Balance and volume

[faonfaon]

Hi,

I would be interested in the way you work on that subject.

Let's say, I have a string orchestra score written in DP in a very classical style.
Now, what I usually do, is to create 5 new tracks (v1, v2, vla, cello and db) pour record the volume of each track.

My problem starts there. If I start with the violin 1, it means I have no audio reference and I never know if I am too loud or not. How do you do for this task ? Is it good for example to have the piano reduction to play with ? I don't know....

If all the volumes of each track are OK, then I have to make a balance between the five parts which is not so difficult at this point.

Thank you for your help and explanations

François

[FMiguelez]

One word: experience.

Actually, I'm not sure I understand what you mean... are you talkng about the MIDI programmng stage of the instruments? Or are you talking about the MIDI mixing stage, or the audio mixing stage?

You probably need to set up a monitoring reference you can trust, like the K-20 or similar. I never bothered with calibrating a specific monitoring level because, by the time I wanted to do it, I already had been using my own metering calibration instinctlively without knowing. I already knew when something was too soft or too loud after years of doing it.

Same with full mixes. Just a quick glance at my hw mixer's levels, and I know how much headroom to leave, if thngs are too soft or loud, etc.

Taking that into account, what works for me is to start composing or programming the individual instruments with no worry about volumes at first. At this stage I just concentrate on making the instruments sing and obtaining a great programmed performance. I mostly care about the individual instruments' dynamic levels at this point. This is NOT a "method", but something I usually do in general, BTW.
SO, as I go, I mute/unmute instruments as they get programmed to check how well they sound together, checking for realistic and nice interpretations of individual phrases, etc. But I'm mostly concerned at the instrument level here. Little by little I start unmuting things and correcting levels at the instrument-family level, but not too much worry yet...
Once I have that, and I'm happy with the credibility of each instrument, I balance everything with a rough "MIDI mix". At this point I make sure the overall music level is good for each family and the whole orchestra. The loudest parts should have plenty of headroom (6 to 12 dB depending on style or the rest of the instrumentation). The soft parts should not be too soft either. It depends. This is where knowing by heart what a real orchestra sounds like really helps! So I keep balancing the sections. Noramlly, by the time I'm ready to print all the tracks to audio, I already have a decent "semi-static" MIDI mix of everything.

Before I get too involved in automation, but once everything starts sounding balanced, I unbypass each tracks' compressors. Make sure the compressors are not working hard, just taming a little of the exaggerated dynamic range for each instrument. I do this as I go. Then, once the general levels are fine, I print everythng to audio. None of the plugins get printed with my technique, so I can always tweak them as needed without commiting. Then i start heavily automating everything. As I do this, I check my individual stem compressors, making sure they are not working too hard. And I mix and mix and mix until I'm happy
If that ever happens, then I wrap it up and I print my final mix & stems for my mastering.

Having your favourite orchestral tracks handy for referencing and comparing is always helpful.

Hope this helps.

[faonfaon]

Ahh ! Thank you FMiguelez !
Your explanations are long and very interesting thank you but what I remember is 'experience' !

I didn't explain so good about my question.
I was talking of the curve of volume after the notes were entered. Sometimes I play the notes using the ModWheel for expression 11 or volume 7 but when the parts are too difficult to play I enter them slowly not caring about the expression and then, second task, I create another track (same as violins 1 let's say) to create crescendos and descrecendos.

But when I do that, I have no orchestra playing at the same time that I could follow and could give me a way to adjust my crescendos, and it is difficult for me to create a good expression. The track is alone. I only have a click.

Hope you understand better my problem ?

[terrybritton]

@FMiguelez - knowing how fast you have to work considering the budgets you deal with, this compact exposition of your working approach is pure gold! It should be entered into the "DP Tips Sheet" thread!

Terry

[FMiguelez]

@FMiguelez - knowing how fast you have to work considering the budgets you deal with, this compact exposition of your working approach is pure gold! It should be entered into the "DP Tips Sheet" thread!

