Connect with Tim

Error Code 7 – A Deep Dive: Part 2

Part 2

This is the second half of my discussion of the piece I recorded with Hollywood Scoring, where we look at the melodic section of the piece from bar 32 onward. Be sure to check out Part 1 so everything here makes sense.

Main Theme

I recently bought a new piano and walk past it on every trip to the kitchen. I remember as soon as I got confirmation that this session was happening, I was on my way to make some vegemite on toast when I stopped at the piano, put my hands down, and the first two bars of this theme came out right away. This is about the limit of my piano skills too. The next time I came by to make a coffee and grab a bikkie I worked out the consequent phrase. The harmony alternates between A minor and Ab Major, with C being the common tone.  The tune follows that harmony. I like harmonic surprises, whether that is a note in a chord you were not expecting, or a whole chord change, and I think I managed to get lots of both into this theme. 

Instead of two four-bar phrases which were probably going to be the same, at the seventh bar (38) I go to a different place. I use a first inversion chord here, which I find is harmonically neutral in this instance, and use a kind of sequence to get to the F major chord at 41. I leave in the Bb from the previous chord, making it an F with an added fourth, a sound I had set up in bar two of the theme (33), with the melodic C over G major harmony. I can hear my harmony teacher telling me that the fourth always replaces the third, but I love the sound and rules are made to be broken, just not the one that vegemite needs butter, leave that out at your own peril. 

To add interest I add notes and lines that provide slight dissonance and resolutions, which is also something I set up in the first two bars. The first tension note is the Bb on beat 3, but you don't even notice it, so I will give it a two out of ten. My scale is from one to ten, where one is consonance and ten is just plain wrong. You should start to get interested when you hear a three. A five makes you smile. The C in the G chord in the second bar of the theme is a five. A six, you start to get a little worried but so long as you resolve you are OK. A little vinegar face may start to appear at seven (four to seven is firmly in Prokofiev territory). Once you get to eight you have to wonder if the composer is deaf, does not know what they are doing, or is just taking the mickey. A ten is like eating vegemite off a spoon with no butter on toast, it just should not be done.

An interesting thing about this tune is that if you just play it by itself, it is pretty basic and a little strange thanks to the half step relationships between the first three notes. They do not exist in any western key. But as soon as you add the harmony, it sounds warm and more like it is meant to be, at least to my ears.

Bar 43 is like a pre-chorus, building towards bar 50. We could call this the chorus, or if we want to get totally geeky we could look on the form as a rondo and this was the B section heading into the repeat of the A, but with a variation. Instead of starting on the A minor and going to the Ab major, I start on an F major, which is close enough. I then have Ab minor going to major but if you look and listen carefully to the flute and clarinet you will hear both thirds going. These notes are B and C which are common enough to the surrounding harmony that they can pedal through it. This time I do make it a traditional eight-bar phrase, but I change the harmony up in the fifth bar.

Bar 58 serves as another transition, or a C section to get us to the final A. If you look at the chord shape from second violin down, you see the familiar voicing with root, fifth, and major seventh that I used in the intro and first transition. In bar 61 the main motif is in the horns. The harmony is built off the descending bass line. I wanted something nice and dark. Gm, Gb, Dm, Db, C, Gb/Bb, Am, Ab all with a Bb pedal. Somehow we end up in G minor, guess I screwed up somewhere, we should be modulating up not down! This is the same melody and harmony as the first A section, just orchestrated in a much bigger way. The coda is like a third sequence of the material in bar 71, but much softer and wrapping us up to get ready for a final resolution in Ab major (Oops, not sure how that happened!)


I start off simple, with lush string voicings and we still have the mutes on from the previous section. That lushness is all from the lower parts. If you look at the cello, we have the root then the fifth, then we have the third in the viola. This gives us the range of a tenth and is the key to this sound. The basses are an octave bellow the cello, giving us a solid foundation. One could fill in the root again in the viola part, but it is not needed at all as we have plenty of that already. I keep first violin all in unison and freely divide violin two, viola, and cello when notes are needed based on voice leading and the sweet spots of their ranges. This is standard operating procedure in this type of voicing and range. A lot of the impression of the ensemble size comes from the top part, so keeping the firsts undivided can allow you to divide other parts and the ear will not notice the balance changing. Only when it gets really high would I go the other way and have the firsts divide.

