Digital: What meters? What reference level?

[ofajen]

OK, I'm curious, how do you meter when tracking? What is your reference level and what do you use to monitor levels when mixing? I'm coming at this from 23 years of recording and mixing entirely in the analog domain. The thing I noticed first in the digital world is the lack of useful meters. The little meter displays in Live 5.2 are some sort of peak meter and there are no calibrations.

I was thinking of splurging on a Dorrough meter, but I happened to have an extra pair of Modutec VU meters laying around (doesn't everyone?) and so I built up a VU meter box and I'm running one pair of outputs (the balanced mix recorder outs) from my Delta 66/Omni to the meters and one pair of outputs (the monitor outs) to the monitor amp by way of a stereo pot (that reminds me... time to order a Shallco pot).

Anyway, I use Ableton Live's test tone to set the VU meters to read 0 VU with a -20 dB test tone. It takes a little gain to actually get the Delta recorder output to drive the VU meters at that level (about 6dB) which suggests maybe they think -14 dB is an OK reference level, but that seems to be too high for mixing anything but stuff that came off of tape.

So, is -20 dB typical? I'm tracking and mixing in 24 bit audio, so I'm wondering if -24 dB might be OK.

Any comments or suggestions?

Thanks!

Otto

 

[mshilarious]

So, is -20 dB typical? I'm tracking and mixing in 24 bit audio, so I'm wondering if -24 dB might be OK.

Anything from -18 to -24 is good. I tend to use the LED meter on my preamp, it has a digital out, half the LEDs lit is -12 so I shoot for that as a peak. That should yield -20ish on a VU meter. I also have an analog VU meter in the box, but I don't use it too much.

When mixing I like the K meter on the UAD Precision Limiter. I turn it on even when I'm not limiting anything just because I like the meter so much

 

[ofajen]

When mixing I like the K meter on the UAD Precision Limiter. I turn it on even when I'm not limiting anything just because I like the meter so much

I'm not familiar with the K meter or the UAD. Is that a hardware limiter or software? What kind of response? and how do you reference to your digital mix output?

I've been doing some level setting on old transferred tracks from the big, bad 3M M-56 and the VU meters I set up last night are really a huge help. I can almost always get to a good starting place on levels by setting the strong levels of each track at -N dB on the VU, where N is the number of tracks that are usually playing at one time.

Thanks!

Otto

 

[bennychico11]

it sounds like you got a pretty good handle on things already.
I'm like mshilarious, I use a software plugin (a channel strip) solely because it has peak and RMS meters on it (RMS meters are very similar to VU meters...although not exactly). And it's a TDM plug...which on my old hardware is the only thing it supports on Stereo busses.
I just set it to bypass to keep my audio away from the DSP of the thing (I swear I hear a coloration of the plugin even though nothing is EQing/Compressing) but the meters still register. It's done me well on most things so far. Keeping my levels around -20dBFS RMS have my tape masters averaging around 0VU...which is what I like.

You could do something similar and look to see if you have a plugin already which has both VU/RMS and peak meters. No need to calibrate digital meters...they're already set how they should...you just need to learn to adjust the numbers in your head. Just keep an eye on peak meters so you know you're not creating square waves in your audio.
My personal reference level is -20dbFS=0VU with running a sine wave through all the gear. I tend to keep my peaks around -10/-9dBFS maximum. I do audio for video though, and that's pretty much the industry standard in the states (-18dbFS=0VU in Europe I believe).
For music I'd probably mix a little hotter, though. Spoken voice is one thing...but music is another. It also all depends on what you are recording. A symphonic orchestra is going to have a whole different set of leveling procedure than your average pop band.

as far as the K-System metering...and for a ton more of great info on leveling in general:
http://digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page_id=36/
http://digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page_id=59/
(specifically the second link)

 

[ofajen]

as far as the K-System metering...and for a ton more of great info on leveling in general:
http://digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page_id=36/
http://digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page_id=59/
(specifically the second link)

Doh! I'm getting old! Of course, K as in Bob Katz. I didn't trigger the memory with K-meter. I bought his book on digital mastering when it came out, and highly recommend it.

I read the link on level practices and checked out the Metric Halo stuff (Spectrafoo). Looks pretty cool and probably a good use of $400 sometime down the road. Channel Strip looks really useful, too. First, I need to get an SPL meter and a good stepped attenuator to calibrate monitoring levels. That's long overdue.

His comment about VU meters devoting half the scale to the top 6 dB is useful, as is the idea of worrying more about message than metering. Taking things one step further, he suggests using 0 VU (on an appropriately selected RMS/ peak meter) as a fixed reference at a monitoring level of 83 dB SPL, regardless of whether you're using a 20 dB scale or 14 dB scale above O VU. If 83 dB is the reference level, we should expect occasional forte passages with average levels rising to about 89 dB or + 6 dB (probably more often with thicker music like metal) and we should monitor them at that level.

So, the notion occurs to me that VU meters would actually be more helpful to me with digital tracks if I treat -3 VU as the reference level (monitored at 83 dB SPL), so that +3 VU is the forte volume of 89 dB SPL. That lets the VU meter show me more information about louder average levels that will sometimes be there with my less-restrained digital tracks. Using 0 VU as reference level on a VU meter wouldl mean that I would peg the meters a lot more and not quite know where the loud sections are. I'm going to give it a try...

Thanks!

Otto

 

[mshilarious]

His comment about VU meters devoting half the scale to the top 6 dB is useful, as is the idea of worrying more about message than metering. Taking things one step further, he suggests using 0 VU (on an appropriately selected RMS/ peak meter) as a fixed reference at a monitoring level of 83 dB SPL, regardless of whether you're using a 20 dB scale or 14 dB scale above O VU. If 83 dB is the reference level, we should expect occasional forte passages with average levels rising to about 89 dB or + 6 dB (probably more often with thicker music like metal) and we should monitor them at that level.

