Overview
|
Overview
|
Seeds etc | User guide | Main Window | Musical note intervals | Scales | Midi in | Analyse sound |
Layers - Parts - Terminology - Arpeggios and Scales
These are layers of ornamentation of the fractal tune. It's built up one layer at a time.
You can read about this in How the seeds build up to make tunes, Here the focus is on a short tutorial so you can make seeds with layers yourself and find out how they work in a practical fashion.
First, though, an important point to bear in mind before we set out on the exposition. Layering is a melody structure concept, and Parts are to do with the instrumentation. I'll do these examples with one part for each layer as that makes it easier to hear the layer structure in the tunes, but the parts aren't inextricably linked with the layers by any means. Many of the fractal tunes distribute the parts amongst the layers in a far more complex fashion.
Let's make a simple layered fractal tune by way of an example. We'll do it one step at a time so that the structure is easy to hear.
Choose File | New , to make the first tune you hear in FTS.
Go to Views | Composer
Now, set the Seed Layers to 1, unselect Cycle tune , and play the tune, and you will hear the basic seed played once.
Note - if your fractal tunes sound quieter than expected, take a look at the trouble shooting section - Why is everything so quiet?
So, let's now increase the Seed Layers to 2. You will hear a more complex tune. Actually it consists of the original seed, played more slowly, and with a copy of it added to each of its notes as a kind of ornament, to make a more complex melody.
To hear this, first be sure to select Parts... | Order of Play | By layer with simultaneous notes . Then increase the Parts for fractal tune from 1 to 2.
You will hear two flutes playing, one playing the original short seed, and the other, faster, playing the version with the ornamentations. This may not be so easy to hear, especially since they are both playing at the same pitch.
So, to make the listening easier, let's change the second voice. You do this from the Parts window. Highlight a part - in this case, the second part, and then select the desired voice from the Voices men, say, Orchestral Harp . Or alternatively, type the desired instrument into the text field below.
Now you'll hear the Orchestral Harp playing the original seed, and the flute playing the ornamentations on it. You could also at this point drop the harp down an octave, which you can also do from the Parts window - use the text field, or the up / down arrows at the bottom of the Octave shift column. Perhaps now you can hear rather easily that the Orchestral harp is playing the original seed, only slower. If at any time in this tutorial you find that you can't hear the parts distinctly, try muting all the other parts by setting the volume to zero. Then once you hear the part you are interested in, reset the others to their original volume and see if you can still hear it.
Try increasing the Layers to 3 now, and you will find the Orchestral Harp plays the original seed with its ornamentation - in fact the same tune that the flute played earlier. You will also find that the flute has added yet another layer of ornamentation to it. Finally, we no longer hear the original seed below it all. To get the basic seed into our tune, add another part by increasing the Parts for fractal tune to 3. Let's choose another voice for this, say, a Contrabass , and lets transpose it down a couple of octaves .
The reason for selecting Parts... | Order of Play | By layer with simultaneous notes selected, or indeed, alternatively, Parts... | Order of Play | By layer , is that with that option, the first part plays the tune with the most layers of ornamentation, part 2 plays it with one less layer of ornamentation, and so on, making the structure easier to hear.
If you select other options in this list, then the notes for any of the layers can be played in any of the parts, so the structure isn't so apparent. Once you are familiar with the structure of the tune, you can try experimenting with the other options here to see what they do too, and then go back to By layer with simultaneous notes to compare. The basic melody line stays the same with all the options - the actual notes to play. Some of the notes get transposed up or down by octaves - that's all - but because of changes in the instrumentation you hear melodies crossing back and forth between the layers giving quite complex tapestries of sound at times.
Now, we noticed that the harp played the same tune as the flute in the previous stage. What you mightn't have noticed yet is that it always works like this with the fractal tunes - at least the ones based on the original idea for FTS. Apart from changes in instrumentation as mentioned, the faster tune each time is exactly the same as the tune in the slower layers played faster
In fact, if you have By layer with simultaneous notes selected so that the instruments pick out the layers, the fractal tune is a canon by augmentation - a canon in which each part plays the tune a bit slower than the one before it. It's a strict canon - though the slowest part in our example here only plays the first three notes of the tune. There are more to come if we increase the number of layers.
