Making Orchestral Pieces in a DAW

[Final Kingdom]

(This question goes for any genre that's strings-heavy, but I figured focusing on orchestral makes it a bit more specific.)

How does one generally go about makes a piece that sounds full, as if it was done by an actual orchestra? I don't know much about other DAWs but I know that in Reason, there are a lot of orchestral sounds to choose from. There are lot of them that have certain sections (such as an entire brass section) contained all in one; and there are others that are individual instruments.

Would it make sense to simply use the sounds that have whole sections together and simply mix them to sound fuller, or would layering individual instruments and then mixing them be the way to go? I've been hesitant to use strings because I feel the final result won't have much depth to it.

Hopefully what I'm asking makes sense, haha. If not, I'll try my best to further explain.

[timaeus222]

Are you asking about creating a wall of sound, because that's what it sounds like.

I personally use the section patches, mainly (i.e. patches with recordings of multiple exact-same instruments together). I panned them and put reverb on them, made a short test piece, and saved it as a template. I wouldn't recommend trying to layer individual instruments (aka solo patches) because they're recorded and mixed way differently than section (or ensemble, alternatively) patches are. The solo patches are generally more exposed and more upfront (more close-miced, but not necessarily drier in terms of reverb, afaik), with a more revealing, occasionally more intimate tonal character than are entire section recordings. They're made to play some sort of leading role by default, so they take more effort to make sit in the mix well than section patches.

Reverb is definitely important; it helps to make use of Decay Time, Predelay, Attack, Spread, Room Size & Width, Diffusion, Density, etc. to place specific orchestral sections in certain spots, depth-wise. Picture

As for making them sound more full, yeah, I do some particular EQ and a little careful compression, but it helps more to layer various articulations to simulate the variety of timbre variations that go on at once in a real orchestra---even if some of the instrumentalists play the same articulation. But to really get an orchestra to sound realistic, you need rational automation of expression CC (tends to be CC11) to create volume "swells" and simulate the semi-randomness of human breath (or bowing) intensities.

So overall, what makes an orchestral piece sound full... at least, in a quick, general, gist-of-it way:

• Section patches for harmonies, layering, and getting dat wall of sound
  
• Solo patches for occasional highlights and actual solo melodies
  
• Proper reverb and panning to emulate actual instrument positions and distances away from your location
  
• Automation or Modwheel event edits of expression CC (CC11) to create volume swells, simulating human breath (or bowing) intensities
  
• Some people might suggest multi-band compression on the Master channel; might be hard to do IMO depending on how much experience you have with them but it can work
  

Maybe my decent example here can help for reference with the CC11.

Edited August 22, 2014 by timaeus222

[Moseph]

In terms of instruments, a typical smallish orchestra will likely have at a minimum:

(woodwinds)

2 flutes

2 oboes

2 clarinets

2 bassoons

(brass)

4 French horns

2 trumpets

3 trombones

1 tuba

(strings)

first violin section

second violin section

viola section

cello section

bass section

(plus some assorted percussion and maybe a harp and piano)

For the strings, each section contains a bunch of the instrument in question, but each of these instruments in the section typically plays the exact same notes. Normally, then, you will use a sectional patch for each of the string sections, since you need that combined unison within the section most of the time. It's also possible to use a patch that combines all of the five string sections into a single patch. You'll normally only use solo string patches if you (for example) need the distinct sound of a single violin playing something that doesn't relate to the rest of the strings (

).

With the woodwinds and the brass, you'll probably use mostly solo patches, though this does depend on your preferences, how much orchestrational detail you want, and how quickly you want to work. Unlike the strings, each of the brass/woodwind instruments is given its own specifically notated line (when written for a real orchestra), though each of the seconds (or thirds or fourths in the case of trombones/French horns) frequently will play the same thing as the firsts. (How often the seconds play unisons with the firsts vs. playing their own line is an orchestrational detail with no real general guidelines.) Having, for example, two flutes playing the same thing creates a problem, though: most sample sets only have a single solo patch for each instrument, and layering the same samples on top of themselves will either make the sound too loud or too phasey. There are many ways of dealing with this. You could have no seconds at all. You could just never have the seconds double the firsts. You can write the seconds a step lower than the firsts (to ensure different samples are used) then pitch-bend the seconds up a step. You can use a different library for your seconds. You can use a single library that has separately-recorded seconds. You can use a patch that has a multiple-player unison for that instrument when you need to double (especially common for French horns). It all depends on your preferences and your libraries.

Sectional patches in the woodwinds and brass that combine multiple instrument types (full brass, trombones + trumpets, bassoon + oboe, etc.) are generally used either because they're way faster than writing with solo patches or because, as a result of being recorded together, they have a better sound than combined solo patches. The limitation, of course, is that you get less control over the details of the orchestration this way.

