The Holy Grail ? (second reprise)


#103

I like te idea of the Qnexus a lot but it has too many ergonomic disadvantges to be a true competitor. I thought the new MCMillen K-baord K4 pro would be the real cracker and price breakthrough (https://www.youtube.com/watch?v=2R_-7DiHmOs), until I found out it's keys do actually not move at all! So there will be no tactile feedback from it at all. Gargle!

The polyphonic aftertouch on the GEM's is excellent. Extra so when using the internal sounds. Via MIDI there is of course always a bit of response degradation because of inherent latency and 127 step resolution but then I am really nitpicking.

The S3 is one of the few of my isntruments that I have yet to see on the inside but since the S2 is the same but only provides a 61 note keybaord it stands to reason that the 76 note S3 wil have ample room inside to install an Axoloti, which should upgrade such a synth into the 21st century killer instrument with ease.

Talking of which: Somebody has bought up the GEM name and rights and has been promoting a new verison of the GEM Promega 2 on the 2017 MusikMesse (https://www.youtube.com/watch?v=hho-ewWgElA). I can only aplaude the initiative. I own a Gem Pormega 3, the then ultimate version, and it's "ancient technology" still blows about everything in so called perfect piano emulations out of the water.


#104

if you define conservatism as awareness of the past and build thing upon that, then i'm very conservative :slight_smile:

phuuh, this is a HUGE topic, so let me try to answer parts of it.

where do i start. maybe some examples that are in my head.

there have been and always will be changes/developments in music and instruments, one influences the other actually. let's take two very important developments of the last 150 years as examples.

one is the equal temperament which is now a defacto standard in most commercial western music. this brought so many nice things to music, the ability to play in every key on a fixed pitch instrument, modulation, enharmonic changes etc. but at what cost? we lost many colours in chords, since they now sound all the same in every key. we lost the sensation (and yes it really is a sensation) of a pure major third or a pure fifth. and, here comes the important point, people only trained to equal temperament will actually perceive a pure third as wrong! there is so much energy and emotion lost if you don't even know what a major third really sounds like, one that does not beat with the root note. what's my point? well, if you only let people hear equal temperament, they of course take it as the only real tuning, and are left out off all the beauty that would be there, or simply think it is wrong...a form of ignorance if you ask me.

on to the second example.

the change of gut strings to steel strings on all bowed instruments in all big orchestras. (i know there are again baroque orchestras playing gut strings, which is great) why was this done? well a steel string has better projection, more volume stays in tune, does not break that easily... sounds great! but again, at what price? it is much less able to reproduce nuances of your bowing, giving you a narrower range of sounds to work with. again, be aware of this and there is no problem. but most people today (even players themselves) are not aware of this, and many techniques are lost. things like bow vibrato etc. are not known anymore.

my point is really that every advancement always comes at a cost, it all has positive and negative sides.

emulation, sampling.... here i see the same basic problem.

nothing wrong with sampling stuff, emulating strings et all! i love it. i really dig synthetic strings! but, it makes me sad when the game of thrones title melody is played by a sampled cello instead of a real one! what makes me really sad though is, that 85% or more of the people listening don't even recognise it as sampled! why is that? well how should they, if all they ever hear are sampled cellos? you might say that this is not a big deal, but to me it is! we give away so much emotion, real feelings that could be transported in the music...

well, that is just a first short answer. sorry i have to run.

p.s. i know for a fact that hans zimmer music is very sample based, even in the end result. usually it's all samples and he lets a violinist play the violin part over the sample part to get some realness :slight_smile: of course this can never replace an orchestra, which in itself is a reacting instrument. (i play often in an orchestra that plays film music, we did several live world premieres including all the pirates stuff, and the conductor (ludwig wicky) knows the business and hans pretty well)


#105

Ahem, it's all very well talking about the fancy sounding stuff, but if you want to do good vocal synthesis it's actually pretty easy once you realise you just need to use noise and a good articulation system. I don't know half the stuff you're all talking about. To me, the "Windowing" technique is the art of watching the girl in the window opposite, all Peeping Tom-like.

