Can someone explain why this happens?

1. I setup a project with a number of tracks
2. My aim is to have the average around -12db on the completed levels.
3. No plugins are used on the tracks just setting levels.
4. Six Bus's used and all fader on Bus's at 0db
5. Pulled fown all fader on tracks and then went about setting levels
6 Main Output set to Mono and only one speaker.
7 Set levels and I have now a nice setup to start mixing.
Setup other DAW and did the same level adjustment but look at the difference in the fader positions and outputs on tracks on the mains the average is -12
Both used the same audio interface. Is i t sonething I'm doing wrong or is that just the way different daws work.
Do faders work better on the audio if pushed up closer to unity and reduce the gain on each track?
" />" />" />

Many thanks jeff for a very interesting reply. This is a short video showing what I did with a -20db file from Digital Domain. The large output is my Rme interface.
What do you think?
http://s1019.photobucket.com/user/fitzj/media/k-systemtest-20.mp4.html?sort=3&o=0
 
 
 

Hi John, that video is so low res I cannot really see anything such as the markings on the various meters etc. Perhaps a high res still shot might be better. Might be wise to alter the skin for the Klanghelm meters too. The cream coloured one is best for ease of reading.
 
Can you get a sinewave tone on there instead of noise as well. It is a bit more precise. I can upload tones at various levels on my Soundcloud if it will help.
 
Not sure if different pan laws are going to create such a difference in the fader positions as originally shown. But Danny may be right too in that they could. Stereo test tones on stereo tracks eliminate the errors.

I was going to mention that you may be mixing and matching the choices of meter ballistics, and now that I see the video it seems as if you are.
 
Make sure that when you compare the meter readings that all the meters are set to the same function and the same ballistics.
 
You have choices like
 
Peak,
RMS,
Peak/VU
VU
 
just for starters.
 
If you use a sine wave you will not need to worry as much about ballistics but you still need to think about frequency response and weighting.
 
It will be helpful to acknowledge that RMS and VU metering are two different things and so you can expect to see small differences between a accurate model of a VU meter and a simple RMS meter. The VU meter is weighted for frequency dependence and the RMS meter is not.
 
Finally, if you are using RMS meters inside your DAW you will need to know if your DAW conforms to the AES17 spec or if it is calibrated to some other spec. For example; You will see a 3dB difference if you compare SONAR's RMS meters to a DAW that uses the AES17 spec.
 
My guess is that your observations are of a combination of factors and that it why it seems like a simple answer will not describe the circumstance. 
 
best regards,
mike
 
 
 

I have tested both my hardware VU meters and all my VST VU meters and they are flat from under 10 Hz to well over 25 Khz so it can be done. I was surprised Mike when I did the freq response test on my hardware meters. I was not expecting them to be so linear over such a huge range.

Make sure you have got the BlueCat meter also set for the same reference level. It looks like it could be reading low but I could be wrong as the video is still hard to see on that one. I have got the BlueCat as well and it is a nice meter. I find I use the Klanghelm (as well as my real VU's) much more but the BlueCat can show the level over time which is useful.
 
Are the other DAW's behaving themselves too and showing the same things.
 
BTW I had to tweak the ballisitic settings on the Klanghelm meters to get them to behave the same ballistics wise as my real ones. The real ones are the ultimate for ballistics but out of all the VST VU's I own the Klanghelm can get the closest. I am not at my DAW right now but I will post up the Klanghelm meter settings to achieve close to the physical perfection. (the 300ms rise time is not quite right on the Klang meters as well, it needs to be something slightly different as well as the fall time) I have devised a special pulsed test tone that is a series of pulses 300 ms wide but longer in between. (to see the fall behaviour better) Music is also a good test for ballistics too.
The ballistics are very important because if they are set wrong they will either overshoot or not reach the 0db VU mark the same way the real ones do. (my real meters are worth over $200 each in case you are wondering and they are large and look great!) Incorrect ballistic settings on VST meters can give a wrong impression of levels with music. (not a continuous sine wave of course) It is not a major error but an error none the less and it can be avoided.

