Hey !
I read recently that mastering engineers prefer to be left anything from -6 db to - 12 db headroom when they're sent a mix. 
Ive been working on a piece today and was very careful to keep the levels down as low as possible.Everything worked great,then I got to the end of the piece where the climax was and yeah,the levels just went sky high.So my questions are

1 Should I of started with the climax first ?? The style of music is orchestral. I had an earth shaking bass drum,screeching strings and a horn blasting out like no tommorow. If I should of started with the climax,would it have been necessary to start on the loudest elements first ?? Like the bass drum ?? And then fit everything on top ??

2 How many people on here use subgroups to mix ?? I think I would of saved myself a lot headache if I had grouped the instruments together and then adjusted the levels for entire sections,like strings,drums etc. The only problem I have with this is a lot of my tracks use automation,so some instruments will follow their own settings on their individual track  and could possibly stick out like a sore thumb when placed in a subgroup. Maybe bouncing the tracks down is the way around that,before placing them in a subgroup ?

3 This my dumb question :) If the levels on the master fader get too high,why not just the trim on the master to bring it all down ??? Whats wrong with that ?? And why cant a mastering engineering do this to a mix also ?? If I sent this piece to a mastering engineer and there wasnt enough headroom to work around,couldnt he / she just reduce the gain of the track by a couple of Dbs to solve the problem ??

Many thanks in advance for your thoughts,ideas,

Kev

purists that record orchestra would suggest leaving maximum headroom for dynamics, which means starting over.


you let the entire piece play thru, find out where the peaking tracks are, set those fader levels at no more than -8db at PEAK, and mix backwards from there.

bring everything down to match the levels required to allow those problem tracks to PEAK without compression or limiting.



if it's a modern mix, then it's a whole 'nuther ball of wax.

you can add limiting.... compression..... track automation...... you can do destructive edits like re-gaining problem areas that just got to be too much...



i always use sub groups to mix.

that's what they are there for.


i'll even automate sub groups.

leave the master alone.

you want it sitting on 0db.

you want to change the individual track faders, so at their collective PEAK, they are no higher than -12db to -14db  on the master fader.

Ok here is my take on it for what it is worth.  Firstly people need to understand gain structure, what you need to think about is when you have say a 20 track mix is think how loud the mix will be once it hits the main out.  This is called summing.  So lets say you have all your indivual tracks all nicely mixed and they are averging (sonar meters) roughly -18DB, you will notice that the main buss is still going into the red.  This is because when you add all your nicely mixed tracks together, their sum is greater than 0DB peak.  What you need to do is give your self headroom, I do this by lowering the trim by anywhere up to -15DB, depending on how many tracks I have summed together, within a mix.  This will give you the required head room to master, makes sense??

Ben

i'm definitely no mastering expert, but my understanding is this:

in sonar, you have either 32 or 64 bits of floating-point resolution to work with. you can raise and lower volume all day long and it's no big deal, because when you mix down to 24 or 16 bit (integer, or discrete), it will amount to the same thing. 

for example (and this is super-simplified), let's say you did 5 things to a wave form in sonar:

+55.3, -66.83838, +89.3939, -22.33, +11.5

and you end up with 87.383. 

now lets say you did 5 other things and ended up with 87.384.

when you mix down to 24-bit, or 16-bit, the resolution is lower, and the .001 difference doesn't matter. you have 87 either way.

now, when you give your mastering buddy your 87, if he has to lower it, he's losing real data that he can't get back, and then he's going to "turn it up" again to normalize volume among all your tracks, ensure a good listening experience, etc. so that loss of quality (6dB worth, if you peaked out) is now getting turned up back to where you originally had it. and if you already had it there, why are you getting it mastered?

so the idea is to give the mastering engineer something he can work with, without having to "turn down" first. in your case, i would say do the following:

turn down the master volume in sonar until your loud part is -6dB. then, using your volume knob on your speakers/amp/stereo/whatever:

1. listen at low levels. see if you still like what you hear.
2. listen at normal levels. see if you still like what you hear.
3. listen at high (not ear-fatiguing, but loud) levels. see if you still like what you hear.

if the quiet parts are too quiet when you listen to it low, then readjust your dynamics to bring up the quiet parts relative to the loud parts, and start over. 

once you like it, send it off to the mastering dude.

