Those that heard the rather controversial piece I put up a few weeks back gave me positive reviews, discounting the inability to separate their political views from the piece itself. However, I digress.

I have decided to undergo a completely different project in my spare time, orchestrating the final fantasy 6, or for the American folks, final fantasy 3 ending, which to my knowledge, no one has done an adequate job of yet.

This is a very long ending, around 22 minutes in length, something which was unheard of for the Super Nintendo.

This is the very beginnings of this project, I have converted the original smc file straight off the cartridge to midi, mapped everything to use the actual note range the instruments can actually produce (I think haha...I have to go through all the notes used and see if any exceed the actual range, already fixed a bunch), and done a very rough mix for instruments that were excessively loud by modifying the velocity of the notes used. Mind you, this is a VERY rough mix, so this is by no means perfect.

One of the interesting things about the SNES is, given its lack of flexibility, it still managed to have a full sound, even with the limited 16 tracks that were made available. In addition, the SNES only supported 16 volume levels, as such, when you listen to this, you are going to hear some instruments fade out and it is going to sound jerky.

This track actually uses only 16 tracks, but many of the instruments are doubled, for example, there are 4 string ensemble tracks, two french horn tracks, three trumpet tracks, and two flute ensembles, so in actuality, it uses nine instruments. Obviously, its going to sound empty in places. There are virtually no dynamics. The articulations are not set (I exported using the sustain patches for the instrument tracks.

Track 1 - Harp
Tracks 2-5 - Full String Ensemble
Track 6 - Timpani
Tracks 7 & 8 - French Horn
Tracks 9 & 11 - Flute Ensemble
Track 10 - Cymbal Crash
Track 12 - Oboe Ensemble
Tracks 13-15 - Trumpet Ensemble
Track 16 - Snare Drum

I dont play video games anymore, but as a kid, I grew up with Chrono Trigger, Secret of Mana, Final Fantasy 4 & 6 (2&3 in the states), etc. The music from Final Fantasy 3/6 in particular sits with me even today, for as crappy as the SNES sounds by modern standards, it added alot of emotion to the game, which was a rarity for the SNES era (music didnt start getting cinematic as a mainstream thing until the PS2 era, to my knowledge, Squaresoft was one of the first companies to try to use cinematic, emotional motifs for their games).

Anyway, like I said, its very rough, but if you want an idea of what this might sound like fully orchestrated, here is the beginnings of the project. I will of course be adding instrumentation, fixing dynamics, and generally, beefing up this 22 minute track to be something that I think will be worthy of todays modern standards. Who knows, maybe Ill spend the time to do the other pieces as well, they are certainly short enough that it would be much less of an undertaking. I imagine, this will take several months to get through, depending on my free time...

https://drive.google.com/open?id=1p_anfRmR8wVtSJ57yZM2jW4dKWESZG4-

 

I love it! Once again am inspired by your orchestration. If you decide to do another I've always liked some stuff in Chrono Trigger

 

Its slow going, got the articulations in and did some audio leveling to try and get a better balance of sound. I will tell you, first off, this is tedious work. Second, it already sounds better than any other versions I have heard and I havent even added anything to this yet.

There are 1073 measures in this, and that isnt even accurate because the time signature changes several times, but the generated midi doesnt have all the time signature changes accurately represented. This will be something I will have to correct as I need to have the measure markers accurate when I start adding my own content.

Articulations have been added for the first 5 minutes and 35 seconds and I did a rough tweak of velocities as well to try and get a better mix. That is measures 1-256 to give you an idea (out of about 1073 measures).

My plan is to release this as a video with 3 audio tracks. The first audio track will be the original audio straight off the SNES. That is a wav file in my current project. The second track will be the original notation, but using real instruments, no notes added or changed, except in the few cases where I had to raise or lower notes or phrases to make them playable by the real instrument. I may also make a few changes to the original notation in that, as I mentioned above, they used multiple tracks for the same instruments in some cases, more on this in the next paragraph. Finally, I will add instrumentation to fill it out and make it sound like something a real orchestra would play. This may involve having different instruments playing the melodies and harmonies than the original, adding moving strings, winds, and brass lines, and in general, making it more epic, without going too epic. It is my feeling that epic in todays world is over the top, but I do think it can be much more dynamic and much more moving than what the original was.

I just noticed the time, I will have to address the multiple tracks for the same instruments at a later time. I will put updates in this thread, until it is ready for release...

 

I'm surprised and impressed that you use full string ensemble patches making them sound realistic. Also you previously mentioned using Spitfre Albion 3 ensemble which made me feel a lot better about the 600 I paid for it and never used much yet. Your knowledge and skillset are appreciated!

 

Well, I am making them as realistic as I can, which will be pretty close. When I begin adding my own touches, it will sound as realistic as I can make it, but I want to have two versions, one where nothing new is added, and another where additional instrumentation will be added, the new instrumentation will allow me to use layering to cover up the limitations in the string ensemble libraries by layering other instruments over the top of the strings at transitions, which will hide the flawed nature of strings. In addition, when I add my own instrumentation, I will separate the string ensembles into individual sections.