One thing one might notice at this point is that at first, the tune goes up and down by whole tones, as does the original seed. But then, at the highest point in the tune, right in the middle of our short piece, it goes up by a larger step of a minor third instead of a tone.
Why is this - well the seed shows the number of steps to go up and down in the arpeggio. The arpeggio in turn is:
Pentatonic (with note names shown for 1/1 = concert pitch C)
0 1 2 3 4 5 tone tone minor third tone minor third C D E G A C
The seed in its original position is 0 1 0 - and a single step at this point is a whole tone. Next, when it gets shifted up one step in the arpeggio to 1 2 1 then it still goes up and down by a whole tone. However, at the next shift up it plays 2 3 2 which goes up and down by a minor third (e.g.. from E to G):
To experiment with this idea, try another seed with this same combination of instruments, whatever you like, let's say, just picking something at random, 0 4 -3 -2 0. It helps if it starts at 0 for these first experiments. For more about how these numbers work, see the How the seeds build up to make tunes, and the introduction to Seeds.
Notice that the Contrabass plays your entire seed just once, and in the other parts, notice how the seed gets transformed as it moves up and down the arpeggio. Try a few other seeds, and try increasing and decreasing the numbers of parts until you are well familiar with the idea of how it works.
Now try upping the Seed Layers to some large number, such as 50 (the maximum), and play the tune, and it will just keep going on and on. Actually it will stop eventually, in principle, but look at the Tune Info window, and you will see that with this number of layers the tune will keep going for some vast number of days - or all practical purposes, for ever.
The tune may eventually begin to get out of range of the instruments, indeed, possibly quite soon if the numbers in the seed are large. At that point one might want to explore Parts... | More... | Ranges... The most important section in this window is the Range - here you can set the lowest and highest note for each part. You can then set the tune to bounce back into range when it goes beyond the range of the instrument. Of course it will then no longer be a strict canon- but it will keep playing at least and these bouncing melodies can also be very attractive..
Note that if you now set the number of Layers back to 2, the Contrabass (or whatever you have for the third part) will stop sounding even though you have three parts still in play. Since we have Parts | Order of Play | By layer with simultaneous notes selected, the third part is set to play the third layer, and because we now have only two layers, there isn't anything for it to play. I mention this because sometimes it can cause confusion, if the number of layers is low and one wonders why one of the parts isn't sounding. If you try other selections from the Order of Play menu at this point so that each part doesn't need to play only one layer at a time, you will hear it join in again from time to time.
As you may realise from this tutorial - harmonically the fractal tunes are really simple in concept as they just use parts in unison or at octaves. Technically, everything that happens consists of passing tones, you could say, of parts moving in and out of unison or octave equivalence. I'm not sure why the tunes should sound so harmonically rich and complex with such a structure; it's just the way it turned out... The tunes also stay within a particular tonality all the time, most of them. Some exceptions there such as major chords round the cycle of fifths - Bach temperament
Shortcut tip : Note that you can change the voice, octave shift, etc for any number of parts simultaneously in the Parts window and other windows of the same type. You do that by highlighting several parts simultaneously.
The way to do this is to click on the first entry you want to highlight, then use Shift + click on the last entry - or alternatively, use Shift + down arrow to extend the highlight. It's Ctrl + click to add another line to the highlight without changing the ones you have highlighted already, or Ctrl + click then Ctrl + Shift + click to add a section of several rows to the highlight in one go.
Now change the value in the usual way, and you will find that they are changed in all the highlighted parts simultaneously. For instance, if you select a new voice, it gets selected into all the highlighted parts. Similarly if you change the octave shift or the volume.
As you will see in the Parts window, you use these to choose what instruments to play and what volume to play them at, and various other parameters relating to your instrument(s). In twelve equal programs you would normally use channels here - you would select an instrument to play in a particular channel, and set its volume and so forth. However, because of the retuning requirements in FTS, normally we can't use the midi channels in quite this fashion, at least not directly. Instead we work with them through Parts as an intermediary.
