Yes, it is time to talk about how to mix sample based orchestral works (sometimes referred to as MIDI mock-ups). There is a lot of content about this everywhere, yet people seem to struggle finding something that is simple, easy to follow and that works, so this is it!
Most importantly, some libraries promise a good “wet” sound out of the box which is fair enough, but in reality you need to be able to treat the dry signal correctly because: a) some of the best samples out there are actually dry, b) if you’re blending samples from different companies they were recorded in different rooms, therefore their close mics are the minimum common denominator.
Is Orchestral Music different?
Let’s start by acknowledging the similarities between this and other genres when it comes to mixing:
- you still need a template
- you still need references
- you still balance your levels, then pan, eq, compress, add FX (in this order)
- you still shape sounds in context, not in isolation.
What is different though is the “wet” component of the sound, which is the single most important aspects of orchestral mixes and if we don’t get this right, our mix will never sound convincing. Because of this we need to get th reverb right at the beginning of the mix, not towards the end (see Mixing With a Plan)……it’s not a nice effect that we dial in at one point: it accounts for 50-70% of the total sound of our mix and we treat it as an actual “far mic” whether it is a real room mic that comes with our samples or an artificial one that we created by routing our dry signals through an Impulse Response.
Panning is another aspect that might need a slightly different approach compared to what we would do in many other genres.
First we will import our tracks into a mixing template, for orchestral projects we are going to need a specific template and we can obtain it by tweaking a generic Mixing Template, it still revolves around empty tracks, bus groups and FX but there are a few things to bear in mind….
- our bus groups are the sections of the orchestra: strings, brass, woodwinds and percussion. Quite often is not necessary to group the keyboards together, group the harp with the woodwinds.
- For tidiness and consistency across the different projects the template should emulate an actual orchestral score with woodwins at the top, then brass, percussion, keyboards and strings, all layed out from high to low within their respective sections.
- As for the FX chain, a good convolution reverb is all you need for most of your mixes.
If we intend to use an advanced panning tool, such as the VSL PowerPan or Waves S1 we might want to open one instance of this plug-in on each and every audio track, if we have enough CPU, otherwise we will use the regular channel panners.
So here’s our starting point: one instrument per audio track, layed out from top to bottom as we would see in an actual orchestral score, one advanced panner on each track (optional) and outputs assigned to the respective bus groups.
For the first part of this article we will make a few assumptions:
- all tracks are stereo
- all tracks are audio
- they are only dry/close mics.
We will of course see also what happens with mono tracks, with wet stems and MIDI projects so if you’re mixing your own mock-ups you don’t necessarily have to bounce everything into audio, but let’s talk about the simplest of scenarios for now, please refer to the above list and let’s dive into the main big concepts.
Setting up the Reverb
Go to your strings bus, activate a post-fader send to the convolution reverb and set it at -6dB, it’s a random value but it gives us headroom for future tweaks, do the same with all the other sections, the individual keyboard tracks and solo intruments (if present). Listen for one moment to the string section alone, it’s just a test, quickly check the inner balance of the levels, then go to your reverb channel and open your reverb, set it to 100% wet and try a few different impulse responses…..these can be a concert hall, a scoring stage or a church. What you pick depends on factors such as the density of the score, the instrumentation, the compositional style and your personal preference. You can change the type of IR later on during the mix if what you choose wasn’t quite right. Please remember that whilst your reverb is set to 100% wet the send level on the strings is -6dB and you can adjust the amount of reverb that we dial in from the send level, listen to the section alone and establish now a reasonable send level at which the strings sound natural and “enough wet”, note down the send level. Do the same for all the other sections, one at the time: set the post fader send to the convolution reverb and dial in the same amount you chose for the strings. We will be able to quickly adjust these values so don’t fret, now the whole orchestra sounds reasonably wet and it’s the starting point of our mix. Try to start your mix with a reverb that is not extreme, neither a super short scoring stage nor a cathedral, unless that’s what you want and there’s a reason for it…..start with a medium sized concert hall or scoring stage with a tail between 2.5 and 5 seconds and then experiment later on.