So listen, here's how to make a good vocal synthesizer the 'Axoman' way. Ya takes yer noise and you uses it to makes a roughness control, and ya blends it with yer pitch and frequency and all that other lovely stuff. Then ya takes yer envelopes and uses them to makes them pitch-bendy sounds with it. The hardest part is getting throat crackle to kick-in with the right amount of pressure, cause that in itself needs to be modulated correctly. Then you just stick it through formant filters etc.

Testing the quality of your vocal synthesizer is a technical matter in it's own right. And here at Axoman Industries, we take this very seriously and have an organic listening system in place. We strongly believe that all vocal synthesizers should be trested on a four-legged friend. If the four-legged friend bolts for the door, we must go back to the drawing board. But if it howls along happily, we know we're onto a real winner, the sort of stuff the peeps over at Yamaha could only dream about.

Axoman - Pushing the boundries of advanced vocal synthesis :checkered_flag:


#106

I have an urgent question here for Marc as he obviously knows his Yamahas.

How do you go about disconnecting the ribbon cables in those things?

I have a Yamaha PSR-510, an arranger keyboard from the early 90s. I love it to bits cause it's one that was designed very well due to having an excellent one button per function design (something the manufacturers appear to be allergic to these days). Anyway, it's developed a slight problem where holding down one of the keys stops the other keys from sounding, even though that same key works fine on it's own. I'm assuming some dirt has gotten into the key mechanism contacts, so I'm trying to dismantle it.

Problem is, no matter how hard I try, I cannot figure out how to disconnect the weird white ribbon-connector sockets they use in this thing. They look like normal connectors to me, like they just pull out, but I'm concerned if I pull any harder, I'll break the soldered joints off the connector or something like that. They're so solid it feels like they're all glued in place!

Sorry for the OT-ness, but how do you disconnect cables on these early 90s PSR keyboards?


#107

28 posts were split to a new topic: Analogue vs digital


#120

A post was merged into an existing topic: Analogue vs digital


#121

this thread is in the software category, the main categories are for discussing Axoloti (e.g. Axoloti software).
can I suggest a couple of things:
- more general discussions, such as analog vs digital should be in the 'lounge'
- new topics are preferred over taking one topic in many directions, it allows other users to follow/search in the future.

I'll leave this topic in software for now, as I think its original intentions were to be about using Axoloti, but I'll probably move it to the 'lounge' as well, if it diverges again... as it has little content specifically about Axoloti, seems to keep veering to general discussions ... I guess topic title tends to lean It this way.

btw, absolutely no criticism intended, I'm as bad as anyone... and happy to have wider topics discussed in the lounge, just want to keep the main categories about Axoloti, as this is this forums main focus.

your help is appreciated, as I don't have a lot of time/inclination for moderation and splitting topics etc.


#122

A post was merged into an existing topic: Analogue vs digital


#123

The point about us all drifting off a bit is well made. Not that the results cannot be interesting. Establishing that the digital versus analog discussion is by now more based on personal preferences and established clichés then on actual fact was a good one.

I would however like to restart the initial subject again: What more can be done about taking the Axoloti beyond it's initial accent, that of offering a very flexible sytem for already established synthesis methods?

Is an anybody out there for instance trying to establish less main stream concepts? We've seen people take it in the direction of FM, modeling and even Buchla type territory has been touched. Is there however somebody who is moving away from all conventions and simply inventing his own new take on synthesis.

if so I'd love to know all about it.


#124

I had an idea way back to create 'Rotation Synthesis' or a 'Rotation Oscillator' if you like, but it was just a weird thought I never attempted to actually do anything about, due to not being a math wizz. Wouldn't know where to start, but by 'Rotation Synthesis' I mean using whatever means there is to create something that produces sound based primarily on a circle, and where the interesting sound variations 'the synthesis' is made by distorting a spiral that is formed within the circle.