The key with all of this is, as I keep saying, dynamics. The attack on strings tends to be ridiculously long, so, the secret to making this work to your advantage is to have the strings swell (get louder, then softer) for each chord transition. So, when the strings hit a new chord, the attack has a natural volume intensity they will build to, then, on the release, you need to have the strings decrease gradually in volume. This way, the volume doesnt just go from soft to loud, then immediately jumping back to soft to loud, etc. The transitions will sound natural. The volume the strings stopped playing a note should be the same as the volume they start playing another note.

This is the part of MIDI that no one likes. The fiddly bits are tedious. However, if you want realism, this is what must be done. Anyway, over the course of the next week or two, Im going to try to finish the articulations, and getting the volume balanced. I may in fact swap some libraries for other libraries. Right now, Im using Vienna Symphonic Library Synchron Special Editions, but we will see if perhaps other libraries might work better, I want to get the original with nothing added, as realistic as I can get it before adding my own touches, because I really think it important to have the original notation to reference from, even as a listener.

I just got my Icon G2 control surface extender, so now I can do 16 channels at a time instead of just 8, so that will make the part where I start composing, even more fun. <3

 

Okies, just uploaded the first version, only the first 11 minutes have been modified. All of the articulations are in place, everything is balanced, but there is no dynamics yet. This means, the strings dont swell yet and the volume doesnt change much, except when new instrumentation comes in and goes out. Since the strings dont swell, they dont sound quite realistic, but they are pretty close considering they have no dynamics.

Velocities do change throughout the first 11 minutes and in a few isolated places, the volume of instruments is programmed to jump to another volume in order to maintain good volume balance, but these volume changes are subtle and you shouldnt really notice them. What I mean by dynamics is, volume doesnt gradually increase or decrease yet, and each section of the piece has not yet had the volume come up or down in relation to the rest of the piece, all I am doing on this first pass is programming articulations and getting a good balance of the instruments.

A couple of notable things here, I changed the harp to a softer harp, less harsh on the ears. The percussion actually uses drum rolls now in places where it was previously smashing a note in rapid succession. Staccatos and marcatto (crossfaded with sustains) are now in place, and, as previously mentioned, instrument volumes are now balanced so that one instrument isnt notably louder than the others. The result is actually astonishingly good, compared to the nearly untouched one I uploaded.

https://drive.google.com/open?id=1LllRbM_YaoM1l5nj_HRWmATpXWwY-CJS

 

This is cool on several levels. The original composer was able to articulate all of these parts in the 8bit realm making timeless classics, and also that you were able to deconstruct and recompile something that sounds very realistic

Are you saying dynamics are something you normally add later in the process?

I think I'm catching some of the concepts you are mentioning as well

 

Yes, dynamics are usually the last step, and there is a big reason. If you think about a music piece like a story, it ebbs and flows. Emotion and tension build, and then releases, builds, releases, over and over again, that is, unless you listen to Debussy, in that case, there is no tension and release, its all just melodramatic, monotone.

Im going to get a bit wordy here, and go into some history, and link some fairly lengthy pieces to listen to...

The amount of tension and accompanying release should and will vary as the piece moves through its various thematic elements, and it is important to recognize that the amount of tension and release are not proportional. This is to say, sometimes, there is a big buildup of tension, and only a small release, leaving tension still built up. In fact, most pieces work this way, big tension, with small releases along the way, until the tension is so big that it requires a huge release.

In many cases, a piece can build tension for tens of minutes at a time and finally release all the pent up tension in one big blast of sound. One such example of this is Ravel's Bolero, a 15 minute piece that is, in itself, a work of art, and by the end of it, if you havent gone insane from the repetitive rhythm that plays over the duration of the 15 minute piece, is a huge dramatic increase in volume, and the trombones will play a huge brass line imitating a person almost yelling, enough already at the 15m 35s mark of the piece.

The big change between the Rennaisance era and the Romantic era of music is big sweeping dynamics. Prior to this, written music did not denote varying loudness's. In the Rennaisance era, the beginnings of loud and soft were developed, but there was no way of denoting any more than loud or soft. The romantic era introduced the idea that it was possible to have something between soft and loud, and even introduced the idea of really soft and really loud. Generally, denoted by pp, p, mf, f, and ff, in order of soft to loud, (pp = pianissimo, p = piano, mf = mezzo forte, f = forte, ff = fortissimo: piano literally translated means soft, mezzo literally translated means medium, and forte literally translated means loud, these are all Italian words). Later, another symbol was developed called a crescendo and decrescendo that indicated a gradual increase or decrease in volume.

In the beginnings of the romantic era, the idea of building tension was a new concept, and some composers who were on the bleeding edge of this idea, wrote music that did not effectively build tension before releasing it, and the result was, well, listen for yourself. Haydn wrote a piece called Surprise, and Ill bet you cant guess what the piece does...