If you look at the ones that alter your level being centered, this can be an issue with loudness. For example, if you used -3dB center or -6dB center, that's exactly how your levels will be affected without panning a thing, understand? I'm not saying it's what your problem is, but it COULD be a contributing factor.
 
-Danny
 

Pan laws in the case of test signals can effect mono test signals but by using stereo test signals (which are effctively same signal in both channels except for pink noise) on stereo tracks they don't apply and that is why it is preferable to use stereo test signals on stereo tracks. As long as the pan pot is centered the correct level will be sent to L and R outputs respectively and pan laws will not effect the L and R extremes.
 
In some programs such as Adobe Auditon or Cool Edit you can alter the phase relationship between L and R sides when making test signals or you can have both sides in perfect phase too. Even so you wont get level variations unless you start summing outputs anywhere but in this scenario we are not.
 
The -20 db FS full range pink noise test signal from the Bob Katz website is also in stereo for that very reason.
 

Hi John, yes that is looking good. And it should be the same in all your DAW's too. That was your original concern but I think you will find all the DAW's will show the same level.
 
The other part of the process is to calibrate your room monitors. If you calibrate your system to say -20 then it is time to put the -20 pink noise (full range) onto the track now and measure the SPL level in the room. I like 85 dB SPL when both speakers are on. (83 dB single speaker only) The SPL meter should be set for C weighting as well. This will sound quite loud with pink noise but with music it is softer and nicer. Some around here will say 85 dB is too loud but it is not for me. Once the SPL meter is showing 85 dB you can mark your monitor volume control.
 
If you decide to work at -14 instead then obviously you recal the meters but the SPL level in the room does not change, it still should be 85 dB SPL. So now you put the -14 dB pink noise (full range) on the track and adjust the level in the room for 85 dB. Except because the system is working 6 dB hotter than before that mark on your monitor output will be in a different place. (lower of course) You will have two marks now, one for -20 and another for -14.
 
85 dB is considered a SPL level where the ear is hearing at the most flat it can be. You can certainly monitor lower but low and high frequencies will drop and it is not advisable to make important mix decisions at too low a level. You tend to push the ends of the spectrum if you do.
 
I have actually got a permanent SPL meter setup in front of me so it keeps me honest! I also check mixes down low (well under 85 dB eg 70 db or so) in mono on a small speaker similar to the Auratone. I find it reveals a lot about your mix. I also like monitoring up very loud too eg 95 dB SPL and 105 dB SPL! But not for too long. That tends to reveal a lot about the bass end and the reverb levels.
 
You may have some issues with bass nulls and peaks in your room where you are sitting. If you find while the full range pink noise is playing and you move the SPL meter even slightly the reading may change wildly. There is another pink noise signal you can use. It is called band limited pink noise and it is pink noise only from 500 Hz to 2 KHz. BUT the level will drop way down due to the low end mainly being removed so the bandwidth limited pink noise needs to be level adjusted again to bring it up to the same level as the full range pink noise. The band limited noise avoids any room low end issues and gives a very accurate reading. I can upload it if you need it only. Mine is level corrected already so you can just plonk it on a track and away you go.
 
If you use the VU's well you can aim to get all your tracks just hitting the OdB Vu mark on your VU meter. Very transient sounds such as drums etc will not make the Vu's reach that level. This is where your peak metering comes in though. But by the time you send all your drums to say a drum buss the VU will start to show normal levels there too.
 
Your buses should also be just reaching 0dB VU as well and you will find your master buss will also just reach that too. When you do this there will never be any clipping or red lights anywhere. You will always have plenty of headroom above. And when you mix say 10 songs they will all be the same level prior to mastering. Nice! If you check the VU levels going in and out of any plugins or effects chains you will never overload them either.