So rather than the traditional method of seeing time as a streigh line, the time is a constant rotation where the virtual playhead line (the point at which you hear it) is drawn from the outer edge of the circle to the centre of the spiral. Everything that goes on, any interactions mathematical or otherwise, are only heard where the line of the playhead is. So the synthesis parameters for such a rotating oscillator would be Speed of rotation (which would give you pitch), Noise (which would distort the spiral), Smoothing (which would smooth the noise), Waveform (which would allow you to place whatever waveshape you wish along the spiral), Density (which would control how many turns in the spiral, therefore controling thickness and density of the sound), and whatever else would work well in such a scenario. You hear the full width of the playhead line rather than a point along a line. And the volume of the line changes from maximum at the outer edge to silence dead center of the circle/spiral.

So, Rotation Synthesis is based on a Rotation Oscillator.

I checked at the time if such a thing existed, and have just checked again. As far as I can tell, there is nothing out there such as a Rotation Oscillator, or indeed Rotation Synthesis. If I'm right (that's IF), and have invented this, then I hereby declare the idea free and open in order to prevent someone from patenting it, because patents should not be allowed to exist :relieved:

So, if it's new, who's gonna be the first person in the world to build the Rotation Oscillator - anyone?
Won't be me, cause I haven't a clue how to do it.

Axoman - The beautiful mind with really shitty math skills! :grin:


#125

@axoman Can you try to draw it and post a picture?
It's very likely that what you're describing is a sinewave oscillator.
(It's not a new idea to describe signals with circles, Euler did it in the late 700)


#126

It would be easier to explain I think.

I got the idea while watching a vinyl record spin. Basically, that optical illusion you get when you see what appears to be a constant supply of new grooves disappering into the center of a record. So imagine the 'Rotation' aspect of it simply being the record. The interesting part, the major difference, is that there's basically a play-head for each turn of the spiral, unlike a turntable where the play-head is a single stylus playing only one groove at a time.

So draw a virtual line from the center of the record, to the outer edge. Add a stylus for each groove in the record and place them all on that virtual line. Start playing and the sound you get is the result of every stylus (play-head) sounding at once. In other words, the density of this oscillator would be immense, completely dependent upon the amount of turns in the spiral.

And what about the grooves of the oscillator? Well, just like a record, they can contain anything you like. A sawtooth, square, noise, a voice, anything you like could be formed along the length of the spiral. But the characteristic sound it would produce would be based completely on the fact that it's a 'rotating' oscillator because the sound it outputs is the sum of every play-head the oscillator has. And remember, the amount of play-heads it has is completely dependent upon how dense the spiral is (how many spiral turns you specify using a density knob or whatever).

So that's the basis of the Rotation Oscillator itself.

As for the synthesis side of it, creating an actual Rotation Synthesis synthesizer, well there's some very different parameters for that. Density, as was explained, would control the amount of turns in the spiral and therefore the amount of play-heads making up the output of the oscillator. Speed, which unlike a standard synthesizer, does not control the speed of linear movement of a play-head over time, instead, it controls the speed the spiral is rotating, which is how you control the pitch of the oscillator. Then you would have more conventional stuff like, say, specifying a saw wave be used along the length of the spiral, anything really.

The difference between a standard saw oscillator and a Rotation Oscillator containing a saw, is that the standard oscillator runs in a linear fashion, letting you hear the single saw as the head passes over it. The Rotation Oscillator is quite different in that it's not based on a linear travel, its based on outputting the sum of every head of the spiral in a constantly revolving, ever decreasing and constantly regenerating process.

