A. Hole (Composer)

"Frankenstage" is what I call a small project of mine that I have been working on lately and which was the reason behind the comparison of microphone positions. My intention have been to try to match those microphone positions and emulate that soundscape created on the scoring stage, so that otherwise dry sampled instruments easily can be blended into the mix with instruments actually recorded at a scoring stage. The goal is not to recreate any specific scoring stage but instead create something inspired by the sound of stages such as MGM and Fox.

What I did was to use the different perspectives offered by CineSamples' sample libraries recorded at MGM Scoring Stage as ground truth and similar dry sampled instruments as reference. I then proceeded doing A/B testing of the signals from the ground truth and the reference with added processing. Most of the testing was performed by ear, simply listening to the differences and trying to match them using headphones, primarily listening to the coloration of the instruments and the color of the reverberation decay while taking into account that each instrument wasn't necessarily a perfect match in performance.

While the processing of the dry signal can be done in several ways with different effects, the way I elected to set it up was guided by my intention of creating something that could easily be duplicated by others. Thus avoiding any specific plugins and by using a freely available resource.

My tool of choice thus became the impulse responses from Samplicity's Bricasti M7 Impulse Response Library that can be downloaded from the website and used in a convolution reverb of your choice. The Bricasti M7 is an algorithmic hardware reverb by Bricasti Design, but unlike previous algorithmic reverbs it model natural acoustics and tries to produce as good early reflections as a convolution reverb that is using impulse responses from real places. Peter Emanuel Roos of Samplicity has created a library of TrueStereo impulse responses using this quite expensive hardware unit and have made it freely available on Bricasti's request. Note that you need a software convolution reverb capable of producing TrueStereo (or ganging two instances together) with zero latency to use it successfully.

I selected two impulse responses from the library after some listening, one modeling the early reflections and another to model the reverberation tail of the room. Since scoring stages such as MGM, Fox and Warner are treated acoustically with wood, I found the Medium & Dark ambiance preset being a good match for early reflections while the Large Wooden Room preset has a similar decay and sound as the reverberation tail.

My set up consists of three separate instances of these impulse responses, one for each microphone position, in a special split sends arrangement. The placement of instruments are thus achieved solely by panning and balancing the relative levels sent to each effect instance while no dry signal is sent at all to the master bus. By using this split sends arrangement I get the best things from both the send and insert effect world. The three instances are shared among all the instruments that need to be processed, just like any regular send effect, but at the same time the effect acts as an insert as it completely replaces the dry signal. It is similar to using a pre-fader send with the fader all the way down, but still being able to use the fader to balance the instrument relative to the other instruments as usual.

I found that the stereo image of the input in the reverb needed to be slightly tightened as the TrueStereo treatment of the impulse responses will widen it, which makes placement by panning easier. I also found that the amount of panning usually need to be slightly different between the signals sent to the different instances. The close signal need to be panned the most, the room signal (emulating the Decca Tree) a little bit less (about half as much) and the surround the least (again about as half as much as the room). The way I balance the sends is to first set the dominant perspective (usually the room) to 0dB, I then set the other perspectives to lower levels (usually between -6dB and -9dB for the close) until it sounds about right. If an instrument need to sound further away I just insert a low pass filter in the signal chain before the sends and attenuate the high frequencies. 

The only thing I am still not entirely happy with in this setup is the surround perspective as it does not yet sound wide enough compared to the ground truth.

This is an example of what the same trumpet fanfare sounds like from three different microphone perspectives on MGM Scoring Stage. This is the natural sound of the room without any added reverberation. First comes the close microphones, then the Decca Tree and lastly the surround microphones. As you can hear all perspectives contain room reverberation of about 1 second. The main difference between them are instead the level of direct sound and early reflections from the instruments relative the indirect sound from the room.