Then, we have the brilliant composer Beethoven who was before his time, he liked to do things ass backwards. In one of his most known works, Beethoven's 5th, he begins immediately with tension, then releases.

All of this is to say, what we learned throughout history is, music should be full of buildup and releases and the primary ways we do that is with chord structure, and dynamics.

The answer to your question is this, the reason I wait until the end to put in dynamics is, everything must be proportional to the piece. In the FF6 ending, if you are familiar with it, you'll recognize that the ending part of the piece should be epic, it should blast you out of your seat, it should end with you saying, YES!!! However, what about other sections in the middle of the piece? Should Celes theme be as loud as the ending? Probably not, but it certainly should build in intensity, drawing out emotion, the strings should soar. Terra's theme on the other hand should remain soft and tender throughout. The Zozo theme should be louder, it should almost feel chaotic, possibly louder than Celes's theme, but definitely not as loud as the ending. The Final Fantasy theme should be loud, as it portrays confidence, the listener, if they have followed the series, will be very familiar with the theme, I would even say, it should be louder than Celes's theme.

The general workflow is this, you start with the small details, get those right, then work your way out to the bigger picture. Much like building a house, the first thing I am interested in is the foundation and the frame, focus on all the little details. Once the structure is built, now we paint (literally) the broad strokes, we do the landscaping, all the stuff that makes the house look beautiful. Without all the little details however, no amount of dynamics in a piece will fix the fundamental and inherent design of the music.

In the case of MIDI, there is another special unique challenge, as the volume of one instrument is adjusted, it may require adjusting the volume of every other instrument for that passage, which will result in the overall volume of that passage being louder or softer, which can really screw with how loud that passage is in relation to the rest of the music, thus, always start by placing articulations, then sound balancing, then doing overall dynamics. You start with articulations because, certain articulations are louder than others. For example, Staccato, Italian for disconnected, or unattached, is typically played louder than a sustain articulation. This means if you want a note to be played short, you will use staccato, but it will be louder than the adjacent sustain notes, and as such, you will have to adjust the volume for the staccato notes to be softer if you want them to have the end result of being the same volume.

Hopefully, I didnt bore you, I find the history of why we do what we do just as important as the music itself, the greats started it all, made all the mistakes, we can learn from those mistakes, and learn from what they did right, and hopefully, evolve music forward into a more perfect listening experience.

 

i need midi for cinematic and orchestral stuff. i have inspire 2 but i am not able to make shit. i think you can help me

 

I always said a lot of things went soft during the romantic era...

This is similar to your previous notes which were super helpful. I've been working on keeping things smoother as well as building dynamics, something non-existent previously here so thank you again

Ah yes, my master taught me this with the paint-the-picket-fense technique once.

Definately not boring, I sat motionless outside the Shaolin temple for days for less then this..
Your notes have actually deconstructed my whole process in a good way!

 

Ha! Indeed ;)

The key is, never give up, never surrender! I lied, the second key is, listen to the greats and emulate them. Very few ideas are new, but creative people know how to take multiple ideas, pick what they want, and put them together into a cohesive thought.

Really spend time on the little details. Its the most tedious and boring aspect of what I do, but, it also gives the biggest results in the end.

John Williams and Hans Zimmer havent created anything new. In fact, they even stole musical lines from other composers. What they did do is take ideas that previously existed and built on them to create something even more grand.

I try everything before I ever make the decision to mask a sound with another sound to cover the flaws in a library. My goal is to solo an instrument and play just that instrument and have it sound natural by itself. If it sounds natural, the whole piece will sound natural.

Velocity crossfade, dynamics, articulations, loading a completely different library for just 2 notes because I just cant get those two notes to sound right, or even moving those two notes and having another instrument play them, those are all things in my normal attempt at realism. Now, in the case of this project, I am limiting my options, I am not adding any notes, or changing the instrumentation in any way, so having another instrument play those two notes is not an option, but generally speaking, its an option you can try.

If you really want to learn the ins and outs of MIDI, grab a midi track for a well known song. Id recommend a shorter piece, and try to make it sound as realistic as possible, with the artificial limitation that you cannot change any notes in the existing MIDI (provided the MIDI doesnt have any wrong notes or blatant errors) and you cannot change or add any instrumentation. Learn how velocity crossfade works, learn how to use the curve tool in your DAW of choice to create natural crescendos and decrescendos. Learn how the library you have sounds and works, become intimately familiar with it. Learn what articulations are, and creatively implement them. Marcatto, Spicatto, two lesser used articulations, but are very powerful used effectively. Using velocity crossfade, you can combine them, or other articulations and really make things sound natural. Musicians rarely use one articulation completely by itself, instead, they interpret the piece and combine multiple playing styles for each note to match the style of the music. Even in an orchestra, each violin player will play a staccato note (short, often loud) very differently from the player sitting next to them.