Other interesting parameters for the Rotating Oscillator are things that controll the density at various stages throughout the length of the spiral. Think of those springs where they have tight winding for the start and end of the spring, but the center part is a lot less dense. Well in a Rotating Oscillator, due to it being based on rotation, editing such things as that would have a massive impact on the sum that is output. Not only because of the changes in density, but also because what is lying directly beneath the play-head of each spiral turn, would change in time due to changes in the density of the spiral along it's length.

If you're thinking about this the way I'm thinking about it, then mind-blowingly dense trance leads would be a peice of cake for a Rotating Oscillator. A rotating oscillator, by it's very nature, has the ability to create dense, thick, lush, all that sort of stuff.

Now one last thing about the Rotating Oscillator is, due to it rotating, I can imagine the density specified by the amount of turns will also effect pitch to an extent. In other words, the more dense the oscillator gets, the lower the overall pitch will sound. Simply up the speed of rotation to counteract this effect, and you're done!

  • Earth-rumbling bass strong enough to rattle your nannas false teeth out!
  • Trance leads so dense they make you cry with emotion!
  • Pads so silky they turn silk-worms green with envy!

Axoman - Ambitious even if slightly off his rocker! :sunglasses:


#127

Yeah more like a fancy wavetable player. Good news is that you can already patch that yourself with standard objects, just don't think it as a spiral, please. Instead of reading the table in one take with an osc/phasor, you can divide its output (i mean the output of the phasor) by an integer (possibly power of 2), and make several copies of the phasor that differ by a constant (math/+c).
At this point every copy of the phasor is used to read the table, but instead of reading it from start to end, they read a little section, sorta like pitched granular synthesis.

You should expect clicks.


#128

Why would it click though?

The design I describe says the volume of the spiral is output so that the outer edge (or start) is maximum and the center (or end) is silent. The idea is so that technically the spiral contains a crossfade due to it's constantly regenerating nature.

Not 100% sure what you mean about not thinking of it as a spiral though, cause that's the very idea. Part of it's synthesis structure is the ability to change the density of the spiral at varying points along it's length. The idea being that it would allow you to alter the harmonic content and phasing inside the oscillator itself, which makes me think it would work extremely well for string-type sounds, considering such paramaters can easily be modulated.

I'm not talking about granular stuff, what I'm getting at has nothing to do with granular synthesis. I'm not sure how taking various feeds from a single sample would work the way I'm describing it, because if you're playing a single sample, I don't see how you can play it back in multiple phases at the same time using just one sample (therefore one speed) - though I'm not saying you can't.

The idea of changing the density and positions of the virtual grooves is that it all changes relative to itself. Then again, yup, it's math so I suppose you could. What I'm saying though, is that it needs to behave like a spiral would behave otherwise it would not be what I'm suggesting. For example, you couldn't just use, say, an ever decreasing in size, array of perfect cirlces cause that would defeat the whole idea of the spiral and the harmonic interactions it would create.


#129

BTW, Sputnki, something I forgot to mention (and really important).

It's vital to understand that although I'm likening it to a record, the way the sound is put onto that record is completely different. On a record, it spins at a fixed speed and the grooves are written to it while it spins. This ensures that whether the grooves are on the outer or inner edge they play back at the same speed throughout the whole track.

The Rotating Oscillator doesn't work like that.

Think of it more like taking, say, a sawtooth wave and engraving it into one long groove (in a streight line), then you take that length of groove and bend it into a spiral. The difference is that the spiral you make will play faster at the outer edge than it will in the center. This is vital to understand because it's what makes the Rotating Oscillator do what it does when you vary the density of the spiral during playback.

So it's taking it's own output (that was based on speed relative to it's density), then feeding it back into itself using a linear fashion, then bending it back into a spiral, and so the process continues. This is why the spiral aspect of it is critical. It's a rotational oscillator with a spiral at it's heart.

Axoman - Inventor of the Rotating Oscillator :grin:


#130

Trying to understand here.. turning the saw into a spiral... Do you mean the user interface, the visual part of the oscialltor must look like a spiral... Or you want to actually change the saw wave into a spiral? If this is the case, how would you go about that?