I could go div. a3 in the violins but that is silly here as it makes a voicing nightmare for the players switching lines. Save the three way for when it is a longer phrase in three parts or the voice leading makes sense. A good way to test for this is to imagine you are somewhere in the back of the firsts and when there are three parts you play the second one, is it easy to read and follow the line and does it make musical sense? I see a lot of inexperienced orchestrators thinking they are being cool going back and forth into three parts, because they can, with no regard for the poor player jumping about.

In bar 32 I had a few options but chose to divide the cello. Firstly, it is the biggest sound, so dividing it even when there are only six players is going to maintain the best balance. Secondly if I had the viola on the fifths, in bar 35 they would hit their open C, which runs a very real risk of not blending well. Also these particular notes are fuller on the cello than the viola. For the other splits, you can see that voice leading made all of these the most obvious choices. While the string section is quite forgiving when it comes to divisi, you can really put some extra butter on that vegemite toast by thinking about it musically.

Notice there is no text saying 'div' or 'unis' anywhere on basic two way divisi! Totally unnecessary in any situation, studio or live. The way we were all taught to indicate for strings is backwards, and I go into it more in 'Divide and Conquer'. The 'div. ord.' text above first violin is just canceling the previous section where the back stands had their own part, now they fold back into their regular section.

Most people would write some swells in a section like this, the natural thing would be up one bar and down one bar. I purposely left this blank so I could demonstrate how a section thinks and plays something like this 'by default.' Let’s look at the first phrase of the first violin. I have given them two pieces of instruction, and that will lead to two other consequences in their playing. Firstly, it is soft, and secondly I want the phrase done one bow to each bar. Being soft, they will want to start with an up bow. That is because the bow is lightest on the string at the tip and they can easily and musically start a phrase. If they start on an up, then of course the next bar is a down, so they are ending the phrase at the tip again, where they started. The physics of this mean that the phrase has a natural swell to it. As they get to the middle of the bow, the sound becomes a little more intense and louder. In the fifth and sixth bar the line rises, and the players can just tell that they should keep going and not diminuendo. Not only can you hear this in the recording, but you can totally see it in the waveform.

What I wrote:

What I could have written to get the same thing, and what they actually played 'by default':

You will note that after all of that banging on about what they do with nothing, I had to go and ruin it with the ugly tie over into bar 42 and the 'senza dim.' text. While I knew what they would do in bars 32-40 with no info and it would be perfect, I also knew what they would do in bar 41 would not be what I wanted. If you follow the bowing pattern, we would have ended on a down bow (and even if we had not ended there, there is a good chance they would have switched somewhere to end on a down). They always like to end soft phrases on a down bow, as it’s the easiest way to create a musical end and release to the phrase. Can they end a soft phrase on an up bow? Sure. But it is just not as natural or as easy. And just like in the opening phrase where the end on the down makes a diminuendo, they will do the same here, but as they play nothing in the next bar they will go down even lower in dynamic. Strings ALWAYS release the pressure on the end of a phrase unless you tell them otherwise. One way is to make it a crescendo; they will then aim to end on an up bow as they end up at the frog, the heaviest part of the bow, and can project a more confident sound more easily than at the tip. But musically I don't want that here, so the only way to do it is to nicely ask them not to do what they were going to do and end without a diminuendo. The other thing string sections always do is end phrases with a final long note early, before the barline. The best way to get them to hold the full bar is to trick them by tying the note onto the downbeat of the next bar. I swear in many cases they still end early!

You will also note in the fully notated version I just wrote poco. As I discuss in my post "Swell Enough," I find that even if I mark a swell as piano to mezzo piano, they will overshoot it. My solution to control the swell is to use poco. Just a reminder, poco is not a dynamic, it is a description of the hairpin, so it goes with the hairpin, not at the end of it, don't crescendo to 'a little' you do 'a little' crescendo. You can actually hear this in bar 43 where I mark pp to p, but it feels louder.

