VCAs - two possible issues / bugs

Based on what another user told me I looked into a couple of use-cases where VCA and connected track-behavior seems to be wrong.

The basic function according to the manual is:

image

(I believe this should probably always be adds, and never subtracts, because of how math works. (+ - = -, ++ = +) Let me know if that assumption is wrong.)

Issue example #1 - repro:

  1. create an output bus with a metering plugin on it. Create an audio track with test generator on it and route it to the output, and a VCA track. In my example at the output bus the signal measures -10dBFS.
  2. create two automation points on the audio track and the VCA track each, make them 6 seconds apart from each other. Lower the second point on the audio track to -oo, lower the first point on the VCA track to -oo. This way they work ā€œagainst each otherā€.
  3. notice the output level as you locate within the range where automation happens. It will read -10dBFS at the start and slowly drop to -oo, because this is what the audio track output level does.
  4. connect the audio track to the VCA track. The level will now read -oo regardless of where you are located. This seems to make sense.
  5. Make a bezier curve on the VCA. Notice the level on the meter, notice the resulting automation curve on the audio track. Here at 3 seconds into the fades, first with a tiny movement of the VCA curve, second with a deeper move, but bezier in both cases:

Intuitively it seems both cases should yield -oo. If the VCA adds an offset to the audio tracks level then adding a lower negative value you should mean an ever lower resulting audio track value. Imagine the maximum allowable value is 100 and the VCA is at -50, the result should be -50 when added together. If the VCA is at -70 the result should be at -30. And if the Audio track started at -50 the result should be 0 in both cases, not increase in the second. It should bottom-out at -oo.

This seem like a clear error.

The next one is weirder…

Issue example #2 - repro:

  1. As before, create an output bus with a metering plugin on it. Create an audio track with test generator on it and route it to the output, and a VCA track. In my example at the output bus the signal measures -10dBFS.
  2. create two automation points on the VCA track, make them 6 seconds apart from each other. Lower the second point to -oo.
  3. create two automation points on the Audio track, make them 3 seconds apart from each other. Lower the second point to -10.
  4. Place the playback head slightly after the second point. Here is my reading:

The horizontal red line is for reference.

  1. Move the second automation point (at -10) back to the previous point. Notice how the resulting automation line rises, along with the output level:

As you can see the resulting line - at the same point where the playback head is located - is now higher, along with the output level.

But as the manual states the VCA value should be added to the original track’s value, and in both of the above cases, at where we are located, that audio track value is -10dBFS.

Am I missing something or is the above just wrong?

More, stranger weirdness - Continue from where we were above:

  1. Click on the flat automation line on the audio track before the playback head and verify the new automation point is -10dBFS. Output value is the same.
  2. Click on the flat automation line on the audio track after the playback head and verify the new automation point is -10dBFS. Output value is the same.

  1. Use the ā€œundoā€ function to undo step 7. Look:

Hey, just for poughkeepsie and giggles;

  1. Hit ā€œredoā€ to get that second automation point back:

So here we have 5 cases where my audio track’s fader is at -10dBFS, my playback head is parked at the exact same spot, and my output values are -38.5, -34, -34, -38.5, -38,

Please tell me I’ve been smoking too much cocaine while eating mushrooms, and now I’m both dumber than usual and hallucinating.

2 Likes

Not sure I’m with you, but with the VCA bezier curve and looking at the fader movements in Mixconsole it will be observed that the VCA fader raises to a higher level than the audio track as they work with each other, this implies that the level is perhaps correctly being multiplied by the VCA channel resulting in an audible signal before silence again when the VCA and Audio track again cancel each other out.

And it shouldn’t.

Look at the automation line that exists between the two automation points. Then drag the bezier-point between the two downward. If you look at the curve then at no point is it higher than the previous line between the two anchor-points on either. If the bezier curve is never higher than the straight line it implies that the level should always be lower than a straight line.

Do you not agree?

Can’t possibly say that, but I think it is working as intended. Unless Someone’s been putting Shrooms in my coffee.

Can you please explain to me why pulling down a bezier curve should increase the value at any point on that curve rather than decreasing it, relative to a straight line?

Here’s a bezier curve I grabbed from Wikipedia, and I added the red line to illustrate the difference between the starting point (red line) and what the result is after creating the curve:

The manual says the VCA value is added to the controlled tracks value. Red line value added to audio tracks value. Why would a lower value (vertical axis) yield a higher value added to the controlled track?

Maybe the straight line automation point (from unity gain down to -oo is misleading or indeed ā€˜not implemented correctly. On the automation track in Nuendo. i.e The two Straight lines as you have shown, cancel each others level out exactly between two points as an internal calculation (with no automation points along the path) all in relation to the levels in the mixer. If you add in an automation point this value will be added to the other point (VCA) at any given position resulting in the volume at position. Now I’m going a bit fuzzy!

Sorry, I’m just not seeing your point or reasoning.

will go through your steps again another day, it’s a bit strange……

Well ā€˜adding’ together the two values gives the reading here:

ignore the meter value of 4.0 (meter was not reset)

I know. Your example isn’t really what the thread is about.

I would prefer it if we could stick to the actual possible issues I described.

So yeah, can confirm that adding the same value points (-10) to the audio tracks at new position of fader automation this gives a different value as when there are no automation points (when the VCA is a straight ramp). If the VCA is a bezier curve then adding those points does not change the value of the straight -10 line. If I am not understanding this, it’s probably best if I exit here:-)

Edit: I get it now after reproducing and see the problem! However I will leave it with you as I cannot explain it any further than you have shown.

You are not understanding it unfortunately.

Oh well, I will exit here….maybe Steinberg can chime in on your observation. Have bookmarked to see if any developments or comment.

Don’t shoot me down here Mattias, For your first point.

As previously shown by you, when two straight lines as opposites go against each other within Nuendo they cancel out, whatever code is going on under the hood, those values are adding to give -oo all along…..I don’t think this should be happening. In my test at 5 seconds point (sinewave 0dBFS) the audio track with a straight line from 0 to -oo reads -30.7 (vca not enabled)

The VCA track reads -17.9 going from -oo to 0db @ 5 seonds. Combine those two values and you get -48.6. not -oo which is what happens when you enable VCA

Change to a bezier curve in the VCA and I believe in effect we are settings things to operate correctly, i.e In this specific context of the VCA track working with the Audio track we perhaps cannot compare the bezier curve being lower at a specific point to the straight line. i.e just because it appears lower it does not mean that the value is lower than when it was (or should have been) with the straight line.

Personally I think the way the Straight lines work against each other to give -oo is wrong.

That really is it from me, I am just trying to work it out in my head further.

All the best

Except the value of the fader IS lower. Just try it. You can even link an audio track to it and see the change.

Why should it make sense that the value be -oo regardless of where you are? (these are straight lines in the image) This value really should be -48.6 on the audio track at the 5 second point as the VCA is at -17.9 and the Audio Channel is -30.7