo signal into a MS-Signal (say, M on the left channel, S on the right channel), then apply different EQ settings on M and S, then decode the whole thing back to LR.
To start the task, I added two Calf Stereo Tool plugins into my master strip, choosing “LR > MS” in the first one and “MS > LR” in the second one. I would expect that the overall sound does not change, but it does. Then, I generated a sine (with another plugin) which I can pan within the stereo range, while observing what the first Calf stereo tool does (on its meters):
Sine centered: The left output channel (M?) is louder that the right channel (S?), but I would expect that the M channel is loud and the S channel contains no signal.
Sine on left side: both output channels have the same loudness, while I would expect no signal on M, loud signal on S.
Sine on right side: both output channels have the same loudness (but less than with the sine on the left side), but I would expect no signal on M, loud signal on S.
Well, maybe I didn’t understand the whole concept of MS (Mid/Side), or maybe I only don’t understand the plugin? Therefore I kindly ask you whether there is some documentation on the Calf Stereo Tool plugin or on the Stereo Matrix concept which might help me understand the issue.
Middle / Sides - or even Lat(eral) / Vert(ical) to use the old Fairchild terminology - is just the sum and difference of the left and right signals.
Great, this is what I actually believe. So the “LR > MS” conversion can be described with M=L+R, S=L-R. This is revertable with L=(M+S)/2, R=(M-S)/2. I think usually the division by two is ommited, I added it just for mathematical correctness
Yet I don’t unserstand why the Calf Stereo plugin does not work as I expect it to do.
It is not out of the question there are bugs in it. I remember using it a while back for M/S decoding and finding a few issues I never bothered to report to them, sorry:(
I filed a bug report to the calf team. Even if there is no bug, I hope that the devs can point me to a useful piece of information.
you dicussed 3 usecases. But you think wrong in 2 of them:
1st: L=R In fact this is strange. If you used an external meter: Can you exclude possible high capacity between audio signal cables or some other mechanism that strays signal between the audio outputs? how high is the ratio of the amplitudes at master?
2nd: L=1 R=0 -> M=1 S=1 ->same amplitude, same phase!
3nd: L=0 R=1 -> M=1 S=-1 ->same amplitude, inverted phase!
Or did you remap the signals?
I use an M-Audio Audiophile 2496 soundcard with an Headphone amp and a connected headphone. I did the measuring completely within Ardour, using multiple buses. Therefore we can exclude any capacity-related effects. The amplitudes are around -20 dbFS.
Regarding the 2nd and 3rd case, you’re right. I already noticed my error.
Regarding the 1st case: I still wonder why the S signal is not zero (or at least near to zero). However, I also tried Steve Harris’ “Stereo Matrix: LR to MS” plugin, which behaves the same way. So did I misunderstood something here? By the way: I also created L+R and L-R signals by myself, using a lot of buses (an no plugins at all beside the sine generator). I fed the centered -20dbFS sine signal into the thing and found that L+R is -14dbFS and L-R is -∞dbFS, which is what I expected.