Sytrus Tech Study Intro – Operators

Sytrus Tech Study Intro – Operators

This Post has taken me most of the week to do, as there is a lot of info here, but it is my last for while, so I made it a Biggy, as I’m moving house and have painting etc. to do, but should be able to do more quite soon as there should be no break in the Internet connection anyway.

Ok, after looking at Harmless, I feel a little more prepared to delve into Sytrus, at a basic tech level, as I was a bit overawed initially. At least I feel I can get some of the simpler, more obvious stuff logged for now, and just slowly chip away at it over time, as there is lots of functionality in this amazing Plugin. If you buy this at 160 Euros or so, it represents a big investment, do you need to be able use it the best you can. Like everything in FLS, it just takes time, perseverance and patience to get to know something…and there is a basic Sytrus Tutorial in the Help Index, but not much online by image-line themselves that I could find.

The whole design is based around a matrix, X-Y grid structure, particularly for the 6 operators (oscillators) control knobs structure on the right side, for example the, OP1 tab volume out knob is the only active sound for the Default patch, on the right side, set at 100% for the default sine wave, you can see playing in Wave Candy:

and hovering over the grid reference knob at 1:1 shows this knob is Op1 FM feedback control. The same control for OP2 is at grid ref 2:2 and so on, but these are currently dimmed, so inactive:

You can de-activate knobs with an Rclick. If done on the OUT knob for OP1 above, it stops it being heard.

I’m going to spend some time in OP1 as it is important to understand the fundamental waveforms that all else is built on, and as there are 6 identical operators, knowing 1 well is knowing all well.

In any OP, major fundamental changes to base waveform can be made by selecting one, some or all of the shape modifier buttons above the sine wave screen:

This half sine sounds like:

halfsine.mp3

Slider Section – Manual:

———————————————-

Shape Modifiers

The modifiers are a set of parameters which can help you in adjusting the operator shape to suit your needs. Alt+click a modifier to reset it to its default level:

  • Shape (SH) – Allows to morph the shape to sine, triangle and pulse (via the other modifiers you can also achieve saw and other shapes).
  • Tension (TN) – Sets the tension (distortion) of the base operator shape.
  • Skew (SK) – Allows you to “skew” the base operator shape.
  • Sine Shaper (SN) – Applies a transform which is useful for certain patch types (such as bells).
  • Pre-Filter (FL) – Applies a smoothing (low-pass) filter to the shape. This can reduce potential aliasing artifacts but makes the sound less bright.
  • Noise (NS) – Adds a real-time noise to the filter shape. This can be useful with plucked mode (which requires rich spectrum) and drum/percussion samples.

——————————————–

You can further slowly modify this to a Saw wave via a Triangle, with the first slider:

Sounds harsher as we know:

saw.mp3

Be careful with this slider – if you continue to 100% you will get a flat topped pulse, so other than a click – silence:

Do this with the default sine wave to make a square wave and familiarise yourself with its harsher tone:

square.mp3

This is a stereo view I thought it might be good to know about later at FX addition to see right-left shape and phase differences.

The Tension slider, next along, can alter the shape further, to squared pulses:

The Skew slider deforms the waveform slantwise, bending leading edges:

To have a better look at the Sine Shaper slider effect, I Default to the Sine, then add a little:

With these last two waves, you can start to hear higher harmonics present.

To try and understand the Prefilter slider, (which seems contrary to visual logic at first, as it seems to “tidy” a waveform), you need accept that square waves are harmonically rich compared to sines, as it takes many sines added mathematically to create a square wave in theory.

http://en.wikipedia.org/wiki/Square_wave

Note that the square wave contains only odd-integer harmonic frequencies (of the form 2?(2k-1)f), in contrast to the sawtooth wave and real-world signals, which contain all integer harmonics.

There is a helpful graphic to help picture this here:

http://mathworld.wolfram.com/FourierSeriesSquareWave.html

FourierSeriesSquareWave.gif
Remember the odd-even harmonic slider section in the Harmless Post? The above implies using those odd integer harmonics sliders to create a more square-wave sound.