As far as getting a full sound goes, it's largely a matter of how you distribute your material across the orchestra. You might think that writing a bunch of separate lines and inner harmonies would sound fuller than a bunch of unisons and octave doublings, but this isn't necessarily the case. Especially in modern film scores, power and fullness, particularly in the brass section, tend to be achieved by having massive numbers of instruments all play pretty much the same thing. As a general principle, don't be afraid to give the same material to multiple instruments/sections at the same time -- doubling things like this will give you a rich and complicated timbre without overcomplicating the musical lines themselves. I think never doubling anything is the single most frequent mistake I hear novice orchestrators make.

Edited August 21, 2014 by Moseph

[Final Kingdom]

But to really get an orchestra to sound realistic, you need rational automation of expression CC (tends to be CC11) to create volume "swells" and simulate the semi-randomness of human breath (or bowing) intensities.

Sorry if this is a dumb question, but what does CC stand for?

[timaeus222]

Sorry if this is a dumb question, but what does CC stand for?

Control Change.

http://nickfever.com/402/production-tips-and-resources/midi-cc-list/

[mickomoo]

for me panning and reverb help. Mastery of CCs (which I don't have) as well as mastering your particular samples articulations can help.

[Cryha]

Question about automating CC11, or I guess two questions.

Is it just volume control? I read the thing timaeus linked to up above, which says it's a percentage of volume control (but it did not say which percentage).

If so, and if I'm doing everything in FL, can I just automate the volume knob for a track in the step sequencer (edit: pattern editor?) and thereby accomplish whatever I would have done by automating CC11?

Just making sure I get things right.

[Final Kingdom]

Control Change.

http://nickfever.com/402/production-tips-and-resources/midi-cc-list/

Oh okay, that was simply enough. And thanks for the list.

Question about automating CC11, or I guess two questions.

Is it just volume control? I read the thing timaeus linked to up above, which says it's a percentage of volume control (but it did not say which percentage).

If so, and if I'm doing everything in FL, can I just automate the volume knob for a track in the step sequencer (edit: pattern editor?) and thereby accomplish whatever I would have done by automating CC11?

Just making sure I get things right.

You raise a really good question.

[timaeus222]

Question about automating CC11, or I guess two questions.

Is it just volume control? I read the thing timaeus linked to up above, which says it's a percentage of volume control (but it did not say which percentage).

If so, and if I'm doing everything in FL, can I just automate the volume knob for a track in the step sequencer (edit: pattern editor?) and thereby accomplish whatever I would have done by automating CC11?

Just making sure I get things right.

It's similar. If you don't have a sample library, which generally have CC11 available, you can certainly automate the volume knob instead (I would actually use the volume multiplier knob within the Wrapper since you have a wider capable range). For example, with soundfonts you would have to do it that way because they don't have CC programmed into them, even though volume for them is just amplitude of the currently played sample.

CC11, from what I've read, is an "additional" volume control; what that means is that you can adjust the volume one way in your DAW, but if there comes a time you need to simultaneously have volume swells AND have it reach a specific volume during the swell that happens to be louder than the current maximum (expanding your dynamic range), for example, then CC11 can help with that. Another way to say it is that it's a local volume control, rather than a global volume control.

---

In short, you can use the volume knob; it's just a little less flexible with only that on hand.

Expression is the 2nd standard control in Midi to change the output level.

Volume is meant for "global settings", let's say for an entire track or for parts.

Expression is actually an attenuation control, as it defaults to 127 = max. You can use it to create more local volume "expression" curves within parts. Which makes it essentially.... an expression control

The combination of these two makes it very handy to setup a good default volume setting for each instrument and still be able to control the output level without overruling this volume setting.

Edited August 22, 2014 by timaeus222

[Moseph]

Possibly not the case with Reason's orchestral samples, but in addition to modifying the level, a lot of sample libraries use CC11 to crossfade dynamic layers so the sound of the instrument also changes when you change the level.

[SnappleMan]

Your most important mission is to think about polyphony and orchestral section relationships at all times.

Consider violins as an example. An orchestra will have two violin sections, each with 11-20 players (sometimes more sometimes less), so that means in total they have around 20-30 voices. So what does that mean?

1. Violins can play a 30 tone chord. That chord will sound terrible and be all over the spectrum, that's why this usually isn't done.

2. Violins can play a single note with 30 voices. This is a nice sound that sounds rich and alive. This is done often. Violin sections will sometimes play single note lines.

3. Violins can play a nice 3-5 tone chord. This will sound better than a 30 tone chord, but might still be too much depending on what's going on with the rest of the orchestra.

4. Violins can play a 2 voice harmony, split between first and second violins. Leading tone on first violins, harmony on second. The second violins are often used to support the first violins.

