OK dumb question…I just put a SuperVision instance set to VU on my master. Before that, I have a UAD Ampex tape plug in. So I see VU levels on the Ampex, and Supervision. The Ampex is reading normal levels (actually they were a bit hot and I trimmed the input to the AMPEX back a bit). Problem is as you can see in this screenshot, the VU meters in SuperVision are barely moving. Why is there a discrepancy between the two sets of VUs?
Different calibration of the VU meters. In supervision you can configure different scales for the display, and the default is “VU db”. If you set that to “VU dbFS” (with an “offset” of “-18”), it should be similar to the meter on the plugin, assuming that it is calibrated to 0VU=-18dbFS, too.
Thanks figured it might something like that, but I cannot find out where to change it’s configuration…?
Is the signal going out of the track with the Ampex unchanged? I mean, you didn’t accidentally change the level between the output of the plugin and the input to supervision? And you are measuring the output in Ampex, right?
The cog wheel at the top of the Supervision window configures the selected module.
Supervision’s VU meter actually does behave differently in its default configuration than other VU meters I use (e.g. VUMT). This is a -18dBFS 1KHz test tone through VUMT and Supervision:
Switch the scale to “VU dBFS”, et voilà, same behavior as VUMT:
Interesting. I think the labeling is likely not very good here. It looks like the meter when set to default basically just makes -20VU the ‘reference’, and it then shows absolute level measured as VU. So it’s a 1-to-1 situation: an input of -6dBFS average will read -6VU on the meter.
When switching to VU dBFS the reference level is what you set using “offset” and it ‘shifts’ the entire scale accordingly, so that if you set it to -18 for example that same input signal of -6dBFS no longer “fits” on the scale. Since our “-18” = 0VU a signal of -6dBFS is far higher than the additional 3dB we have available above 0VU on the meter (because -6 is 12 hotter than -18).
So I suppose it’s a somewhat useful default if you want the ballistics of a VU meter but also want to see more of the signal. By this I mean that if you’re mixing a TV show for example and you need to hit -24LUFS (which is a sort of (average’) then using a ‘normal’ -18dBFS = 0VU calibration your meter headroom is only 9dB. Because you have 3dB above 0, and then you have (-24)-(-18) below that . So if your average is -24LUFS then all peaks at or above -15 will peg the meter far right. And if you have a loud passage where the average is about -18 for example then any peak 3dB louder than that will peg the meter. Not very useful. But if you’re using default the peaks all fit up to 0VU because that’s 0dBFS practically speaking. On the other hand your average signal will be below what the meter is showing since the lowest part is -20VU, not -24VU. Mixing for streaming with an average of -16LUFS for example might make sense though. You’d see that average and all peaks within the meter range.
But anyway… it makes sense but isn’t really labeled or explained that well.
Anyway seems that Supervision VU meter is not correct at all. Sometimes the result is so different. I am tested it with TB pro audio and Klanghelm (the results are the same except Supervision VU). Klanghelm works perfectly! Any chances to see that Steinberg will correct VU meter? Attached the screens
And what were your Supervision VU settings?
-18 Dbfs. Iam also tested it via tone generator
And how is Klangheim set in terms of reference value (i.e. 0VU = what?)…?
Klanghelm and Pro audio show me the same result… and Supervision always show me the different picture 
I tested it on the kick drums, guitars etc. My attached screenshots show it…
I understand, but what is the reference level of the Klangheim?
-18 dbfs. So the same settings as in the Supervision
Interesting… Is it the same if you do a mono channel?
Same settings here and Supervision and Klanghelm VUMT show a similar response.
Note that peak hold, peak feedback and the ‘maximum audio performance / sample accurate display’ switch (on the settings toolbar) all have an influence on the response of the Supervision display. Perhaps try switching it to sample accurate display and you may see a difference.
[EDIT] But there may possibly be an issue with the rise time of Supervision’s VU meter.
Honestly, don’t understand. What I need to change in the Supervision and where? In the settings?
I’ve only tried this with iZotope’s “Insight 2” set to VU for metering, not an actual VU meter plugin. But when I did they showed very similar results (hard to tell judging between needle and bars).
One thought - if one of the plugins induce latency it’s maybe possible that the second in the chain is reading an ‘older’ value while the first in the chain is reading a more ‘current’ one… (?) Not sure if this is likely, but one quick way to test for that is to have one source feed two groups in parallel and have each plugin on its own group instead of in series in the same path.
I’d also run both program material and different generated waveforms to test.