This slider straightens small ripples by – I assume – adding higher harmonics of the fundamental, as you increase the value, making anything other than a pure sine, look neater. E.g. from:

To:

BUT, anything with a horizontal component (a partial square) is harmonically rich, so you see 49 harmonics in the Hint bar at only 28% slide. This goes to 16000+ harmonics at 100% – which of course you can’t hear separately – but is more of a technical interest than a practical one, as I’m trying to understand the degree to which the programmers of FLS have gone to include such detail.

You can hear the first few added, as few as 6 harmonics, but the changes soon become merged.

The last slider in the group just adds noise – a common use for this is in Dance tension builders:

http://www.youtube.com/watch?v=x4moxxf_sI0&feature=relmfu

It’s really worth playing around with these basic shapes for a while to get a feel for the sounds they produce. I chopped up a note and used the Riff machine on this shape to get a cheap PC game type sound:

It sounds like:

PCGame.mp3

Even another use for the Noise slider to add urgency:

UrgentNoise.mp3

Turning on the Plucked button affects the attack so it is more percussive and removes the noise element – explained later.

A very basic but important control to jump to now is the frequency ratio number set at 2, as multiples of this alter the octave of a note, doubling by 2,4,8,16….a 0 set here gives no waveform, and a 1 drops an octave, 0.5 is -2 octaves etc.

You need to know how this relates to Concert Pitch, defined as note A above middle C4 on a Piano, vibrating at 440Hz. When this ratio is at 1, playing this A note will be 440Hz. It shows as the note above C5 on the Channels settings keyboard now the Sytrus default is 1 not 2.

A440.mp3

This sounds as the example on the WikiP page here:

http://en.wikipedia.org/wiki/Concert_pitch

As Concert Pitch definition has changed at times, A = 443HZ for example – there is a frequency offset box also, with 1/100th Hz sensitivity. I experimented with changing this and could barely detect a 7Hz offset by ear from 440Hz, depending on how fast the change occurs also.

It may be helpful at this point to reference some concepts on absolute note/pitch relationships.

I came across 1200 “cents” terms in the Harmless Post (being equal to 12 semitones, where 100 cents or 21/12 = 1.0594631 semitones). So, blatantly copied from WikiP here:

http://en.wikipedia.org/wiki/Scientific_pitch_notation

Table of note frequencies

Frequency in hertz (semitones above or below middle C)
Octave ?
Note ?