As this section naturally builds, I bring in the flute and clarinet, then the horns and bass clarinet to fill out the harmony.

At bar 43 I originally had the bass in, playing the octave with the lower cello part, but I wanted to feel bar 44 more and a good way to do that was to leave out the bass for the first bar of the phrase. It is amazing how much you can gain by leaving things out. At bar 44 I do a neat trick with the clarinets where they oscillate F and A. I wanted to feel a little motion here. I could have had them repeat the notes, but by trading off, I get the harmony and a cool oscillating effect.

with samples

In bar 47 the horns take over the lead. One big difference I often see between how us 'Hollywood' types write and how legit composers and orchestrators write is that we love unison horns. There is something special about how multiple horns blend. We have all heard the stories about avoiding unisons in certain sections (I for sure rant about avoiding oboes in unison) but horns are the opposite. Two horns are fine, three work really well, and right up to six you can write just about anything you would write for one and get a lovely result. I think once you get to eight horns, unless it is something epic, you do start to have to think about it as it gets pretty heavy.

In bar 48, I hope you recognize the pitch set in second violins. I use it to build on the subtle motion I set up in the clarinets. This leads rhythmically to what the woodwinds do in bar 50. Notice in bar 49 I have the cello and bass articulate beat 4 to help push us into the next section and draw our ear to the downbeat. I also deliberately have the melody rest on beat one so we feel the rich chord first.

At that point we kick it up a notch, I have the strings at mezzo forte and now playing two bows to the bar, which leads to a fuller, more intense sound than the last time we heard similar material with a single bow for each bar. Even if I had left the dynamic at mp, it would be louder. The issue with leaving it at mp is that some of the players would ignore that I wanted more bows and still keep playing one per bar. It is much easier to play softer, with good control and blend, using less bow – especially in the studio with a small section. (For more on string section sizes, check out this post). In this case, though, they knew it was me who wrote it and that there was a reason for it!

Listen to how the firsts phrase this line with no information. Each phrase tapers off, and ends a little early, they just can't help themselves! While the strings carry the tune and harmony, I have the horns mostly in unison playing a counter line. I have third horn play with the lower second violin part to bring that rub out a little. The woodwinds are providing some motion. If you look at the example below you can see the top staff is a reduction of the flute and clarinet parts. Many people would just leave that line as is. One issue is that they have to breathe. If I did write this, one could just leave it up to the players to breathe. Please don't insult them by saying 'breathe when necessary.' If the part was in unison many would write 'stagger breathe.' If you read this article you will see my thoughts on that. My preferred solution is to use this type of figure as a starting point to get creative and create new textures. Dovetailing is the most common way. But like I mentioned earlier I like to take it to the extreme, E.A.D., Extreme Australian Dovetailing. Instead of trading off, they all play and split the line, letting notes hang over, creating cool blurs and textures. This way it won't be a problem when they grab a breath as there are still things going on.

You can see in this example I have the flute take the first two notes and hold the second one. The clarinet rests on the first note, plays the second, and then holds the third. Depending on the harmony, it could hold over and have no rest, but in this example it would not work too well to bleed over the harmony changes. We would get some eights and nines on my dissonance scale for sure!

with samples

The bass clarinet plays the roots. I could have given them just long notes or no rests, but again they need to breathe and they would have done that at the end of measures, leaving out an eighth or making one really short, not what I wanted. All of the note changes are well-covered in the low strings, so we don't need them, so we can leave them out and they can breathe in any of the rests. You will also notice there is no articulation on this. I would say 97.765% of people would feel the need to say legato, or put tenutos or slurs and tenutos and write some essay on soft tonguing but maintaining the pulse. But I always start with “what will they do with nothing?” as I am lazy and don't want to do too much work. And the answer is all of the above. There is nothing telling them to play short and it is mezzo piano repeated notes. They need to be articulated to speak, so just give a dynamic and you will get light articulation and full duration.