Kev, you may be overthinking it. The mastering engineer simply wants to make sure he's got some headroom to work with. This is especially important for pop/rock tunes that lack dynamics. Not so much for classical, where the ME is not going to do much compression or EQ. If your dynamics are pretty good, he might have no problem working with only 3db of headroom.


1) Yes, start with the loudest passage, and make sure no peak exceeds about -6db and you'll be fine. This will be plenty of headroom for the ME to raise the quiet parts if necessary without having the limiter engage. If only one or two peaks exceed your target, handle them in the mix with volume automation.


2) Subgroups are common practice and make mixing much easier. In classical music, ensembles are typically treated as one instrument. The conductor doesn't point to one violinist and say "you! get quieter". Rather, he points to the whole violin section and says "everybody! get quieter". The conductor doesn't have enough hands to issue instructions to every member of the orchestra individually.


3) Yes, turning down the Trim/Gain slider on the master bus is indeed a perfectly acceptable way of bringing levels down. If you watch track levels from the get-go, you won't have to do that, though. 


The standard argument against using the master trim this way is that some plugins don't like to be overdriven, so it's better to keep levels under control all along the signal path. Not many plugins have that problem, though, and you're probably not using them for classical music anyway.

Hi !


Start over again  doh !!! You're prob right.


So - 8 db for a peak. Thats a good reference point to try out,thank you ! It is a modern mix in the sense theres drums and bass guitar in it. I did subgroup both of these and apply compression and limiting,but the strings etc I left with no compression or very little.  I used big cinematic hits and I found myself using compression on these,and these appear to be the problem areas in the track,where they are leaving big stonking peaks behind them :) But ok - 8 db,I'll try that and hopefully this will give me some more headroom to work in.


Thanks for your insight,

Kev

batsbrew


purists that record orchestra would suggest leaving maximum headroom for dynamics, which means starting over.


you let the entire piece play thru, find out where the peaking tracks are, set those fader levels at no more than -8db at PEAK, and mix backwards from there.

bring everything down to match the levels required to allow those problem tracks to PEAK without compression or limiting.



if it's a modern mix, then it's a whole 'nuther ball of wax.

you can add limiting.... compression..... track automation...... you can do destructive edits like re-gaining problem areas that just got to be too much...



i always use sub groups to mix.

that's what they are there for.


i'll even automate sub groups.

leave the master alone.

you want it sitting on 0db.

you want to change the individual track faders, so at their collective PEAK, they are no higher than -12db to -14db  on the master fader.

BenMMusTech


Ok here is my take on it for what it is worth.  Firstly people need to understand gain structure, what you need to think about is when you have say a 20 track mix is think how loud the mix will be once it hits the main out.  This is called summing.  So lets say you have all your indivual tracks all nicely mixed and they are averging (sonar meters) roughly -18DB, you will notice that the main buss is still going into the red.  This is because when you add all your nicely mixed tracks together, their sum is greater than 0DB peak.  What you need to do is give your self headroom, I do this by lowering the trim by anywhere up to -15DB, depending on how many tracks I have summed together, within a mix.  This will give you the required head room to master, makes sense??

Ben

Hi Ben


You lower the trim on the individual tracks or the master fader ? Is there a differance ?!