Ok I have uploaded the band limited pink noise and it is level adjusted. Now just note that it is only at -14 db FS. If you want to make a -20 version just subtarct 6dB of gain. It does basically the same thing as the full range pink noise except that because it is band limited any low end issues in your room environment wont influence it in any way.
 
https://soundcloud.com/jeff-evans
 
I am only going to leave those test tones up there for a little while so anyone else interested should download all of them and put them into a folder. They come in handy for various things from time to time.  I find another good thing to do as well is to take a commercial track eg Steely Dan and re adjust it so it just peaks the VU's at your chosen ref level. It is handy sometimes to put music to a very precise ref level onto a track just to see how everything is shaping up in your system. Interesting in that the Steely Dan's CD 'Everything Must Go' was mastered only up to -12 dB rms which is pretty cool. No wonder it sounds so damn nice. I only had to lower it by 2 dB to make it a -14 dB track instead. And then you would take it down another 6 dB to make it a -20 dB track instead.
 
I still would not mind knowing what ref level produces +4 dBu from your audio interface. That is an interesting test. If it turns out to be -18 then you might as well stick to -20 as a chosen K system reference. -18 is very common amongst a lot of pro digital gear too. Interesting that where I teach sound engineeering they bought an SSL AWS948 console and it came set for -20. But it is easily changed in that console. My digital mixer has chosen -14 and so has Harrison Mixbus for some reason too. I can change my mixer but Harrison Mixbus is fixed at -14 for now. I would like to see that variable perhaps in a future update.
 
It has been a pleasure to help you John and others. I am a bit of a level and calibration nut when it comes to all this. I find once it is done it makes such a difference. It means everything you turn out is to a precise level. I really care about helping others make better recordings. I get a lot of satisfaction from it.

So much to take in Jeff. You really know your stuff. Thanks for the education. Hopefully I can put it into practice as well as anyone else who has been reading this thread.

Jeff
I still would not mind knowing what ref level produces +4 dBu from your audio interface.

 

fitzj
Thats great Jeff
I have two RME devices. The Fireface 800 and the Babyface.
The output for the Babyface says 
Maximum Output Level @ 0 dBFS Line: +15 dBu
Maximum Output Level @ 0 dBFS Phones: +8 dBu
 
The Output for the Fireface 800 says 
Input/Output level for 0 dBFS @ Hi Gain: +19 dBu
Input/Output level for 0 dBFS @ +4 dBu: +13 dBu
Input/Output level for 0 dBFS @ -10 dBV: +2 dBV
I will do a check  on this at the weekend.

 
The info listed by RME hints at the fact that it is in the ball park. If you want to know the specific details it is easy to take a measurement.
 
If you want to measure your hardware's dBu output to find out what reference level produces +4dBu send a 1kHz sine wave out and measure it with a true RMS Volt-Ohm-Meter set to AC voltage. A dBu level is measured "unterminated" so the super hi impedance of a VOM is appropriate for a direct measurement of the voltage. Play back a 1kHz sine wave through the RME and adjust the level on the RME until the VOM reads 1.227652988 volts (or close to that) and then look at what the meter says on the RME.
 
The reading on the RME meter will indicate what reference level produces +4dBu.
 
Then if you have real analog VU meters you can see if they are calibrated so that the 1.227_vAC (a.k.a. +4dBu) reads as 0VU on their scale.
 
Most top of the range pro gear will be designed so that -20dBFS on the digital meters will equal +4dBu at the outputs.
 
There is a lot of high quality prosumer gear set up where -12dBFS equals +4dBu at the outputs. This usually correlates to the fact that the analog outputs on that gear do not have as much head room as might be encountered on top of the range gear where +24 thru +26dBu output before clipping is common. As transfers to analog tape become less common there is less practical need to have such high output levels and so a lot of high quality gear has lower Max output levels. The extra output capability was primarily used to saturate tape recordings and get a fat sound during the transfer and to maximize the SNR.
 
best regards,
mike