0

1

2

3

4

5

6

7

8

9

10

C16.352 (?48)32.703 (?36)65.406 (?24)130.81 (?12)261.63 (±0)523.25 (+12)1046.5 (+24)2093.0 (+36)4186.0 (+48)8372.0 (+60)16744.0 (+72)
C?/D?17.324 (?47)34.648 (?35)69.296 (?23)138.59 (?11)277.18 (+1)554.37 (+13)1108.7 (+25)2217.5 (+37)4434.9 (+49)8869.8 (+61)17739.7 (+73)
D18.354 (?46)36.708 (?34)73.416 (?22)146.83 (?10)293.66 (+2)587.33 (+14)1174.7 (+26)2349.3 (+38)4698.6 (+50)9397.3 (+62)18794.5 (+74)
E?/D?19.445 (?45)38.891 (?33)77.782 (?21)155.56 (?9)311.13 (+3)622.25 (+15)1244.5 (+27)2489.0 (+39)4978.0 (+51)9956.1 (+63)19912.1 (+75)
E20.602 (?44)41.203 (?32)82.407 (?20)164.81 (?8)329.63 (+4)659.26 (+16)1318.5 (+28)2637.0 (+40)5274.0 (+52)10548.1 (+64)21096.2 (+76)
F21.827 (?43)43.654 (?31)87.307 (?19)174.61 (?7)349.23 (+5)698.46 (+17)1396.9 (+29)2793.8 (+41)5587.7 (+53)11175.3 (+65)22350.6 (+77)
F?/G?23.125 (?42)46.249 (?30)92.499 (?18)185.00 (?6)369.99 (+6)739.99 (+18)1480.0 (+30)2960.0 (+42)5919.9 (+54)11839.8 (+66)23679.6 (+78)
G24.500 (?41)48.999 (?29)97.999 (?17)196.00 (?5)392.00 (+7)783.99 (+19)1568.0 (+31)3136.0 (+43)6271.9 (+55)12543.9 (+67)25087.7 (+79)
A?/G?25.957 (?40)51.913 (?28)103.83 (?16)207.65 (?4)415.30 (+8)830.61 (+20)1661.2 (+32)3322.4 (+44)6644.9 (+56)13289.8 (+68)26579.5 (+80)
A27.500 (?39)55.000 (?27)110.00 (?15)220.00 (?3)440.00 (+9)880.00 (+21)1760.0 (+33)3520.0 (+45)7040.0 (+57)14080.0 (+69)28160.0 (+81)
B?/A?29.135 (?38)58.270 (?26)116.54 (?14)233.08 (?2)466.16 (+10)932.33 (+22)1864.7 (+34)3729.3 (+46)7458.6 (+58)14917.2 (+70)29834.5 (+82)
B30.868 (?37)61.735 (?25)123.47 (?13)246.94 (?1)493.88 (+11)987.77 (+23)1975.5 (+35)3951.1 (+47)7902.1 (+59)15804.3 (+71)31608.5 (+83)

 

So, A4 = 440Hz = C4 + 9 semitones, where C = 261.63Hz – nasty numbers.

It must be understood that there a many arbitrary systems and tables for note relationships for historical reasons, music having been around a lot longer than science for one, so the current Concert Pitch does not relate to Scientific Pitch, which makes C4 = 256Hz, so that all the octaves of C match perfectly with binary maths (256 = 2^8), and 256 is divisible by 2 right down to 1Hz, which keeps numbers nicely out of the fractional realm.

Because human sound perception is subjective, psychological and non-linear (volume perception is base 10 logarithmic – so “twice as loud” is pretty meaningless at best). No human related absolute numbering system is possible because of this subjectivity.

Given that the average human ear frequency response is given as (very) roughly 20Hz to 20,000Hz, the lowest note on any instrument one could hear from the above chart is C0 for an individual with “normal” hearing – though many people would hear frequencies lower than this – a car idling at 900 RPM = 900/60secs = 15Hz for example – most people can hear that. C-1 would be only 8Hz or so, but most people wouldn’t hear that, so C0 at 16.352Hz it is. This makes sense as:

———————————

http://en.wikipedia.org/wiki/Psychoacoustics

The lowest frequency that has been identified as a musical tone is 12 Hz under ideal laboratory conditions

—————————–

The lowest I can hear with my laptop Sennheiser headphones at full volume, is C0, as it is the minimum on the Channel settings keyboard anyway, for this Sytrus default sine wave, at 0dB output – more of a moth flap than a musical note.

This is interesting having the FLS Channel keyboard going from C0 to almost C10 – I am completely deaf to A9 in my left ear, but not A9flat or A9# – weird – must be my tinnitus in that ear?

Be careful you don’t deafen yourself if you try this yourselves – it’s very piercing that high.