Listening to the sections
Like a conductor: ask your strings to play and listen to them, adjust the levels accordingly. If you are mixing someone else’s composition you may want to hear some reference material and the rough mix as well before you balance the levels. Balance the levels for the woodwinds, then the brass and the percussion. As always, it doesn’t really matter how a section sounds in isolation so you will probably tweak this levels later on but bringing all the sections together is way easier if they already have some sort of inner balance rather than raising up the instruments one at the time. Having said that, if you find it easier to level up the instruments one at the time there’s no reason why you shouldn’t do it……the only thing that matters is that you can balance all the levels to your satisfaction, with a signal that already has the reverb we chose and you can achieve the first milestone of our mix: a good static balance.
Panning and Layout
Three aspects determine the stage placement of our instruments: panning, eq and reverb, once we understand this we can place instruments on our virtual stage simply, effectively and with no effort.
In a normal mix we would primarily handle mono signals, we would generally hard pan instruments if we need them to stand out and we would more or less start with an LCR approach, clarity is very important in a general mix and we don’t like a what they call a “mono blob”. Orchestral mixes are different, we still aim for a certain degree of clarity but if we are trying to achieve realism we need to listen to real orchestral references and acknowledge that instruments never sound hard panned……the room responds to the orchestra in a very complex way and blends the sounds together, so there will be some overlaps.
Look at several orchestral layouts and try to mimic the placement by narrowing and moving the stereo width of each instrument until it sounds where you want, strings generally range across the entire stereo field with 1st violins to the left, followed by the 2nd violin (center – left), violas (center – right) and cellos (right) with the double basses behind the cellos. This is not the only possible layout and by all means not the most versatile in the context of a modern mix, it creates imbalances with too many low frequency to the right hand side…….a very useful alternative (see the picture below) features 1st violins to the left, second violins to the right, with violas center – left and cellos center – right and double basses at the back (centered)….this layout is more versatile, sounds more balanced and creates the foundation for a better mix and, for those who care about realism, it is possible in real life scenarios: some conductors actually prefer it. Now anything that we do with our mix is aimed primarily at getting a good, pleasant and powerful sound, realism shouldn’t be an obsession, but as a matter of fact our taste was influenced by orchestral recording and we tend to find credible mock-ups more pleasant to some extent, at the same time we musn’t be afraid to break the conventions to obtain the sound we want: listeners will only judge the quality of the result and not the theory behind it therefore we always need to be open minded, acknowledge the limitations of our tools and make adjustments where needed.
Double basses can have their spot at the center of the stage “behind” the other strings although we will need to tweak the reverb and eq settings in order to bring them behind, let’s just pan them for now. We can pan woodwinds and brass according to our layout of choice but there’s another important concept that we must introduce in order to achieve the desired depth: far instruments sound less panned…..this means that if our french horns are intended to be behind our strings to the left hand side they will sound a bit more centered than the strings, this is what happens in a real life scenario, this is what recordings tend to capture and this is what we can do IF we want the brass section to sound deep and behind the strings, again it doesn’t have to be that way in each mix but this panning technique is part of what we need to do if we are aiming for a three dimensional sound.
Percussions with a lot of energy in the low end belong to the center, panning the timpani to the right or left hand side doesn’t add to the realism but it makes the whole mix sound off balance and can only result in that excess low energy towards the sides being trimmed off at the mastering stage, either with a multiband stereo imager or a mid-side eq, so save yourself a lot of time and pan all the low instruments towards the center: bass tuba, double basses, bass drum, taiko and timpani….try to make room for them without hard panning any of them.