Thanks for your post,


Kev

jacktheexcynic


i'm definitely no mastering expert, but my understanding is this:

in sonar, you have either 32 or 64 bits of floating-point resolution to work with. you can raise and lower volume all day long and it's no big deal, because when you mix down to 24 or 16 bit (integer, or discrete), it will amount to the same thing. 

for example (and this is super-simplified), let's say you did 5 things to a wave form in sonar:

+55.3, -66.83838, +89.3939, -22.33, +11.5

and you end up with 87.383. 

now lets say you did 5 other things and ended up with 87.384.

when you mix down to 24-bit, or 16-bit, the resolution is lower, and the .001 difference doesn't matter. you have 87 either way.

now, when you give your mastering buddy your 87, if he has to lower it, he's losing real data that he can't get back, and then he's going to "turn it up" again to normalize volume among all your tracks, ensure a good listening experience, etc. so that loss of quality (6dB worth, if you peaked out) is now getting turned up back to where you originally had it. and if you already had it there, why are you getting it mastered?

so the idea is to give the mastering engineer something he can work with, without having to "turn down" first. in your case, i would say do the following:

turn down the master volume in sonar until your loud part is -6dB. then, using your volume knob on your speakers/amp/stereo/whatever:

1. listen at low levels. see if you still like what you hear.
2. listen at normal levels. see if you still like what you hear.
3. listen at high (not ear-fatiguing, but loud) levels. see if you still like what you hear.

if the quiet parts are too quiet when you listen to it low, then readjust your dynamics to bring up the quiet parts relative to the loud parts, and start over. 

once you like it, send it off to the mastering dude. 

I get the gist of what you're saying. So basically its a big no no for a mastering engineer to turn down a mix as it results in a loss of data,and therefore quality....ok I guess thats the explaination I was looking for :) Yep that really pushes home the point about getting the mix right before sending it off for mastering.


Gonna try this approach today.Thank you for your post and ideas !!!!!

Kev

bitflipper


Kev, you may be overthinking it. The mastering engineer simply wants to make sure he's got some headroom to work with. This is especially important for pop/rock tunes that lack dynamics. Not so much for classical, where the ME is not going to do much compression or EQ. If your dynamics are pretty good, he might have no problem working with only 3db of headroom.


1) Yes, start with the loudest passage, and make sure no peak exceeds about -6db and you'll be fine. This will be plenty of headroom for the ME to raise the quiet parts if necessary without having the limiter engage. If only one or two peaks exceed your target, handle them in the mix with volume automation.


2) Subgroups are common practice and make mixing much easier. In classical music, ensembles are typically treated as one instrument. The conductor doesn't point to one violinist and say "you! get quieter". Rather, he points to the whole violin section and says "everybody! get quieter". The conductor doesn't have enough hands to issue instructions to every member of the orchestra individually.


3) Yes, turning down the Trim/Gain slider on the master bus is indeed a perfectly acceptable way of bringing levels down. If you watch track levels from the get-go, you won't have to do that, though. 


The standard argument against using the master trim this way is that some plugins don't like to be overdriven, so it's better to keep levels under control all along the signal path. Not many plugins have that problem, though, and you're probably not using them for classical music anyway.

Well its orchestra but there are some modern elements in it too,like drums,bass,electric guitar,of which I do compress heavily.For orchestral instruments I might use a little compression over a subgroup,but mainly the only processing I use is the reverbs and delays.  I finished mixing the piece yesterday (I posted it on the songs forum if u want to hear). I left in the end about -3 db at the premastering stage. I felt like I hugely compromised the mix though by bringing some of the peaks down at the end of the track,it feels like there should be more impact. Definately hear what you and others are saying about the importance of starting with the loudest parts first now !!!! We live and learn :-) I sent the mix to an friend who did some mastering for me,and the track came out pretty good in fact,but next time I mix,I'll be paying more attention to giving the loudest parts their due in the mix.

Thanks for your post,some great and informative points there.

Kev

I don't see that as a "start over again" situation. Just jump to the end and pull the master fader down to the correct level for the loudest part. 

Everything else will be lower and like was said above, the ME has the ability to bring the low stuff up.  What the ME can not fix is if the loud parts clip and you get distortion as a result.  

If I was going to master this myself I would use something like a volume envelope in the master to control the fader automatically.  As long as the resultant exported wave sounds fine.... to me that's what matters.