The buttons next to the sliders top to bottom – Manual

—————————————————

Misc Settings

  • Center – A DC offset removal.
  • Declick – Applies a special filter to avoid clicks when starting a voice (useful for operators with sharp attack).
  • B.Limit – Band limits the waveform.
  • Pluck – Plucked mode turns the operator into a plucked string simulation unit. The unit basically starts with the original spectrum as set by all harmonics and modifiers then a gradual “damping” occurs which filters that spectrum to a sine wave with the pitch value of the operator. For this mode to work it is recommended to start with a spectrum rich operator (if you start with a simple sine there is no audible effect). The damping amount is controlled via the DAMP articulation target (see articulation below). Note that an operator in plucked mode cannot be frequency modulated (FM) by other operators, so its FM row in the matrix will be disabled while plucked mode is on (however ring modulation /RM/ is still available).
  • Phase (PHS) – Sets the initial phase offset for the operator when a voice is started (or global offset, depends on the next option).
  • Global – Enable this option to use a “global” phase for the operator in all voices (i.e. sync the phase between all voices for the operator).
  • Volume (VOL) – Sets the operator volume. The modulation levels in the matrix cannot be automated, however this volume property has the same effect for modulators and can be automated.
  • Pitch Envelope/LFO Amount (PE) – This property sets the range the pitch articulators operate with (the default value is 12 semitones).

—————————————————

More depth on some of the above:

DC-Offset:

Technical aspect regarding average value of an AC signal – if you are really interested:

http://en.wikipedia.org/wiki/DC_bias

There are general guidelines for Patch editing here:

http://www.image-line.com/support/FLHelp/html/plugins/Sytrus_pres.htm

Band limit waveform:

http://www.slack.net/~ant/bl-synth/5.band-limited.html

A band-limited waveform is one whose harmonics are kept below a particular frequency. If a waveform is band-limited to the Nyquist frequency and then sampled, no aliasing will occur.”

Why Aliasing/Nyquist Limit is important:

http://www.slack.net/~ant/bl-synth/3.nyquist.html

The simplest sound wave is an oscillation between two amplitudes. A sampled waveform thus needs at least two sample points per cycle. Put another way, the wave’s frequency must not be above half the sampling frequency. This limit is called the Nyquist limit of a given sampling frequency.”

If a sine wave higher than the Nyquist frequency is sampled, a sine wave of lower frequency results. This effect is called aliasing.”

This would result in a completely wrong frequency being reproduced by sampling – for those frequencies above the sampling rate Nyquist limit. This is the main reason why CDs are sampled at 44.1kHz – over twice the human ear audio range of 20Hz-20kHz, so a minimum of only 2 samples taken, as explained above, available for the highest frequency of 20kHz.

As an ex IT/ Telecomms Tech, I really want to mention something else interesting relating to human perception and sampling rates here, about the digital telephone system. If we sample CDs to 44.1kHz for quality, why do we not expect TV/Audio quality speech down our phone line? Why is digital telephony of such lower sound quality than TV? Well, it’s for historical necessity/technical/financial reasons mainly.

Earlier copper line based systems nor associated electronics could not sample sound data, transmit audio packet data, then re-integrate it fast enough in real time, over long distances, without huge expense, until relatively recently, for a full 20Hz to 20kHz bandwidth spectrum to be viable en masse. Only with optical fibre or cable TV systems right to the door is it now possible for telephone/video conferences at CD sound quality possible, though broad band speeds over copper wire are getting faster all the time. 3Mb/s line speeds give adequate YouTube video experience for pre-recorded A/V material, and some real time TV without much buffering problems.

For the first digital telephone systems of the early 80’s, necessary compromises were accepted so that a minimum quality (at least as good as the analogue system it replaced obviously), sufficient to convey human speech clearly enough to be understood became the minimum standard. This is based on the ear being most sensitive to frequencies between 1kHz-5kHZ – a bandwidth of 4kHz:


Not surprisingly, most human conversation takes place at these frequencies also.

This means that a sampling bandwidth of only 8kHz (8000 samples per second, about 1/5 CD quality) is sufficient to transmit the bulk of human speech nuance. This was also an easily achieved digital line transmission rate over the existing analogue copper telephone network. It became the basis for the 64kHz ISDN line rate as, when multiplexed together, eight (8 x 8kHz = 64kHz) telephone conversations could occur seemingly simultaneously over 1 line. Eight consumers now paying eight line rentals, for only 1 main line with a gadget on each end, even if they don’t speak on it, means R&D payback time for the Telcos to boot.