The directional close microphones has a very narrow stereo image and contain a great deal of direct sound for obvious reasons, the sound is colored by the room but the tail is really only heard when the trumpets stop playing and is otherwise mostly drowned by the direct sound. The close perspective of each section is used to balance an usually otherwise unbalanced film orchestra or to bring out a section or soloist in the final mix for a more intimate sound.

The Decca Tree with its three omnidirectional microphones above the conductor podium has the widest stereo image while the trumpets still has a quite directional placement in it; as the setup consist of a left and right channel microphone separated by some distance and a center microphone that fills the "hole" in the middle of the image and add direction. This perspective contain a good balance between the direct and room sound and is the basis for the final mix.

The surround microphones are a regular omnidirectional A-B pair set up behind the Decca Tree but closer to the ceiling. As you hear in the example the big difference between the surround perspective and the Decca Tree is the "hole" in the middle of the wider stereo image, making it lose direction in the placement of the instruments, the sound of the trumpets seems to bounce between the right and left channel. Obviously the balance between direct and room sound favor the latter but there are still some direct sound as the surround microphones are indeed omnidirectional. However, notice that the higher frequencies of the instruments are attenuated due to the increased distance and being drowned out by the room. The surround perspective should not be brought up too much in the final mix, otherwise the stereo image phantom center will be lost, just enough to widen the image.

The final mix is a balance between these three perspectives and depends on the cue in question and the general sound of the soundtrack, anything between intimate to epic. Additional reverberation is also commonly added by digital reverb with a decay time between 1.5s and 2s that is either fed the full mix or just the surround microphones.

For fun, compare the score of Back to the Future (1985) to the score of The Avengers (2012) by the same composer and scoring mixer (Alan Silvestri and Dennis S. Sands) although at two different scoring stages (Warner Scoring Stage and Abbey Road Studio One). The first is an example of classic scoring stage only sound while the latter is an example of epic sound with added reverberation.

Further reading:

The most important to keep in mind about intervals is that they are quite few when you start to group them together and study how they work.

The first thing to note is that all intervals (except for the bloody obvious perfect unison and perfect octave) inverts to another interval. Seconds inverts to sevenths, thirds to sixths, fourth to fifths. The inverse of a major interval is a minor, why a major third becomes a minor sixth and so on. Which is why voice leading by inversion is a powerful device.

The second thing to take note of is that the inverse of a strong progressive motion is a weak retrogressive motion. While we can use strong motions to indicate tonality, we can just invert them to weak ones to avoid it.

Normally the bass line follows the root of the chord in a progression. But it can also be made independent to create a more interesting bass line melody and harmonization.

An independent bass line is indicated in regular chord notation with a slash after the chord name and the intended bass tone, such as Cmin/E♭ (first inversion of C minor). This is obviously the same notation used to indicate the inversion of the chord, but if you would be playing it on piano and harmonizing in the left hand as usual you would execute the independent bass line by inversion of the chord in the left hand if the line consisted of chord tones. However, the independent bass line is also free to use non-chord tones and are notated in a similar manner to inversions, such as Cmin/F (a C minor with an added F as the lowest tone).

So while the chord progression moves in either strong or weak motion in thirds, four and fifths the bass line can move in major and minor seconds or thirds. This device is especially effective together with chord progressions in weaker motions.

Mode mixture is the borrowing of chords from a parallel mode. Modes that are parallel share the same tonic. While there are many other modes this device usually borrow chords from the parallel minor or major. For example, C minor is the parallel of C major, D minor is the parallel if D major and so on. All triadic chords with the same diatonic function can be used in place of another, except for the dominant and the leading-tone, and seventh chords can be borrowed freely as well.

So while writing in the key of C major we could borrow the E major chord from the parallel C minor key and use it in place of the regular E minor chord via mode mixture. The borrowing from minor while in a major key are a more common practice than the other way around.