But from the description you gave, I chime in with @Sputnki and say that it sounds like a wavetable oscillator of some sort.


#131

Take the grooves of a record, completely unwind it, and put it in a straight line. If you then run a stylus along that groove at a constant speed, the recording will get faster and faster even though the stylus is moving at a constant speed. This is because the record had it's grooves cut in a 'spiral'. meaning the outer grooves are larger in circumference than the inner grooves.

This proves that the geometry of the groove can effect the pitch. If proves that if the groove were in one streight line it would behave differently if it were then wound into a spiral. And that's the idea behind it, to take what is generated at a constant rate, wrap it into a spiral, and rotate it. This has the opposite effect, because rather than recording in a spiral then unwinding it, you're recording in a streight line then winding it into a spiral. It's this spiral that causes one aspect of the oscillators behaviour.

The other aspect is that there's a stylus for every groove in the spiral. So if the density of the spiral is low, say only six turns, then you get 6x stylus to output the sound simulatneously. If the spiral is super-dense having something like 120 turns, then you get 120x stylus outputting the sound simultaneously, cause like I said, the oscillator has a stylus outputting sound from each groove simultanously.

Now you take those two aspects, add them together and you get the Rotating Oscillator. Because due to the time-shifting caused by it being a spiral, it means you can change the characteristics of the harmonics by altering how close together and far apart the grooves are as they rotate. For example, if you modulated the groove spacing across it's surface by using a sinewave, the density of the grooves would get progressively closer and far apart across the surface. Remember, you have to draw a virtual line from the outside of the circle to the center of the spiral, and no matter how many stylus you generate due to groove density, they all sit on that same line.

If you were to do this with a real record, the sound would be a mess, but the sound from each groove would sound as if it was playing at the same speed. That is not the case with the Rotating Oscillator, simply because the sound was not made in a spiral formation, but rather a streight line that was then wound into a spiral. This means that the pitch will differ from each stylus that is outputting sound, therefore giving you control over the sound, over the harmonics, over the phasing etc.

This is what makes me think that such an oscillator would totally rock at string instruments when modulating the groove density across the width of the virtual playhead line. I can imagine it would also work well for air/breath in vocal synthesis, stuff like that.


#132

Might be better to call it a Spiral Oscillator.


#133

Not related to the previous post, but has anyone ever considered making a sort of PMCK Physical Model Construction Kit of Axoloti objects?

Would really love to be able to use a complete orchestra of physical instruments, and even sound effects. All the usual stuff like brass and strings are essential, obviously, but as well as all that, I've always wanted to hear a convincing 'old church bells' type of sound. The sort you hear in the distance from really old village churches where they go in and out of tune as they continue to sound - you know that really old, quaint sound those types of bells make.

Would love to be able to do that in Axoloti convincingly, and the same goes for those really huge Gongs, like the one they strike at the beginning of those old movies.

Pianos as well. Can't remember the name of it but I came across a demo of a really impressive modeled piano once, it was even available for Linux. I was tempted to buy it but the licence agreement didn't allow using the sounds generated in it to be used as samples in a commercial product, so I never bought it.

Would be awesome to have Open Source physical models of that quality so that we could generate various pianos that can then be baked to samples for use as a rompler or whatever. So yup, would be cool to see a PMCK project go on around here, where basically, the only objects allowed to be part of the PMCK are strictly physical models of complete objects and instruments, or parts of the models of instruments so that we can mix'n'match and build our own by linking-up the patch cables.

PMCK = Physical Model Construction Kit


#134

pianoteq probably. no point in sampling it imho, because the point of modelling a piano is that voicings actually sound different every time when you play them on pianoteq, depending on dynamics of each note. since every string is calculated, that's possible. if you sample it, you loose that "natural" part of piano playing. that's exactly why most sample based pianos are hard to play for professional pianists.