This principle has extended since, to today’s 150 million or so “simultaneous” phone conversations that can be carried over 1 pair of optical fibres, pulsing at 40GHz (fourty thousand million bits per second)

Anyway – enough lectures…

Pluck: The Manual description above explains why the noise disappeared from my example earlier – filtered back to a sine the same frequency as the fundamental.

—————————

  • Phase Offset (PHASE) – Sets an offset to the operator phase (graph bottom = -100% offset, graph top = +100% offset). If you vary the phase via the LFO articulation unit, the result is a vibrato effect.

————————–

I had to think about this a minute, using 2 Operators to check what happens– basically, this can set a wave cycle up to +/- 100% behind its normal start cycle, demonstrated by turning on OP2 and seeing the 2 sines add in phase to twice the height of 1 alone. If one is then set to 50% or 180 degrees out of phase then they cancel each other out. One can be set to LFO slightly in/out of phase with another to generate phaser/chorus/flange effects.

OP1 at 0%:

OP1+2 at 0%

Op1 at 0% + Op2 at 50% – Total cancellation:

To really prove this, I set Op2 at 25% which corresponds to 90 degrees phase difference, and got an amplitude about 1.4 times larger than each alone, which is what I expect from the Excel maths example I used in a prior Post where I briefly mentioned AM modulation, where 2 sines added create a 3rd Resultant sine:

http://stevepedwards.com/wp/?p=1503

y=sinX y=cosx (and composite in black = amplitude modulation), and derives at 45 degrees at an amplitude of 2sin45 = 1.414), the square root of 2.

This phase difference won’t be apparent by ear, as the resultant is another sine wave of the same frequency, but if 2 other types of waveform are used – say 2 identical Saws at 90 degrees out of phase, you get a new wave:

To give an example of aTremolo effect using 2 sines, you have to turn on the LFO of one OP in its lower Tab. As we now know, shifting identical sines out of phase just produces another sine of fixed amplitude, so to get Tremolo (constant variation in amplitude) using sine waves, one has to continuously phase drift over time using a lower period cycle which constantly changes the offset of one to the other, producing a changing sine of the same frequency, but varying amplitude between a maximum and minimum – or Tremolo:

It sounds like this = Tremolo:

Tremolo.mp3

Obviously this is academic as it is much easier to produce this effect by using LFO on the Volume itself.

The 3 knobs in the OP modifier section and the 4 ADSR controls at the bottom can be Rclicked to have Automation Clips set for them. I sped up the Tremeolo LFO speed using the SPD knob in the last example. I don’t know why, at this stage, the beginning of the note is un-modulated, perhaps to do with the notes Attack envelope not being LFO affected?

The Drop menu by the Lemon:

The options here only relate to the active OP tab – to Reset OPs all at once use main Default – Rclick in the next Preset arrows.

Just sidetracked again – something I forgot yesterday, re the Arp Presets named Formant, in FLS – a sound related tech term:

http://en.wikipedia.org/wiki/Formant

Ok, long but enjoyable Post for me so far – where do I start with the next section? The Manual.

Let’s Default for a start.

I made a 1 bar C6 note over a 2 bar loop in the Piano Roll, as a reference:

——————————————–

Articulation Section

The articulation section allows you to apply an envelope, LFO, map keys, velocity or unison mode voices to a set of predefined properties. To see the full list of envelopes/mapping in the articulation section and their meaning, please check the envelope editor page.

The filter mode supports the following articulation targets (controlled parameters):

  • Panning (PAN) – Sets the operator output panning (graph bottom = left panning, graph top = right panning).
  • Volume (VOL) – Sets the operator output volume (graph bottom = no output, graph top = maximum output).
  • Modulation Influence (MOD) – Defined the amount of ‘influence’ other operators (as modulators) can have on the current operator (graph bottom = no modulation, graph top = maximum modulation).
  • Pitch Offset (PITCH) – Sets an offset to the operator base pitch (graph bottom = -100% offset, graph top = +100% offset).
  • Phase Offset (PHASE) – Sets an offset to the operator phase (graph bottom = -100% offset, graph top = +100% offset). If you vary the phase via the LFO articulation unit, the result is a vibrato effect.
  • Plucked Damping (DAMP) – This target has effect only if the plucked strings mode is enabled for the operator. This envelope/mappings set controls the plucked string dampening amount (graph bottom = 100% dampening, graph top = no dampening).
  • Oscillator Harmonics Editor (OSC) – A feature-rich harmonics editor and external sample analyzer lets you adjust the oscillator for every operator in Sytrus.