I build this section to a forte then drop right back. Notationally, note that I did not say subito pp in the top violin part. From what I see out in the wild, most would, but it is so obvious what is going on as there is the forte at the end of the bar. The other not so obvious trick I pull here is that we go from ten violins on the top line at the forte, to six on the pianissimo. Bar 58 is marked div. a3 which means the violins go into three groups of six players. A few other ways to make this note softer would be to drop out the front stand or two, that pushes the sound back away from the tree and of course reduces the number of players as well. I could have the back stands take it. The voice leading for some of them would not be ideal so they would have to drop out for a bit to shift. If this were coming from a bigger, longer, more intense section I would just stop and start again. It is really hard to get a beautiful controlled soft note with all of the adrenaline and energy built up from playing loudly, so getting a fresh start leads to a guaranteed result and no one in the theater will know the difference. The actors did not shoot the movie or scene in one take, so we don't have to either.

To contrast the previous section, I thin this one out. The wire in the top violin part glues it to the previous section. I have the shape from the opening in the rest of the strings. To add some color and interest I have the violas slide their notes. I have learned that while slides are fun, it is easy to overdo them and things can sound cheesy pretty quickly. By just sliding one part you create something that is not expected, nor easy for the listener to work out right away especially if it is on the inside of the voicing. To help this one come through a little more, I have them play sul pont. 

As the woods have been in a supporting role, I finally feature them here. A cascade finishing with the horns joining the bass clarinet on a dark version of our main motif. Notice how beautifully the woods play this line. We could use many words to describe it, but it took none to get this, just the right dynamics, articulation (or lack thereof here as it is all slurred) and orchestration. The bass clarinet continues the line, handing it off to the horns. Having no slurs on the horns here and the mezzo piano gets me the perfect articulation, look mum, no tenutos! Notice how musically they end the phrase at bar 62 and breathe, all together, 100% predictable.

I wanted the strings soft, but intense, so no slurs. I will admit that I originally had this at mezzo piano but it was too loud. So I asked them to play piano, but fight the urge to slur the notes. I fixed this score to match what was played. I always am weary of leaked session scores and people getting too excited, there are so many changes made at sessions, some huge, but the scores are never fixed to show that. The only example is when they are edited for a live to picture performance or published version. But even then they will often don't match the original recordings. I have conducted many live to picture concerts and found heaps of little things that were not fixed.

We did another pickup at 61 to get the mutes off. I did a similar trick to last time; the bass clarinet finished on its own and did not play on the pickup. In bar 64 I had the clicks go to eighth notes to help set up the people who have sixteenth notes in bar 65. I kept the eighths for bar 66 then back to quarters in bar 67. In general, players do not like eighth note clicks unless they really need them. Even then I find their idea of when they need them is different to mine. Like I mentioned in my conducting article, every player in the band knows when it is out of tune, but since they can all play out of time, together, sometimes they don't know they are off. At slow tempos subdividing the click helps bring them back onto the grid, but once they are on it, the eighths are often not needed, as was the case here.

The Final A

This is the climax so I pulled out all the orchestration tricks. The violins have the tune in octaves with two bows per bar. The basses hold down the low end while the cellos play an arpeggiated counter line. I use non-chord tones on strong beats to create good dissonance, resolving each one.

But the fun happens in the violas. Just like I did with the woods, I split what would be a cliche and boring line up into something unique and interesting.


Beat 3 of bar 66 looks like a mistake, but the chord in that bar is Fadd4 and when I was working this passage out I heard the rub in my head at the point, which is rare, usually it is just voices. If you want to hear another example, check out this cue from Maya and the Three, you will also hear an epic bass slide.

The horns outline the harmony, playing triads. If this was a large orchestra with full brass I might have the horns all in unison on the top line, or some other nice counter line, and the trombones playing the pad. And in a case where I had this ensemble but four horns, I would have the top line doubled in this case and 99% of others with this texture. Like I already mentioned, horns love unisons and it just makes everything better. Seeing as it is the top line that we hear most, making it better by having the extra player on it just makes sense. I know there are all sorts of other theories about what note of a chord should be doubled. Even if I did subscribe to those theories, which I don't, it is just not always practical. A part could be skipping all over the place and if you then start to smooth it out to make sense – which is an OK move to make – then there was no point to it in the first place! (Again, more details on the French horn are here.)