That's because a single Part selected in the Parts window may need to be played on several midi out channels at once. This help explains the reason for this, and also introduces the notion of a part and how it is used in FTS.
In practice, the parts function pretty much like channels if you are familiar with those already, indeed with some extra features that the original Midi channels don't have as well (such as the ability to play a different tuning in each Part should you so desire)..
To get to other related sections of the midi relaying help click on the links above. To find out how they get used in the fractal tunes see Layers (earlier on this page) or Making Seeds Arpeggios and Scales.
If you want to know a bit more about the background to this idea of retuning to a new tuning using midi, see Microtonal music in MIDI
Playing the notes - Other ways to retune in Midi - Relaying from Midi In - Relaying from a keyboard via midi in - Are there enough Midi out channels for the retuning?
This is a diagram to help show how things work in FTS - for those who like diagrams :-).
FTS part ® Several Midi Out channels depending on the pitch bend needed
You want FTS to be able to play your part on several channels because each midi out channel has to have only one pitch bend in it at a time. That's because if you use the pitch bend wheel on a music keyboard, it bends all the notes in play in that channel at once.
To take an example: the notes for a just intonation major chord 1/1 5/4 3/2 all need different pitch bends, so need to go to three separate channels. FTS might play those on channels 1, 2 and 3 - though the numbers for the three channels it uses depend on what has happened before.
To get a feel for how this works, try any of the fractal tunes that use scales other than twelve tone equal, and show Out | Notes In play . You will see the notes in play for each part, the channels they are relayed to, and the current pitch bend in cents for each channel.
You can also hear the reason FTS does this for yourself - try unselecting Out | Options | Okay to change channels for pitch bends . FTS will now play each part in the same numbered midi out channel, for example, part 1 will be played in Midi channel 1. Now, try to play a major chord as 1/1 5/4 3/2, maybe from the PC keyboard. FTS will need to clip the notes, and only one note will play at a time because of this restriction to one pitch bend in play at a time in the channel.
FTS also changes the pitch bends for the channels as seldom as possible. That's because some notes such as harp may continue to resonate for a fair while after the note off and this lets them resonate for as long as possible.
The other usual approach is to retune all the midi notes using a tuning table. There is a midi standard for this - the MTS sysex - which is implemented in a few synths and some of the new soft synths and FTS can use this. See MTS tuning programs. Many hardware synths however use custom sysexes to change the tuning tables, and FTS doesn't have any option to work with those as there are so many of them. Use SCALA for these instead as it can handle most of them, as one of its specialities.
The rest of this section is by way of a musing about possibilities for the future. I don't know of any implementation of it, though it is in the Midi spec. So for practical matters, skip on to the next section Relaying from Midi In.
Actually, the midi spec also includes an option to add pitch polyphony (which would make FTS's work much easier), but it hasn't had wide acceptance so one can't use it. The idea here is that pitch bends don't have to affect notes already in play as they do now. What if the bend were to only affect new notes?
That would be great for what we do here - except for one drawback - since the pitch bend wheel will only affect the pitch of a note when you first sound it, it can have no effect on notes already in play. This isn't much use for improvised pitch shifting - you don't hear the pitch for the note until after you change the pitch bend wheel position - and then at that point, once it is played, it is set in its current position and whatever pitch it has, you are stuck with that, and can't change it any more.
So, either you can bend notes using the pitch bend wheel, or you can have unlimited pitch polyphony, but it seems, not both. Everyone seems to have opted for the idea of bending notes already in play when they design a synth or soft synth - which is no great surprise as it is the natural choice indeed.