Then I think I would look for a software K meter. UAD is not the only one out there. It does exactly that, set 0 at K-14, K-20, or K-12, and when the average goes too high above that it's red.

For tracking though, again the only concern from my point of view is plenty of headroom for peaks. Final average level is not necessarily a relevant concept when you are tracking a single instrument or voice.

 

[bennychico11]

If 83 dB is the reference level, we should expect occasional forte passages with average levels rising to about 89 dB or + 6 dB (probably more often with thicker music like metal) and we should monitor them at that level.

where are you seeing this?
I'm reading that he's saying 0VU=83dBSPL= -20dBFS RMS....per channel (86dBSPL in stereo)
He says this 83dBSPL should be the forte passages. So 0VU should be your loud passages...I think only on fortissimo passages you might get near +3dBVU (which will give you your 89dBSPL with BOTH speakers). Maybe that's what he was talking about?

Also, just make sure you don't disregard your peak meters. It's good to watch both (and you have to watch peak meters when working with digital)

Final average level is not necessarily a relevant concept when you are tracking a single instrument or voice.

although something you should keep in mind in the back of your head...remembering how many instruments you are needing to sum together to get the final average. Summing 24 channels of a rock band together will be a lot different than summing 2 channels of a solo piano.

 

[ofajen]

where are you seeing this?
I'm reading that he's saying 0VU=83dBSPL= -20dBFS RMS....per channel (86dBSPL in stereo)
He says this 83dBSPL should be the forte passages. So 0VU should be your loud passages...I think only on fortissimo passages you might get near +3dBVU (which will give you your 89dBSPL with BOTH speakers). Maybe that's what he was talking about?

You are correct, I think! I got confused with the mono/stereo thing... I thought everything was referenced to the 83 dB SPL of a single speaker, but of course, when talking about an actual performance SPL (orchestra fortissimo SPL), we would compare the total SPL produced by two speakers.

Also, I should have used the term "fortissimo". O dB (83 dB SPL mono, 86 dB SPL stereo) is forte, +3-4 dB (89-90 dB SPL in stereo) is fortissimo. Sorry, I was up late last night.

I still think it might make sense to use the VU meter with the reference level at -3 VU when dealing with digital tracks, so that I have a little more useful metering range above the reference level.

Is there a good software plug in meter that does K-14 and K-20 that is a lot cheaper than Spectrafoo?

Thanks!

Otto

 

[mshilarious]

although something you should keep in mind in the back of your head...remembering how many instruments you are needing to sum together to get the final average. Summing 24 channels of a rock band together will be a lot different than summing 2 channels of a solo piano.

It shouldn't matter that much. Digital attenuation will yield the same result as tracking at a lower level, again provided there is adequate headroom at all stages. On the other side, there still is no reason to cut a solo piano hot, since there is more dynamic range available than needed.

It is true that a classical musician will not tend to be 6dB over soundcheck

 

[BigRay]

I use Dorrough Meters and Samplitude/Sequoias metering visualizations..

 

[Reggie]

I think you can pretty much do whatever you want. Especially if you are mixing for CD releases or whatever. At work I use -14 as my reference level, because that is what we settled on with our bin loop duplication system (cassette duplication). I also use Dorrough 40-A meters. Kind of an expensive extravagance if you ask me. I've gotten to where I can get a good idea of what is going on by just watching the peak meters and the recorded waveforms.
You are going to have plenty of headroom in digital if you use anything between -12 and around -20. And if you are not planning on sending the stuff out to be mastered, you will want to bump up your final mix levels by quite a bit in the end anyway.

 

[ofajen]

Source for stereo attenuator

Can anyone point me to a good source for a stereo stepped attenuator with 1 dB steps to calibrate monitoring gain?


Thanks!

Otto

 

[ofajen]

Can anyone point me to a good source for a stereo stepped attenuator with 1 dB steps to calibrate monitoring gain?

Uhh... never mind. I finally got in touch with the engineer at Shallco and they're going to make me the pot I need. I ordered a stereo pot with 20 1 dB steps plus an "off" setting with total attenuation.

Thanks!

Otto

 

[ofajen]

By way of update:

Selling some big tape decks has helped fund a few useful items and given me more room in the studio!

The Shallco stepped pot should be here in a few weeks. I got an Extech SPL meter, so I can already calibrate levels, I just have to play that pink noise file every time I change.

Along with the peak meters in Live (and Peak), the VU meters are actually quite handy. I had to run through something to get about 3 dB of gain to drive the VU meters at the right level and happened to use an old DOD eq. So, I now eq the VU input signal with a curve that is the inverse of the Fletcher-Munson equal loudness curve at 80 phon. The idea is to take Bob Katz' point about VU meters responding linearly while our ears have a sensitivity curve and try to reverse it, based upon the curve that corresponds to the audio reference level (83 dB). The 80 phon curve happens to be about the flattest one. The lower ones have a HUGE bass rise. Anyway, the eq tries to compensate for our ears' typical lower sensitivity in the bass and high end and increased sensitivity from 1k to 6k. I use 1 kHz as my reference tone and the VU eq is flat at 1 kHz. So far, it seems to make the VU style meter much more useful than the "true" VU spec meter.

I checked out the demo version of Bias Peak and agree that it looks like the way to go on the Mac side, so I've ordered a copy. I did suggest to them that they consider putting in K-system metering options in the next version. If anyone else wants to make the same suggestion, that might help.

Thanks for the help, things are working great, other than the slowness of my 800 MHz G4! Live 5 is absolutely insanely cool ...

Otto