Eq and Compress in context
More than ever, with orchestral mixes it doesn’t matter how one instrument sounds on its own: do not spend time equalizing your first violins in “solo” mode only to find out that the sound you achieved doesn’t translate as expected in the mix. Equalizing and compressing in context sounds hard but it’s only going to be hard in the beginning, it is in fact the fastest and most effective way to process your sounds so if you have too much “porridge” in the low end start by choosing your main low instruments and trim off some excess bass from the secondary ones whilst listening to the whole mix, if your violas have nasty resonances or your oboes sound a bit nasal carve out those frequencies and hear the mix change as you do it. Also if your percussion has too much energy it might need dynamic processing and it has to sound good together with the rest of the orchestra so just try a few different compressors and see which one does a better job at taming those transients without sounding too obvious in the mix.
Don’t forget that reverbs often need eq too, but generally you don’t want to loose the low end of a nice convolution reverb as it adds a lot to the depth of the orchestra, it’s best to find the room main resonance(s) and do some surgical eq to just carve out a bit of lower mids or lows without altering the natral ambience too much.
Until now all sections shared the same reverb send levels and were panned without any depth added to them, now that we’ve done most of the work with eq and compression (if needed) our mix should sound clear enough so let’s establish the relative depth for each section, it’s best to work with two sections at the time in the beginning: remember that distant instruments have more reverb but also sound slighly darker.
Hear your strings and woodwinds, have the reverb send levels handy and insert one eq on each group: now slightly shave off some of the high frequencies and add more reverb if you want to push back a section, emphasize the highs slightly and dial in less reverb if you want to bring it forward. Simple as that. You will be amazed how small changes (-0.5dB from 8 to 20Khz and 1 or 2dB more on the reverb send) can really push one section behind the others…..you can work the sections in pairs (strings + woodwinds, woodwinds + brass, brass + percussion) or just start with one section and add the others one at the time. If the stereo image of one section is too wide it will be difficult to push it towards the back with reverb and eq alone so in some cases you might have to add a stereo imager to that section and make it slightly narrower.
Harsh strings are very unpleasant to hear so always make sure they have enough of the right reverb: if they need more reverb don’t be afraid to add some, if the other sections already sound good and the mix is well balanced you may as well leave them a bit drier…..you can always push them towards the back of the stage with eq and stereo imagers alone.
Saturation, distorsion and exciters can add a lot to orchestral sounds, use them very sparingly: a little saturation on the brass section can make it sound fatter and more powerful, these tools can sometimes add sparkle to the percussion or the strings depending on the mix in question but they should be really used only to add that 5% extra glow when needed….if we start adding them everywhere we will just ruin the mix: if the whole mix needs extra punch and fatness, we can add a tiny bit of saturation to the mixbus.
Don’t be afraid to change the Impulse Response in the middle of a mix or even towards the end. If you’ve set everything up correctly as explained above, your sections are all feeding into the one main reverb and you can change the IR for all of them in just one click, listening to different IRs can be very refreshing and you may find that you were using the right one already or that the whole mix sounds better with a different IR, which could belong to the same category (e.g. swapping a concert hall for a bigger/smaller one) or to a completely different one (a scoring stage or a church).
Some reverbs give you clarity and a very detailed sound, other reverbs will cover up the sounds more but also make everything smoother and homogeneous so it all depens on which direction you want to take and what works best for the score you’re mixing.
If you don’t have a good convolution reverb you can use an algorithmic one, some people prefer them actually and you will find some very good algorithmic reverbs out there from very affordable and even free VST to very expensive outboard gear. Be aware of the possibility of a slightly unnatural aftertaste especially when using stock reverbs that come with your DAW or other affordable plug-ins: it doesn’t mean that you can’t achieve a great mix though.
Algorithmic reverbs can also be combined for a slighlty different effect: you can either use a converb as an insert on each section bus for the early reflections only (lower the tail to zero) and get the tail from an algorithmic reverb inserted on the mixbus OR you can use the whole converb as explained before and then add some extra smoothness by putting also an algorithmic reverb on the mixbus. This is not always necessary, these are just options and you might want to try them and see what they do to your mixes.
These are a key component of any good mix, when it comes to orchestral samples though nothing sounds more unpleasant and fake than a loud sound at a low volume (the opposite can work actually) so ride those faders if you have to but be aware that healthy and good sounding dynamics should come from the performance and the arrangement itself. If you are mixing your own works you will overtime learn how dynamics translate to the final mix and how to put dynamics in the performance so that you will actually end up playing/programming more expressive mock-ups and doing less volume automation work in the mix.