One could imagine the ultimate in flexibility here would be if one had the unlimited pitch polyphony implemented, but with a simple switch on the keyboard which would let the musician switch between pitch polyphony and pitch bending in the usual fashion, as desired (this setting would also be changeable using a sysex or maybe one of the Midi controllers). The pitch bending could then apply to notes already in play just as it does already in midi, but as an extra pitch bend on top of the current tuning of the note. So you could sound notes by one system, and then flick the switch and bend them from their existing pitches using the other system. Then if you wanted to keep your new pitch as a fixed pitch in the scale, flick the switch back again while the note is still bent, to return to pitch polyphony.
If anyone were to implement such a system it would make the task of FTS and other retuning programs far easier, and I think it would be acceptable within the Midi spec... (I'm not sure if it specifies what behaviour is expected when you change the system in the middle of a note which is still sounding - the one suggested here seems a reasonable idea for it but there may well be other ways that will be better).
If you are midi relaying, or playing from a midi keyboard then the situation is like this:
Midi In channel ® FTS part ® Several Midi Out channels depending on the pitch bend needed
Now the question is - how do the parts played relate to the midi in channels? There are two options here.
In | Options | Select part by input chann , FTS plays each midi in channel in the same numbered part. Midi In channel 1 is played as Part 1. This is useful if FTS is used to retune a score from notation software. FTS will be able to relay each midi in channel to a separate part, which is as desired, because usually a score is set up using several parts, each of which will be assigned its own instrument and other settings and its own midi channel.
The notes for these parts get relayed to any of the channels from FTS because of the requirements of a new channel for each new pitch bend - but they will sound as intended, played on the desired instruments and with the desired settings for controllers, pitch glides, and so forth.
So in this case it is:
Midi In channel ® Same numbered FTS part ® Several Midi Out channels depending on the pitch bend needed
This is also the best setting for View | Retuning midi player , and any other situations where several midi in channels are used.
You will notice that the Parts window has part 10 set to non melodic percussion when you use the standard settings for a GM synth or sound card - and that is also to help with this situation - because when relaying from midi you will have part 10 set to relay the midi in channel 10, and in General Midi, which is usually used by notation programs to play the scores, channel 10 is set to be the non melodic percussion channel. This means that non melodic percussion in notation programs gets played as intended in FTS, also the same applies for GM midi files played through the retuning midi player.
The other way FTS is used though is for playing or improvisation from a keyboard. It is possible to use several keyboards as manuals, in which case one would probably still use In | Options | Select part by input chann .
However, if you wish to use FTS to retune a single keyboard at a time then it is useful to be able to divide the keyboard into regions, each with its own voice - indeed, possibly each with its own scale and arpeggio too In this case you would use
In | Options | Select part by keyboard regions . which is the standard setting for View | Midi Keyboard retuning . There are various preset keyboard regions in In | Options | Keyboard options | Preset regions - am d for further refinements, one uses In | Options | Kbd regions .
One way to start exploring some of these possibilities is to try the Presets for the Midi keyboard. Many of these use keyboard regions. For instance, several of them use two keyboard regions, for the left and right halves of the keyboard.
Then you might have:
Midi In keyboard region ® FTS part 1 or 2 depending on whether note is in left or right half of keyboard ® Several Midi out channels depending on the pitch bend
So, the general situation here is:
Midi In keyboard region ® Several FTS parts depending on the region ® Several Midi out channels depending on the pitch bend
To deal with all these various possibilities, we need an intermediate notion of some sort. We can't call it a channel because it mightn't be the same as either the midi in or the midi out channel. I use the word part because it is somewhat like the part of an arrangement, and so far, it seems the closest analogy I can find, and the most likely to be helpful for understanding the concept.
There are only 16 MIDI channels, or 15 melodic ones (in GM mode). At first sight this doesn't seem very much perhaps. However one is less likely to run out of parts than one might think as a modern soundcard will normally be able to play several different voices simultaneously in the same MIDI channel, and most scales have less than 15 notes in them. Even if the scale has more than 15 notes, one will seldom want more than 15 of these played simultaneously - bearing in mind that any notes at octave intervals can be played in the same channel. This means you can just set up one midi channel for each scale degree and send all its notes on that channel, whatever part it originates from - so you may never run out of channels at all in this situation
You can run out of channels for the notes more easily if your soundcard or synth is restricted to one voice per channel at a time - this is fairly common for a soft synth or synth. It can also happen if you vary the effects from one voice to another including the pan position left or right - as that is set for a channel as a whole. It can also happen if you use large scales, or non octave scales with your dense chord clusters.