—————————————————

The ADSR knobs in the bottom left of OP1, have to be ON for each Tab first by clicking in the Envelope radio button, before moving them affects the sound.

Panning (PAN) – Sets the operator output panning (graph bottom = left panning, graph top = right panning).

Starting with the Pan/ENV tabs selected, I twiddled a bit to try to work out what is what here (before reading the Manual!). In Wave Candy, I could see the Pan of the note started to fade on the left then came back to both LR channels, so worked out that the bottom of the Envelope window is Speaker left and the top is Speaker right.

I checked this idea, with the Decay part of the slope, to make the note fade away from the right speaker, after it stopped playing, and moved the start (Attack) section to center, so both L+R channels start equally also:

You can see the note has stopped playing, but its Pan Decay envelope stays constant Left while fading away from the Right.

Volume (VOL) – Sets the operator output volume (graph bottom = no output, graph top = maximum output).

The same vertical LR logic does not apply here, as you effectively set LR volume in the Pan tab before, so this Tab affects the note volume in terms of its ADSR envelope, as a normal graph, 0 to max volume bottom to top. This is apparent when you click the Env ON button, as you now get a slower fade in Attack with this default curve than when off:

Ow

Modulation Influence (MOD) – Defined the amount of ‘influence’ other operators (as modulators) can have on the current operator (graph bottom = no modulation, graph top = maximum modulation).

To be honest, I could not see how this works – I tried setting parameters in OP2 that should affect OP1 then changing this OP 1 Envelope but could not get a result that makes sense to me at this stage.

Pitch Offset (PITCH) – Sets an offset to the operator base pitch (graph bottom = -100% offset, graph top = +100% offset).

Self evident really, this Tab, if you centralise the line, no pitch changes occur, and each horizontal is 1 semitone, so you can effectively slide between four notes with one key press – 1 for each ADSR stage.


Phase Offset (PHASE) – Sets an offset to the operator phase (graph bottom = -100% offset, graph top = +100% offset). If you vary the phase via the LFO articulation unit, the result is a vibrato effect.

This is similar to the example I gave with the Tremolo earlier, but uses only 1 OP.

If you turn on the LFO with this Tab on also, you do indeed get the Vibrato (Pitch variation or Frequency Modulation = FM) effect, and it is best to turn the LFO SPEED up a little so it is more obvious. The section spiked in the graph is the sustain section, so vibrato will last as long as a key is depressed. On Release the vibrato stops even if the note remains.

Vibrato sounds like a whammy bar on a guitar – a pitch modulation:

Vibrato.mp3

Plucked Damping (DAMP) – This target has effect only if the plucked strings mode is enabled for the operator. This envelope/mappings set controls the plucked string dampening amount (graph bottom = 100% dampening, graph top = no dampening).

I tried some Presets to find a suitable example of a distinctive Pluck, and settled for Mando, where you can hear the Attack, then switched on Dampening with the default graph, which starts at full bottom, so the Pluck Attack is totally cut out leaving just the sustain.

Oscillator Harmonics Editor (OSC) – A feature-rich harmonics editor and external sample analyzer lets you adjust the oscillator for every operator in Sytrus.

A couple of interesting things to note here…the first bar is the fundamental note, the second being the octave. You can have no sound, or as many harmonics present, from 0-100% mix for each as you want, with each one up to 100% out of phase, if set below the X axis, which should really generate some complex waveforms.

The shape of the final waveform is drawn out for you in the window above also.

There are loads more things to cover in this Plugin that I will have to look at in later Posts, but I have to send this now before moving house tomorrow.