continuing tangent from bottom of this page:
http://www.mnml.nl/phpBB2/viewtopic.php ... c&start=45
For me. The primary difference is INTERFACE.
Many FM machines built from circuits are actually based on digital things. Buchla, the king of FM actually migrated to digital oscillators. But the main difference with his designs, are the the availability of points for user modulation. You could probably knock up something similar to a Buchla OSC in Max MSP, waveshapers etc etc. But I don't think it would be the same.
Guess it's personal preferences and user experience that makes one favouriable to another.
I know for certain that DX7 and friends are kinda quirky semi interesting peices of machinary, mainly designed for cheap polyphony, rather than deep musical expression. Buchla had it nailed down. Yamaha were business men (yeah yeah hippy this or that)
Not sure about the technical differences between the 2. Probably there are some. Analogue FM sounds much deeper to my ears. 2 oscs sounds much richer than a whole array of digital voices.
analogue and digital FM. differences/ similarities
-
oblioblioblio
- mnml maxi
- Posts: 2556
- Joined: Wed Apr 19, 2006 1:38 am
- Contact:
Re: analogue and digital FM. differences/ similarities
alrite then, geek central here we go.
what i read somewhere is that with analogue FM, as you increase the mod index (amplitude of the modulator) the preception for fundamental pitch changes. i would imagine if true, this is due to the very nature of an analogue oscillator and because it has no fixed phase but it just flows in time. so changing its frequency by turning a knob affects its relative phase, no? whereas with a digi osc, the phases of both carrier and modulator stay fixed unless you shift them, resulting in a steady sense of pitch. basically this is the theory i would like to put to the test.
so lets do it. i suggest the following - whoever has analogue oscillators capable of fm - set a non resonant low pass at around 100Hz-ish and then run a simple 2 sines algorithm through it, changing the mod index from modest to more agressive values, and we'll compare the outputs. i say for convenience we use a simple harmonic ratio, 1:1 or 1:2, sine carrier and sine modulator. ok? i will do the same test in Max/MSP just in half an hour and will post my result for comparison.
i dont know for a fact if this is true or not, but i want to understand. so anyone willing to participate or shed some light on the subject - welcome and thanks. anyone with a DX7 or a CZ of some kind could join too... lets just find out the truth, because if it is true, this sh!t will make a real difference, right?
lets also please stay away from claims like "analogue fm sounds better" because they seem to be different in nature and what one can do the other might can not. agreed?
hmmm, any idea why so? it may be related to the exact subject of this discussion, lets hope. cheers, im about to do my digi FM example now.
what i read somewhere is that with analogue FM, as you increase the mod index (amplitude of the modulator) the preception for fundamental pitch changes. i would imagine if true, this is due to the very nature of an analogue oscillator and because it has no fixed phase but it just flows in time. so changing its frequency by turning a knob affects its relative phase, no? whereas with a digi osc, the phases of both carrier and modulator stay fixed unless you shift them, resulting in a steady sense of pitch. basically this is the theory i would like to put to the test.
so lets do it. i suggest the following - whoever has analogue oscillators capable of fm - set a non resonant low pass at around 100Hz-ish and then run a simple 2 sines algorithm through it, changing the mod index from modest to more agressive values, and we'll compare the outputs. i say for convenience we use a simple harmonic ratio, 1:1 or 1:2, sine carrier and sine modulator. ok? i will do the same test in Max/MSP just in half an hour and will post my result for comparison.
i dont know for a fact if this is true or not, but i want to understand. so anyone willing to participate or shed some light on the subject - welcome and thanks. anyone with a DX7 or a CZ of some kind could join too... lets just find out the truth, because if it is true, this sh!t will make a real difference, right?
lets also please stay away from claims like "analogue fm sounds better" because they seem to be different in nature and what one can do the other might can not. agreed?
oblioblioblio wrote:Buchla, the king of FM actually migrated to digital oscillators.
hmmm, any idea why so? it may be related to the exact subject of this discussion, lets hope. cheers, im about to do my digi FM example now.
-
oblioblioblio
- mnml maxi
- Posts: 2556
- Joined: Wed Apr 19, 2006 1:38 am
- Contact:
can't do recording right now. my computer is dead.
depends on how the analogue oscillator is made.
if it is made right... not sure of the exact terminology... possibly AC coupled... you can get a decent amount of Mod index up when using linear frequency modulation (as opposed exponential), until the frequency of the modulation osc begins to affect the fundamental freq of carrier waveform. Enough to do plenty of work with timbre.
I'm not sure why Buchla moved to digital. I think he's a pretty forward moving person, and i think he was experimenting with digital generally. Preset storage in particular. I think every person who uses his current synth (200e), speaks very highly of the oscialltors. They also have a lot of analogue parts, such as waveshapers. But possibly you are right, maybe there is a lot more leeway to experiment with the actual frequency modulation in digital domain. If I recall correctly you can go pretty much all the way to white noise kinda sounds, and stll preserve the fundamental frequency of the osc.
I think the analogue waveshapers that can be found on Buchla circuits are a big part of the sound. They allow pretty complex ways to affect timbre. Very interesting harmonically, and work extremely well in either the carrier or the modulator.