By this stage I should not need to explain what is happening in the flute and clarinet. I wanted to control where they breathe, so I have some rests to help them. I have the bass clarinet working with the basses. It is a good idea to have this octave here as the cello are now playing something else.

with samples

Check how well it combines with the woods to form a percolating text that provides motion without sounding like a riff or cliche ostinato.

Woods & Violas

Bar 71 is the biggest moment of the piece. The violins end up in the highest range of the piece. I also took out all of the slurs so every note gets its own bow, leading to an even more intense sound. The viola and cello play sixteenth notes, taking over the motion from the flute and clarinet. The notes they are playing are nothing special, the same as they would have played normally, but by instructing them to play 'on the string,' the notes are connected and form a motion texture as opposed to a driving riff if they were to play it 'off the string.' 

I go into more detail here, but in a nutshell bowing technique falls into two types: on the string, where the note starts and ends with the bow 'on' the string, and off the string where the note starts and ends with the bow above, or 'off.' Traditionally the default, with no other indications, is to play on the string. Technically I would not have needed to say anything here, but as this was a session with studio musicians who are used to playing lots of riffs and motor rhythms off the string, I marked 'on' to make sure no one played it the wrong way. In many cases these days I use this to my advantage, knowing what they do based on context. I just write the eighths or sixteenths without articulation or indication and the players will play it off the string or 'off-ish' if we want more attack. I just finished a day in the studio and I did not indicate anything for the 8th notes and every one came out exactly as I wanted. I also never tell them what actual stroke to play, that is stuff they sort out and don't even think about, it just happens organically. In fact in the entire day, I did not change or mention anything to do with note length.
It is not in this piece, but one thing to watch out for is you have a rhythmic passage with accents and normal notes, the accents will be longer than the normal notes. This can sometimes lead to rhythmic issues as the bow movement is not balanced. If they are having trouble, I ask them to dial back the accents and it will even the timing out. As with everything bowing, it is all physics.

You will notice that I tie everyone over onto beat one of bar 75 as I did 42. If I did not they would end a little early with a slight release, I really think I am repeating myself now. At this dynamic and intensity, you don't have to tell them to avoid the diminuendo as you do at softer dynamics as I had to at 42. I also did not need to write anything fancy for the viola and the cello. When strings have a passage of short notes followed by a rest like this, the last note will naturally be longer as they have nothing to get ready for in the next bar. It is very natural, so natural you probably never noticed it before. 

I have the cellos divided, half on the sixteenths and half with the basses. I add some harmonic interest by changing the chord inversion and add a tasty slide in bar 73 as we head to the last big note.

The flute and clarinet are playing the same notes as the viola and top cello sixteenths, and this helps glue it together. The bass clarinet is on its own filling in the gap between the bass part and the cello motion. They have some sixteenths for interest and as a way to blend with the viola and cello sixteenths, but they also work with the bass part.

The horns are in unison until the last chord. Note the slurs are carefully placed. Horns can perfectly legato tongue, sounding similar to a sostenuto bow that just changes direction with no emphasis. This is what you will get articulation wise with no other information from pianissimo to mezzo forte. Once we hit forte they will articulate the notes more to push them out, slight gaps will start to appear in musically appropriate places and be more exaggerated at fortissimo as the tongue must stop the previous note in order to articulate the next one.  This effect gets amplified more in the lower brass as it takes more energy and breath to articulate. It is also observable in all wind and string instruments. The energy to articulate a loud note has to come from somewhere, and that by default is robbed from the end of the previous note.  If you add an accent, the detachment from the previous note will be more pronounced.

When horns slur a second it sounds normal, like any other wind instrument, but when you get to thirds and above with slurs you get this lovely harmonic smush between the notes. Not a gliss where they use their fingers, it is subtle on small intervals, getting more pronounced on larger ones. As the horn is so long, most of the notes they are playing are in the mid to upper register, so what sits on and above the stave when transposed are all high partials. This makes the horn quite difficult to play as many notes close together can be played on the same fingering. All brass instruments share this same issue, but their standard playing range is in a lower set of partials. Of course intonation is better on different fingerings as they are based off different fundamentals. While probably not evident to most listeners or orchestrators, horn players make lots of context based decisions on what fingering and side to play on.