For more about this, see Situations where one can run out of channels in the Musical intervals page.
This section is for those who wonder about this terminology, or feel they want the reasons explained clearly.
Q Why is a distinction made between arpeggios and scales - why not just have scales?
A. If you only have scales, then you will find that you have many ways of retuning the same scale, e.g. a major scale. You will then have to provide separate tunings for each one - e.g. a Quarter Comma Meantone scale on C with the wolf at G# to Ab etc. This is completely impractical as there are hundreds of twelve tone tuning systems, and hundreds of arpeggios listed for each, each of which could start on any of the twelve notes of the tuning (except of course in twelve equal where all positions will sound identical)..
Then, perhaps some may find the particular use of the word Scale a bit puzzling. Musicians often talk about playing a piece in, say, the scale of D minor, meaning, the key of D minor. This doesn't tell us whether it is in twelve equal, or quarter comma meantone, or in Werckmeister III - the terminology ignores tuning distinctions. However, in this field of the study of tuning systems, it is usual to use the word Scale to refer to the precise tuning used. Discussions, literature and other programs use the word in this sense, and that is how we will use it here. There doesn't seem to be any reasonable alternative (Tuning is too broad in its meaning), nor it seems is there much risk of confusion, as I don't know of anyone who has got confused as a result of this use of the word. So, we can reserve the word key for the other usage as I did just now.
As for Arpeggio - these are usually referred to as Modes in this field - and here it seems there is somewhat more possibility of confusion because the more common use of the word Mode is to refer to a rotation of the scale, as in Dorian mode or Lydian mode etc which are all just the ordinary major scale - except that you change which note you think of as the home note of the key. So for example, you can talk about the Dorian mode without any reference to any background master scale - just say that you start the major scale from D instead of C and that gives you the Dorian mode.
The problem is that Mode as used in modal music is so similar to the intended meaning as to be potentially confusing - familiar - but not with the exactly the same meaning you expect it to have.
Here, you see, arpeggio always has to mean steps in an underlying scale. E.g.the major scale as:
0 2 3 5 7 9 10 12
in your underlying twelve tone scale . Calling that a mode blurs the distinction between the modes as in Dorian and Lydian mode etc and the idea we are using here where you can just select any notes from the twelve tone scale, indeed in any order too. They don't have to be rotations of major scales like the traditional modes of modal music. So a musician with that background is likely to have a too restricted view of what the arpeggio could mean if one were to call them modes instead.
In FTS they get used as freely as arpeggios - they needn't ascend all the time but can descend for a while in the middle and ascend again, should you wish, as is sometimes useful in the fractal tunes. They can have negative numbers at any point (to go below the 1/1) also repeated notes, and you can make an arpeggio that ascends one way and descends another way.
So, maybe the way to think of it is that Arpeggio includes Mode in its tuning sense as one of the meanings, but is extended beyond the normal usage even in tuning circles. Tuning specialists would have no quibbles about a mode that ascends one way and descends another since that is common in Indian or Arabic music, but you might get a few raised eyebrows perhaps if you start calling it a mode with repeating notes and negative numbers...
So, in FTS I use Arpeggio as a more general term which any musician will understand with the intended meaning immediately, and with the intended generality too. In practice, since I have changed from Mode to Arpeggio, I have found that users seem to understand well what it means - though some indeed don't like the word - but no better alternative has been suggested to date.
Unfortunately there just doesn't seem to be any widely understood word that exactly conveys the meaning, and Modes is too specialised - that would be okay except for the double meaning which does cause some confusion to musicians who haven't specialised in tunings. Also there is no alternative replacement word for the more common meaning of Modes, as there is with Scale and Key. But, if some better word does come up for Arpeggio then I may change it.