One thing that I fully expect analogue to be better about is fluid modulation of pitch, especially of the modulation oscillator. Very interesting sounds to my ears. Maybe some differences with sidebands also. Cos I guess the methods between digital and analogue are very different.
Man, this post makes me want to play on machines and not give the tiniest sh!t about how they work. ha ha. But i think it's useful info.
You didn't mention too much about interface type concerns. For me this is primary... i never really got into that side too much with Max. I did some nice little patches but not really the same experience of connecting signals and creating a talking machine.
depends on how the analogue oscillator is made.
if it is made right... not sure of the exact terminology... possibly AC coupled... you can get a decent amount of Mod index up when using linear frequency modulation (as opposed exponential), until the frequency of the modulation osc begins to affect the fundamental freq of carrier waveform. Enough to do plenty of work with timbre.
I'm not sure why Buchla moved to digital. I think he's a pretty forward moving person, and i think he was experimenting with digital generally. Preset storage in particular. I think every person who uses his current synth (200e), speaks very highly of the oscialltors. They also have a lot of analogue parts, such as waveshapers. But possibly you are right, maybe there is a lot more leeway to experiment with the actual frequency modulation in digital domain. If I recall correctly you can go pretty much all the way to white noise kinda sounds, and stll preserve the fundamental frequency of the osc.
I think the analogue waveshapers that can be found on Buchla circuits are a big part of the sound. They allow pretty complex ways to affect timbre. Very interesting harmonically, and work extremely well in either the carrier or the modulator.
One thing that I fully expect analogue to be better about is fluid modulation of pitch, especially of the modulation oscillator. Very interesting sounds to my ears. Maybe some differences with sidebands also. Cos I guess the methods between digital and analogue are very different.
Man, this post makes me want to play on machines and not give the tiniest sh!t about how they work. ha ha. But i think it's useful info.
You didn't mention too much about interface type concerns. For me this is primary... i never really got into that side too much with Max. I did some nice little patches but not really the same experience of connecting signals and creating a talking machine.
right heres the files:
http://www.sendspace.com/file/i8tx8c
http://www.sendspace.com/file/tekmtm
test01 - ratio 1:2, test02 - ratio 1:1.
i guess what im talking about can be better heard in the test02 file, because of the 1:1 ratio. timbre changes, but in both demos theres a clear sense of pitch, as in, it all stays in one key. i want to find out if in the analogue world this will not be the same. supposedly the key has to shift.
in both demos i applied a S&H LFO to the mod index, being too lazy to twist it by hand. nothing complex, no note changing, just varying the mod index to demonstrate the way digital FM behaves. LPF at around 200Hz, no reso.
yeah i once heard steevio talk about linear & exponential FM, but i never really understood whats the difference. is it perhaps a difference in the amp envelope of the modulator, exponential or a linear one? i doubt it because imo that wouldn't suffice for two different types of FM. if not, then what is it?
lets hope someone does the test then... steevio? thanks
http://www.sendspace.com/file/i8tx8c
http://www.sendspace.com/file/tekmtm
test01 - ratio 1:2, test02 - ratio 1:1.
i guess what im talking about can be better heard in the test02 file, because of the 1:1 ratio. timbre changes, but in both demos theres a clear sense of pitch, as in, it all stays in one key. i want to find out if in the analogue world this will not be the same. supposedly the key has to shift.
in both demos i applied a S&H LFO to the mod index, being too lazy to twist it by hand. nothing complex, no note changing, just varying the mod index to demonstrate the way digital FM behaves. LPF at around 200Hz, no reso.
yeah i once heard steevio talk about linear & exponential FM, but i never really understood whats the difference. is it perhaps a difference in the amp envelope of the modulator, exponential or a linear one? i doubt it because imo that wouldn't suffice for two different types of FM. if not, then what is it?
lets hope someone does the test then... steevio? thanks
the signal side of Max is the sole reason i got into it in the first place. dont really care about interface, because once you figure out how to connect stuff the way you want to, then its just turning knobs. one can do digital FM in various ways, plugins and whatnot... then it all comes down to interpolation algorithms for the slight differences in the sound they create... and the quality of the envelopes ofcourse, which possibly makes the biggest difference in my experience.oblioblioblio wrote:You didn't mention too much about interface type concerns. For me this is primary... i never really got into that side too much with Max. I did some nice little patches but not really the same experience of connecting signals and creating a talking machine.
-
oblioblioblio
- mnml maxi
- Posts: 2556
- Joined: Wed Apr 19, 2006 1:38 am
- Contact:
mmm linear and exponential fm is not complicated. it threw me for a bit.
linear is basically what you expect, similar to digital FM. Exponential basically puts the signal affecting oscillation frequency through a function converter, which makes the way the modulation signal is added to carrier frequency exponential. You could do this easily enough in MaxMSP. Exponential aways fucks about with carrier frequency. I like to use this one for slow amplitude modulation of oscillator frequency, LFOs and stuff. You can also feed the ouput of the osc into this. Sometimes sounds nice... though makes the osc pretty unstable.
