Sure - it's adding to the learning process.szalonykp wrote:Hey, Compyfox - could you help me interpret the statistics?
I use ToneBoosters EBU Loudness for measuring True Peak (and average signal strength according to the EBU R-128/ITU-R BS.1770-x specs of course). Voxengo should by now have good routines to show correct dBTP values by now, but things can still be off due to small intersample peaks that not every meter might register.szalonykp wrote:As I understand they mean that:
1) I was overshooting the True Peak Volume (I was aiming for -3dB with Elephant (oversampled if I remember right - that should catch the inter-sample peaks...). Do you know any free or cheap plug-in that would help me with metering?
I opted for dBTP in the statistics to show what's really going on instead of going for regular dBFS (digital peak).
Not necessarily, which is why I added more values to the statistics to analyze (in an objective way - listening tests are still needed of course), what might go wrong with your track.szalonykp wrote:2) my average RMS and max RMS are the highest of all mixes - does that mean than my mix have the lowest dynamics?
In this case:
Dynamic Range (DR, declared mathematically, not with a tool), RMS avg/max, SLk Max and LRA.
Let's look at the values:
What is still most used to analyze music these days is RMS - but it only tells you so much.
Your track has an average signal strength (RMS) of -12,26dB (RMS), the max average signal strength is -9,43dB (RMS).
Now, if you wouldn't have told me, that you used Elephant as limiter (in either form), then I would have assumed that you tried to do some pre-mastering or utilized heavy bus compression. Or, that you used a higher worklevel than the others.
The SLk Max value with -10,9dB somewhat aims at that direction, telling me the same two things:
a) this track has been mastered already and is release ready
b) maybe the user has used a higher reference worklevel
So let's check the other values if they confirm that.
The DR (Dynamic Range: Range from RMS avg to max digital peak) shows me a value of about 9,2 - that already aims at hot and dense mixes, but is (IMO) still somewhat acceptable for busy electronic tracks (keep in mind: modern overcompressed commercial mixes would show 5 or even 4 as DR value!). The LRA (Loudness Range, basically from lowest average signal strength to highest average signal strengh) shows me how dense the mix is. Since this is still hard to define what the "ultimate" value is for such a mix (the same flaw with the DR meter), it's more like a "guesstimation". But 3LU shows me that nothing much fluctuates. So the track has to be fairly dense mixed, or that there is overall a lack of dynamic movement. Another indicator for heavy bus compression/limiting.
Now let's take a look at the dBTP max value:
It shows -2,98dBTP. Well, if you shot for -3,00dBTP with Elephant, that could for example be an indicator that the limiter is not really made for limiting content in the EBU R-128/ITU-R BS.1770-x realm. But more important is the roughly -3dB compared to your average signal strength.
It... looks healthy to me if I compare this value with the DR value (avg signal strength to maximum digital peak). But it also indicates that you're cutting off too much transients just to reach a specific value.
So there you have it.
I assume you tried to shoot for -3dB since I wrote in the Mix Challenge Guidelines that a digital peak shall ideally not exceed -3dBFS if you utilized a proper reference level. And you usually don't exceed that if you used -18dBFS = 0VU as reference. Unless you created very strong transients with transient designers and such.
But if you aimed at pre-mastering with this mix, then your better route would have been -1dBTP to retain the transients.
In the end, only you can tell us if you merely used a higher reference level, or did some premastering already. Going by the plain numbers, we can only make assumptions.
Which route is the right one?
Sadly, there is none - everything has to work together. Of course you can aim for mixes with the highest possible DR and LRA value, or the lowest possible SLk value. But does that mean that the track is better mixed in the end? Not necessarily. Mixing is part technical, and a huge part "gut decision". You as mixer need to find the proper balance. The numeric values are then only "bonus" if you might have messed up somewhere, or if your track adheres to certain release standards.
Yes and no - in very basic form, it shows you how dense a track was mixed (lowest avg to max avg signal strength measured over the course of the whole song). Or the other way around, how "jumpy" your production is (interesting for broadcast mixes from voice over with no ambiance to a wall of explosions).szalonykp wrote:3) LRA tells the same thing, isn't it (that the mix lacks dynamics)?
Absolutely not. DC offset can be removed if the track is going into the mastering stage.szalonykp wrote:4) should I worry about DC-Offset (it's -100.0 dB instead of minus infinity)?
It's more like an indication how much you "abused" analog emulations, and whether or not you might have utilized proper low cut per track. Which can cause problems with some metering tools.
I hope this was a helpful answer.szalonykp wrote:If you could help me with that - it would be greatThanks!
.