Due to this quality in horn slurs, I really think about what I want them to do and it is then not uncommon for a unison passage I am writing to have different slurs in the strings, horns and woods or anyone else playing along as they all react differently to them. For some more obvious examples, check this new version of the Disney Logo that I orchestrated and conducted featuring Dave Everson on principal horn and seven more of LA's finest.

At bar 76 we do another pickup to put the mutes back on for the coda. The tune starts in the first violin but then gets picked up by the flute. Note the flute has a repeated note under a slur. They legato tongue the repeat. Unless you have been sloppy in your writing and have a track record of missing ties, this is how they will play it. Most people would feel the need to add a tenuto to the repeated note. For years now I have not been doing that and no one has ever failed to articulate or asked if I missed a tie. I took the clicks out here and the just followed me. A very natural, and unmarked ritard. happens.

If there is one thing I would do differently next time it would be the cello in bar 77. My voicing is good, I am heading towards the last chord, a low tenth voicing, almost as low as you would go with this shape including the low fifth to not get too muddy. But the slide in the bass has no sonic room. I would give the bar a 7/10 on my dissonance scale thanks to the minor 9th. A few solutions I can think of would be to have the bottom cello play in octave unison with the bass, but then that is not the best voice leading to the last chord. Alternatively, I could just have the cellos in unison in that bar. I could also move the violas up a bit; the minor 9th and the slide are a bit much together, so maybe one of them has to go.

Take a close listen to the ensemble as they end the last note, they all do a little dim and end. The strings all end up on a down bow, finishing at the tip. The woods hear what the strings are doing and follow. This is not in the score, but it is just how an ensemble plays. Just like I knew at the beginning that there would be a light start and a little swell on the first long note, I also knew at the end here they would do this little diminuendo. Having no click at the end also makes them listen to each other a little more.

One Last Thought, keep it clean.

In this whole piece there is not one tenuto for anyone, but especially take notice of the horns and woods where I know many would feel the need. In my film orchestrating and arranging I rarely use them to control the attack of a note in the woods and brass. In general I see composers and orchestrators using tenutos way too much. The topic is too deep to go into here, but please read my post on this topic so you can tenuto responsibly!

I hope everyone notices and remembers is that there is no descriptive or emotional text anywhere in this cue or any of my work, the page is always clean and uncluttered with the only information being what is absolutely required. I have found text is just not needed in 99.537% of situations. We can use many words to describe what we think or feel when we hear the music, but I do not need to put them on the page for the players in order to inspire that result. I just need good notes, orchestration, the right dynamic, slurs and articulation, if needed of course. I am not leaving it 'clean' for the players to do their own thing, I know exactly what they will do using my concept of the "orchestral default." Unsurprisingly when I see a score with an over abundance of descriptions there are also a lot of – drumroll please – tenutos!

I know in some earlier modern music, especially when people wrote by hand, the score and its calligraphy are as much a part of the art as the music itself. But that is not the case in the scoring and pops world. It is a goal for me to get the results I and the composer hiring me want with minimal information on the page. I know I am extreme with this, but after hundreds of hours of recordings and performances I have proved my point. Unfortunately, this means my scores just don't look as cool on Insta.

And finally, I know I do say 'context' a lot and it is a big subject, but all I can say is that once you have done this long enough and had the opportunity to do the experiments I have, you can predict the contexts with great accuracy. I hope to have more time to write blog posts about my findings. It could be the vegemite gives me a special sixth orchestration sense, you should all try it.

Thanks to Noah, Adam, Michael, Ed, and the whole Hollywood Scoring team for giving me the time to record this. To the amazing players led by Ben Jacobson who pulled off this recording in 25 minutes, To Max Karmazyn for bowing demos, to Thanh Tran for the score and part printing, and to Adam Michalak for fixing my dodgy edits and doing a fantastic mix. And as always, to Ryan for editing this and making me sound almost intelligent.

1 Comment

  1. Brilliant, mate.


Leave a Reply