A harmonic progression is a series of chords where the root moves by certain intervals and a diatonic harmonic progression is where the chords in the progression are made up of tones belonging to the scale of the current key. There are a lot of theory and fancy names attached to the different diatonic progressions as they were at the heart of all tonal music during the common practice era. But most of that can for all practicality and purpose of writing modern music be ignored. Instead we should focus on the essence of these progressions, on how and when to use them.

Traditionally strong progressions have been favored by composers, with the strongest of them all being movement of the root down by a fifth and its inversion of movement up by a fourth. Such movement establish a strong dominant relationship between the two chords and result in a clear resolution if it ends on the tonic.

This progression can be extended to create a series of motions down by fifths (or up by fourths) known as the circle-of-fifths progression.

To which further possible progressions can be added.

Authentic progressions are movements to the tonic either from the dominant or the leading tone. The plagal progression is a movement from the subdominant to the tonic. While a half progression is a movement that ends on the dominant and never reaches the tonic. And a deceptive progression is a movement off the dominant to something else, the submediant in a diatonic progression, but can be any of the vi, VI, ♭VI, ♭vi, ♭V, ♭II or ♭ii in a chromatic progression.

These are the diatonic harmonic progressions that has traditionally been used in music during the common practice era as well as in jazz and popular music. We should keep in mind that these progressions are strong indicators of the tonality of the music. Conforming to this progression framework will define the tonality of a phrase. Especially the extremely strong dominant-tonic movement, which is why other harmonic relationships are favored in modern film scoring such as chromatic deceptive progressions, chromatic mediants and tritones to avoid clear resolutions.

Another common progression is the deceptive resolution which avoids the dominant-tonic relationship by moving mainly by weak retrogressive movements instead of strong progessive, such as a diatonic mediant and a fourth down or a fifth up (as a strong movement would be the other way around). The only strong movement in the progression would be between the dominant and the submediant (up by a second) if the progression is repeated. But it is also the least strong of the progressive movements and will not establish any dominant-tonic relationship either.

A harmonic series is the sequence of all multiples of a base frequency. The tone from an instrument is made up of a fundamental frequency, usually perceived as the musical pitch of the tone, and series of overtones that whose relative strengths contributes to the timbre of the instrument. If we lay them out as musical pitches on the staff we get the following.

Above is the harmonic overtone series up to the twentieth overtone on C. There are some useful characteristics of the series that you should observe that makes it a key device for orchestration and voicing. It tells us why octave doubling in the bass, open voicing in the middle and close voicing in the treble sounds natural and is the general advice to avoid muddyness.

Note how the intervals between successive overtones gradually becomes smaller, why chord voicing that mimics the structure of the overtone series, with greater intervals at the bottom and smaller intervals at the top, will have a clearer and more resonant sound. The fundamental tone is also repeated in each octave why doubling in octaves always will work. Also take note of the first eight pitches of the series and how they relates to the scale degrees of the major scale.

Using spacing that emulates the overtone series is a good idea for any orchestration. But you can also skip the chord tone just below the top note if you wish to strengthen the top note as a melody note.

Further reading:

We can continue building chords by stacking an additional third on top of our triadic chords and creating seventh chords. While the number of possibilities increases the most important seventh chords is the dominant seventh, that occurs naturally in the chord built upon the dominant of any major diatonic scale, and is used as secondary dominant in tonal music.

These seventh chords are just like their triadic cousins so called tertian chords as they are constructed by stacking thirds. As there are three intervals of thirds, that can either be minor or major, in a seventh chord there are eight possible combinations but only seven of them are common in western music. The augmented augmented seventh chord missing here is rarely used as it is consonant with the chord root.

Study these seventh chords by constructing them on any chosen pitch as root and listen to what they sounds like, both played together as a chord and broken as a melody. What emotions do they awake in you? Also study how you can use different inversions to move between seventh and triadic chords.

However, be aware that these seventh chords and other four note chords such as ninths and elevenths with "color" tones, the staple of the jazz language, can make any film score sound dated. So stick primarily to simple triads to achieve a modern film sound.