These are just a few things you should be aware of, but they don't all translate well into using samples in a DAW. When using orchestral samples in a DAW you have to consider the following:

1. Orchestral samples are usually sampled by section, so you get one patch for first violins, one patch for second violins. This gives you 2 'real' voices (as compared to a live orchestras 30). If you play a 3 voice chord, you're already adding another 10 violins, which is already too big. Playing a full 5-6 voice chord means you're using 80+ violins, this is very unrealistic and doesn't work.

2. Break up duties among first and second violins, think of each of them as monophonic whenever you can, this will definitely add to the realism, and INTELLIGENTLY thinning out the harmonic content of your music really makes it sound stronger and bolder.

3. If your library has real sampled divisi, use it. LASS and Hollywood strings have sampled divisi. In this context Divisi means that instead of sampling the entire first violins in one shot, they broke the musicians up into groups of 3-4, and sampled each group individually. Same with second violins (also violas etc). This gives you true increased polyphony, so you can play more complex chords as a real orchestra would.

4. Stay within the most frequently used registers and voice your phrases in the most common structured way to suit the instrument. There is a reason why all orchestral violin writing uses certain scale shapes and voicings. It comes down to what's realistically playable on the instrument and how high it goes.

5. Some libraries let you use the pitch wheel to slightly change the pitch of the samples, not enough to do any interval jumps but maybe +-20 cents. This is useful when trying to emulate a real performance because the faster a violinist plays the more he/she goes out of tune. This is called intonation. The intonation of an instrument is how it stays in tune with itself as it plays up and down its register. So when doing fast note runs you can use the pitch wheel to slightly change the pitch of the notes, simulating the action of a real player going too fast to make sure his/her fingerings are right on the mark.

6. Properly learn the intended uses for short articulations: portato, staccato, staccatissimo, spiccato, pizzicato. When you listen to an orchestra playing you're hearing naturally flowing changes in articulation. Don't simply think "Well this is a melodic line so I'll use a legato patch." Your lines need to have the proper movement to sound natural, and using short articulations where they're needed is the key.

That's just for the violins. Every section of the orchestra follows those rules. You should listen to a lot of classical music and study how it is voiced among the entire orchestra. Violins can be doubled by flutes for a nice sound. Basses have a really large range and can be used to double cello lines or even play along with the violas. There are many traditional pairings that you can read up on and study. For the most part though, you want to write parts specifically for each section based on their ranges and timbres.

Most sample libraries will be sampled as:

Strings -

First Violins

Second Violins

Violas

Cellos (Celli)

Basses

Brass -

French Horns

Trumpets

Trombones

Tuba

Woodwinds -

Flutes

Clarinets

Oboes

English Horn

Bassoons

Piccolo

Think of each one of those as a monophonic voice to get the most authentic sound. Sometimes they will add Solo versions of each instrument, that gives you a true second voice to use for harmonies if you see fit. So when shopping for libraries, make sure you get ones that have more divisi options. This doesn't mean you should only use single note lines on every instrument, but be aware that you're compromising realism whenever you choose to multiply the number of musicians you're using.

Most importantly you have to learn how to properly orchestrate, and you do this by reading scores. Download some PDFs of Beethoven's symphonies and learn how he voices his chords, pairs sections, designates counter melodies. You don't have to be able to sight read, just learn to read very basic notes, so that you can see the harmonic relationships.

[The_Wizard_Lakmir]

All really good advice in here; although some of it may be overwhelming at first. Regardless of what virtual instrument library you choose to use, you're going to spend a decent amount of time getting it to sound right. You really need to ask yourself what the music is for.

Example: When I don't have a project to work on, I tend to write very ambient, lush and textural sounding music; I use a healthy amount of sustain and dynamic fading, reverb, room delay, etc. This works for the type of music I choose to create, as you can you can fill a lot of space with just a few instruments.

But this would not be the type of thing you'd want to do if you're working on a piece with fast movement and many notes. It'll sound muddy and indecipherable if you apply the same approach. This is why music used to be, and still is for the most part, written for the room it will be played in. Hip Hop wouldn't sound very good blasting in an ancient cathedral, just as Gregorian chanting doesn't sound best in a bathroom. If you're writing a fast moving piece, maybe staccato, or pizzicato and little to no room reverb is what will make the instruments sound best. Let your ears be your ultimate guide, but knowing what you want beforehand can save you a lot of time and stress, which will ultimately make better music.

[Neifion]

To make your orchestral tracks sound more full, two things: reverb (already mentioned a lot) and mic positions. Sections that are closer (ie 1st Violins and Cellos) might have a little close mic dialed in with a lot of room mic. Mid sections (rest of the strings, woodwinds) will be mostly room mics and a little ambient mic. The farther sections (brass, percussion) will have some ambient mic along with the room mic.

Have a listen to this piece I wrote for full orchestra:

https://soundcloud.com/kekopro/heart-and-hearth

It's a pretty full sound, and believe it or not, I used chamber-sized sections for all of the strings (4 1st Violins, 3 2nd Violins, 3 Violas, 3 Cellos, 3 Basses).