What does that db level indicator really mean?

THIS LINK is the one I believe you mean - Headroom? from Sonar

Are all the meter levels you're talking about from the meters in your recording software? If so, then they all reference digital, not analog, whether they're from a MIDI-driven synth (vsampler) or an actual recorded track.

I'm going to talk about two things here. First, dB in general. And second, what meter readouts mean vs the percieved loudness of sound. The latter is more relevant to your problem, I think, so you may want to skip all this dB stuff for now.

First, dB in general:

Like wogg said, dB scales are centered around an arbitrary reference. For digital audio, 0dB is that reference, and it's the absolute loudest point that can be recorded without clipping. This is why digital dB's are always negative.

For analog tape, the 0dB mark is some arbitrary magnetic strength on the tape. Tapes usually don't start distorting much until around 6dB, which is probably where you heard the "6dB analog is 0dB digital" thing from -- but this statement by itself is *not* true.

There are other dB scales, as well, dBu and dBv and others that I don't remember, that deal with electrical signals on wires and voltage levels. Finally, there's the dB scale for human hearing at which 0dB is set at as the "quietest percievable sound" -- which is really just another arbitrary sound pressure level.

OK. So each kind of dB has a different reference point at 0dB. What do levels mean beyond that? Well, dB is defined as 10 * log "power", where "power" is proportional to the power of your signal. Now, it's usually easier to talk about amplitude than power ... and amplitude is voltage, or sound pressure level, or digial amplitude. Since power is proportional to amplitude squared in the physical world, then a dB signal is 10 * log ("amplitude" squared). Now if you remember your math, you can take the "squared" out of the log and put it on the front, so dB is 20 * log "amplitude".

"amplitude" is proportional to your signal's actual ampltude, scaled so that 0dB falls at the arbitrary reference point discussed above.

The effect of all this math is that every time something goes up 20 dbB its amplitude is *ten times greater*. Every time it goes down 20 dB, its amplitude is *ten times less*.

So something that's 40dB up has an amplitude of 100 times more. 60 dB means 1000 times. Etc.

Other good numbers to know: 6dB means doubling. 3dB means multiplying by the square root of two.

The reason we use db is that percieved hearing is more logarithmic than linear. If something is at 20dB, and it goes up to 60dB, it doesn't *sound* like it's 100 times louder. It sounds closer to being 3 times louder (though not quite. )

----------------------------------------------------------------------------------

OK. Enough of that. What you're really wondering about is why your meters don't seem to jive with how loud the sounds seem to be. The reason I went off on dB is because it seems like you're confused about dB as well, and I thought I'd fill you in.

Anyway, back to your meters. If you notice, meters generally have several different modes, such as "Peak," "Average," "RMS," etc. Which mode were yours in?

The reason I ask is because, while "peak" will show you the absolute highs of the signal, the "RMS" mode is much more accurate as to the *percieved* loudness of the sound. Think of a sine wave vs a square wave of the same amplitude. Sure, they both *reach* the same highest point, so both have the same "peak" amplitude, but the square wave has more energy higher up than the sine wave. the result is that the square wave sounds louder. Peak doesn't show you this, but RMS does.

My guess is that your midi track has high transients and high peaks, but they're "thin" peaks, and it doesn't have the "body" that really makes it loud. Guitar, on the other hand, especially when overdriven, has a lot of body and a lot of higher energy in the waveform, and therefore sounds louder. Try switching your meters to RMS and see if that matches what you hear more closely.

If it doesn't (or if you're already using RMS), the reason might be frequency masking. If the guitar part has a much more complex frequency spectrum than the MIDI part, then the MIDI part may natrually have to be louder to "fight for space" in the mix over the guitar. The guitar can be set lower and still be heard because it can fill in the frequency spaces that the MIDI track doesn't cover.

Either way, the lesson is to always trust your ears over your meters!! :)

Ken

There are two reasons why many people get confused about setting levels in a digital recording environment. One is you need to understand firstly what the rms and peak parts of the signal are. The second is adhering to a standard in terms of digital record levels. There are three levels that are well known and are part of the K System metering concept.

This thread is probably much closer to what you want to know but there is quite a lot of reading there but once you get into the nitty gritty of K system signal level flow within a digital system it all makes sense.

http://forum.cakewalk.com...;m=1935407&mpage=1

Calibrated monitor gain is just another aspect of K System concepts. But reserved more for mixing and mastering stages. K System is of course covered much more deeply in Bob Katz's book 'Mastering Audio' K System requires you to get another inexpensive plugin that displays K System metering options such as the Blue Cat Pro here:

http://www.bluecataudio.c...Product_DPeakMeterPro/

Sonar cannot do this type of metering by itself and that is why many people either dont know about it or don't do it. Once you have selected a suitable reference level to work at, and that applies right from setting tracking levels on individual tracks through to mixing buss levels and mastering as well, then you need a metering device that is going to show you an rms level around 0db while you are working at your chosen digital reference level. This plugin can be inserted on your main master buss.

The amount of headroom for peaks is determined by the K level you are working at. 20 db of headroom for a K-20 level, 14 db of headroom for a K-14 level and 12 db of headroom for K -12 level.

Also check out this thread:

http://forum.cakewalk.com...79&mpage=1#1929708

And check out my Post #8 for a really detailed look at setting levels within digital systems.

for the OP:

"The Midi tracks bounce around 0db - 6db on the level indicator before I apply any compression, eq or trim leveling. Fine."

"The live guitar tracks...    When I record the track it bounces around -12db - -6db on the level indicator. One would naively assume the recorded audio is on average 12 db lower than the midi tracks. (which is what 4 times softer?). Is this true?"

Not necessarily true... if the circumstance is simply testing the same playback content at different levels then it is true: 12/3=4.

But when you compare your drum and bass tracks to the guitar tracks you are comparing two different sources and the idea of "loudness" becomes important.

Go read about loudness curves and you will see that often times guitars can "peak" on the meters at a lower level than the drum kit yet the guitar may sound louder. That is because the mid-high pitched frequencies don't need as much energy for us too notice it. That effect is described as "loudness" and scientists have described the effect under many circumstances which are recognized as different "weightings".

This info may help you reconcile the abstract and rather useless info you will learn from a level meter with what your brain is telling you your ears are hearing.

Soon you will hear the character of the music first and simply use the meter to confirm you are not going crazy.

Don't clip... make stuff sound good. Strive for consistency when setting your monitoring levels.

Make sure you hit record when something is going down. :-)

all the best,
mike

" What does that db level indicator really mean?" For me? Usually trouble...