About avoiding subjective terminology like 'i prefer the sound of analogue FM'. I can understand how you wanna avoid stuff like that, but all things considered this is music, and it's always gonna end up being a personal choice of some sort.
Personally, none of this is gonna change my views about stuff. Though I would be curious about some of the results.
I just have a small number of machines, and all of them make MaxMSP cry. Both in terms of sound, and in terms of accessibility, the right balance of expnasive functionality and restricted functionality. And the feeling of using a real instrument. RIght now I'm tweaking a couple of things and then I really feel like I'm set for between 10-50 years with my machine. Finito.
I did some simple demos on my machines before. I really don't like laboratory type tests... I much prefer to hear a machine in it's expressive element. Though I understand that neither test is perfect.
http://www.mnml.nl/phpBB2/viewtopic.php ... &start=420
middle of that page. One demo is FM but probably too much reverb. Mod index stays very subtle, almost imperceitble. I didn't use a VCA between the 2 which would have made it sound nicer. The demo using Wiard osc features FM of a tube filter at about 3.00 minutes in. Digital doesn't behave like that.
linear is basically what you expect, similar to digital FM. Exponential basically puts the signal affecting oscillation frequency through a function converter, which makes the way the modulation signal is added to carrier frequency exponential. You could do this easily enough in MaxMSP. Exponential aways fucks about with carrier frequency. I like to use this one for slow amplitude modulation of oscillator frequency, LFOs and stuff. You can also feed the ouput of the osc into this. Sometimes sounds nice... though makes the osc pretty unstable.
About avoiding subjective terminology like 'i prefer the sound of analogue FM'. I can understand how you wanna avoid stuff like that, but all things considered this is music, and it's always gonna end up being a personal choice of some sort.
Personally, none of this is gonna change my views about stuff. Though I would be curious about some of the results.
I just have a small number of machines, and all of them make MaxMSP cry. Both in terms of sound, and in terms of accessibility, the right balance of expnasive functionality and restricted functionality. And the feeling of using a real instrument. RIght now I'm tweaking a couple of things and then I really feel like I'm set for between 10-50 years with my machine. Finito.
I did some simple demos on my machines before. I really don't like laboratory type tests... I much prefer to hear a machine in it's expressive element. Though I understand that neither test is perfect.
http://www.mnml.nl/phpBB2/viewtopic.php ... &start=420
middle of that page. One demo is FM but probably too much reverb. Mod index stays very subtle, almost imperceitble. I didn't use a VCA between the 2 which would have made it sound nicer. The demo using Wiard osc features FM of a tube filter at about 3.00 minutes in. Digital doesn't behave like that.
Last edited by oblioblioblio on Tue Nov 16, 2010 7:37 pm, edited 1 time in total.
-
oblioblioblio
- mnml maxi
- Posts: 2556
- Joined: Wed Apr 19, 2006 1:38 am
- Contact:
check this series of videos of Buchla Music Easel... 2 analogue oscs, FM confirguration.
Good scope porn as well. ha ha.
http://www.youtube.com/watch?v=I59Xk054EPI
oh bummer. only 1 shows FM. maybe try this one. Pretty ridiculous tonal craziness but maybe shows possiblities.
http://www.youtube.com/watch?v=RO5qRyzY ... re=related
Good scope porn as well. ha ha.
http://www.youtube.com/watch?v=I59Xk054EPI
oh bummer. only 1 shows FM. maybe try this one. Pretty ridiculous tonal craziness but maybe shows possiblities.
http://www.youtube.com/watch?v=RO5qRyzY ... re=related
oh heres what info i found on exponential: Exponential FM, where the carrier is pushed up and down an equal musical interval (therefore more Hz up than down) drifts upwards in its pitch axis as the modulation depth is increased . Linear FM allows the strength of modulation to be increased without the perceived center frequency rising.
i get it now... that means that with exponential, if you have a 100Hz carrier and apply a 1:1 ratio, the carrier will go up to 200Hz, and down to only 50Hz (octave up & down). i totally get it, very interesting. main difference is that an equal amount of Hz up and down is not an equal musical interval up and down - with a 200Hz carrier, +100Hz will make 7 semitones above, and -100Hz will make an octave below. right.
so maybe that was the difference i thought was between analogue and digital fm? or maybe not... someone do the test please
will check the videos and your demo oblio. lets not turn this into a digital vs analogue discussion, i never meant it. i am fully aware of the advantages analogue has.
i get it now... that means that with exponential, if you have a 100Hz carrier and apply a 1:1 ratio, the carrier will go up to 200Hz, and down to only 50Hz (octave up & down). i totally get it, very interesting. main difference is that an equal amount of Hz up and down is not an equal musical interval up and down - with a 200Hz carrier, +100Hz will make 7 semitones above, and -100Hz will make an octave below. right.
so maybe that was the difference i thought was between analogue and digital fm? or maybe not... someone do the test please
will check the videos and your demo oblio. lets not turn this into a digital vs analogue discussion, i never meant it. i am fully aware of the advantages analogue has.