Solo Instruments and Keyboards
We generally want our soloists to stand out, being more or less centered and slightly drier/brighter than the rest of the orchestra. Use the same reverb on your solo instruments, just at a lower send level. Keyboards can sometimes be treated as solo intruments (a piano or a celesta, for example) or can be placed at the desired depth on the stage, each keyboard instrument will have its own individual placement, depth and reverb send level.
Sometimes you’ll find that a random instrument needs to stand out more even though it doesn’t play a “solo” part, a mallet for example, in that case just take it out of the section bus and treat it as a solo instrument to add clarity and detail.
These are preferable in a normal mix: they allow more accurate positioning in the stereo field and take less space overall. Stereo tracks work perfectly in an orchestral mix because we do want some overlap and glue, having said that if you were given mono tracks by your client just pan them by ear and try to place them on your virtual stage as you would with stereo tracks.
Mono tracks can be super handy in one case: let’s say you have stereo samples that come with factory panning settings that you don’t like at all but at the same time you don’t have an advanced panning tool such as the VSL PowerPan, you can either use the standard panner on your stereo tracks or bounce them into mono first, the second option generally results in a more consistent behaviour as you move sounds across the stereo field.
These are great for the ease of use and a great “out of the box” sound, some claim dry samples with reverb will never sound as good. But what if you are using different libraries recorded in different rooms?
You can follow the main method explained in this article and isolate the dry signal (the minimum common denominator) of all libraries. OR you can dial in some of the wet sound as well and see what happens, nobody knows what happens until you try and blend them together: some different libraries sound great together and you will be grateful you have chosen to preserve the room mics as well, in some cases you will use some of the main reverb as well on top of the dry+wet sound or on the dry component alone combined with the room mics in parallel. It is important that you are aware of all the different possibilities and choose the option that sounds best for the given track and with the samples that are available to you.
But how do you proceed?
If you are mixing for someone else and you’re mixing audio tracks you should request the distant mics as separate audio tracks, you will be able to process them separately if needed (eq, saturation etc…) to add reverb to the close mics if needed or to the sound as a whole: this is the best option in terms of flexibilty and control, if this isn’t possible try to obtain at least some of the wet stems, if you’re being given only tracks where the close and distant microphones are already combined there’s still a lot you can do: you can use compression to bring up the transients or to make the room more audible, you can use mid-side processing to affect the signal of instruments that are more or less panned. You can always add more reverb to a wet stem whilst it’s not easy to do the opposite (and requires ad hoc tools) but bringing up a little transient information will make any recorded track less roomy.
Don’t forget to level up the wet stems when you balance all the levels at the beginning of a mix and then tweak the settings later for a more accurate stage placement and more clarity/depth.
This is a great option for flexibility and works perfectly if you’re mixing your own music, it can be less practical if you’re mixing for a client for obvious reasons, but if you’re the composer and have enough CPU power to mix your music whilst the whole project is running then go for it, there’s no need to bounce the tracks.
Make sure you have a very tidy orchestral template and route all the signals through their designated section stems:
- if your instruments are dry use the main method explained in this article
- if your instruments allow inner mic adjustment set those levels when balancing all the signals at the beginning of the mix, you can always do some fine adjustments later on
- if they feature internal advanced panning tools use them to define the stage placement alongside the conventional tools
- when you do micro-automation work towards the end of the mix don’t be afraid to use the key editor if needed, try both volume and velocity automation and see what works.
Hopefully this will clarify some of the main aspects of orchestral mixing and make your life easier: if you learn how to trust your ears, listen to a lot of reference material and rely on very tidy templates both for your compositions and for mixing you will soon find out how small changes can shape the sound of a whole orchestra and that a deep, clear, crisp sound is not difficult to obtain.
If you have questions about this article don’t hesitate to contact the author.