Terry

Thank you, Terry.
But I only work fast when there are tight deadlines to beat. Were not for them, I'd probably be still working on my first piece ever!

Ahh ! Thank you FMiguelez !
Your explanations are long and very interesting thank you but what I remember is 'experience' !

I didn't explain so good about my question.
I was talking of the curve of volume after the notes were entered. Sometimes I play the notes using the ModWheel for expression 11 or volume 7 but when the parts are too difficult to play I enter them slowly not caring about the expression and then, second task, I create another track (same as violins 1 let's say) to create crescendos and descrecendos.

But when I do that, I have no orchestra playing at the same time that I could follow and could give me a way to adjust my crescendos, and it is difficult for me to create a good expression. The track is alone. I only have a click.

Hope you understand better my problem ?

Oh, I see...

Well, some of what I wrote still applies... So your problem is that you don't have something to reference against to set your levels, and you have trouble "calculating" how much volume for crescendos/descresc, etc., especially when you are dealing with one or a few tracks. Did I get it right this time?

If so, then I'd suggest:
I see in your signature you use VSL, like I do.
I REALLY recommend using the cresc./descresc patches everytime instead. They sound so much better snd more real than simply drawing curves with CC automation. There's no comparisson!
Also, those patches automatically give you the natural range of a natural cresc. of the instrument.

These patches are so much easier to balance against the rest of the articulations like this.
Remember that you have LOTS of ways to adapt the cresc. lengths to what you need. There are different length and intensity patches yu can choose from, and if they don't fit, you can use the stretch tool, or better yet, combine sustained patches with a cresc. and use the delay function together with the Cell x-Fade mode. That way, you can tweak the lenghts and they will sound better and more natural. Just make sure your tempo is set and you are happy with it, otherwise you have to re-tweak all that again!

But the bottomline is, whatever method you use, you simply need to get used to "calculating" by ear how much volume or expression is needed and when it is too much or too little. Do not worry too much about it in the early stages, especially when you don't have other instruments that help as reference to set/ balance these cresc. levels against.
As you keep adding instruments, any exaggerated and miscalculated cresc. will jump at you like a sore thumbs anyway. It's qute normal for me to go back and re-tweak these patches or readjust the volume of the cresc. as I add instruments.

Listen to LOTS of classical music, particularly from the solo literature, so little by little you instinctively know how much is too much or too little, and when things sound natural. I suppose there's no way around this. You MUST KNOW the sound of what you're trying to emmulate.

It's all a careful balancing act that keeps progressing and in need of tweaks as you go. What works best for me is to do some basic instrument and family leveling with MIDI, but the heavy level automation usually comes later, once everything is printed to audio.

What I love about my workflow is that I can also start doing basic EQ and compression still in the MIDI stage, so this prevents me from having to re-tweak levels too much when I start automating the audio parts.

If you prefer drawing the CC curves yourself, do not worry if you miscalculate: it is a piece of cake to scale up or down any CC with DP. You get to keep the same shape, more less or more exaggerated, as needed.

Hopefully I got your issue right this time and I didn't bore you to death with nosense...

[faonfaon]

Thank you so much FMiguelez ! Yes, my problem was exactly this one.
And you gave me some advices such as choosing crescendo and descrescendo patches that I didn't know. I will try to do it.
Thanks again !

[FMiguelez]

Awesome!

Another VSL tip for you:
The second best way to emmulate a cresc. or decresc. is to use the sustain patches that have 3 or 4 velocity layers, and use the Velocity X-Fade to control the volume and intensity over time.
This works best more for expression and not exaggerated moments, especially if you don't need a real cresc., but more like little expressive swells.

But for plain cresc./decresc., nothing beats the relevant real patches as were naturally recorded with cresc., especially if you combine them as I wrote in the last post.

Have fun with your dynamics!