Joined: Aug 05, 2007 Posts: 223 Location: Elektron City, Sweden
Audio files: 1
Posted: Sun Sep 06, 2009 12:53 pm Post subject:
frijitz wrote:
/mr wrote:
I'm talking about audio-rate FM, not LFO modulation!
You're a bit of a moving target, it seems.
More of a challenge? But I'm trying to sit still.
frijitz wrote:
First, it is audio rate modulation.
20 Hz - well, not the typical FM synthesis when the two are so far apart, so it's hard to comment on.
All I have tried to say is:
In a typical linear FM synthesis application, the modulator should be able to fade in and out in amplitude without bending the carrier's pitch up or down.
If the modulation is linear and through-zero, this works.
Maybe you were saying the same thing in a different way.
(I was interpreting the mentioned "constant waveforms" as if you didn't change the modulator amplitude. (Even nonlinear FM can be tricked to track over octaves if the modulator amplitude is not varied.) But maybe I was off track.)
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
Posted: Sun Sep 06, 2009 1:28 pm Post subject:
/mr wrote:
20 Hz - well, not the typical FM synthesis when the two are so far apart, so it's hard to comment on.
Why? Again I ask, why do you think linearity depends on frequency?
Quote:
All I have tried to say is:
In a typical linear FM synthesis application, the modulator should be able to fade in and out in amplitude without bending the carrier's pitch up or down.
If the modulation is linear and through-zero, this works.
And it works very well on the Teezer, as I just demonstrated. The degree of linearity does not depend on frequency. I don't know how I can say it any different. If you want to specify exactly which frequencies you want me to use, then I will. But this is getting pretty ridiculous.
Quote:
(I was interpreting the mentioned "constant waveforms" as if you didn't change the modulator amplitude. (Even nonlinear FM can be tricked to track over octaves if the modulator amplitude is not varied.) But maybe I was off track.)
Yes, at first I didn't understand what you meant. But your idea that some large nonlinearity will magically appear above some frequency doesn't make any sense, unless I am still misunderstanding you.
Joined: Aug 05, 2007 Posts: 223 Location: Elektron City, Sweden
Audio files: 1
Posted: Sun Sep 06, 2009 1:52 pm Post subject:
frijitz wrote:
/mr wrote:
20 Hz - well, not the typical FM synthesis when the two are so far apart, so it's hard to comment on.
Why? Again I ask, why do you think linearity depends on frequency?
I don't. I'm not saying it behaves mathematically or electronically different, but when modulating with such a low frequency, the concept of "pitch" doesn't make as much sense anymore. If you hear a 1000 Hz tone one day and a 2000 Hz tone the next day, would you perceive them as 1500 Hz? The FM becomes a long-term frequency change instead of a waveform deformation where sidebands form new overtones. If the carrier and the modulator is within some octaves of each other (which I suppose is the primary reason to implement TZ linear FM in the first place) it's an entirely different thing.
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
Posted: Sun Sep 06, 2009 2:32 pm Post subject:
/mr wrote:
If you hear a 1000 Hz tone one day and a 2000 Hz tone the next day, would you perceive them as 1500 Hz?
Well fine. You are saying 20 Hz may be too low for a fused sonic impression to develop. But when I listen to the demo I can clearly hear the carrier frequency through the first plateau region. Beyond that it is (mathematically) weak and I can't hear it any more. So at 20 Hz modulation you do hear the "1500" in your question.
Quote:
The FM becomes a long-term frequency change instead of a waveform deformation where sidebands form new overtones.
Joined: Jun 29, 2008 Posts: 116 Location: Isle De Mort
Posted: Mon Sep 07, 2009 7:05 am Post subject:
I made another video, this time with much better sound quality
The footage isn't much to look at though, as it was night time.
There are a lot of sweet spots in this module, i like it a lot.
Different settings of the FM amounts & Initial Frequency bring about many changes in the sound. Often mild settings are enough, but it can get really crazy at extreme settings.
Lots of nice harmonics happening in the room, too.
Last edited by Tasmanian Alkaloid on Mon Sep 07, 2009 5:33 pm; edited 3 times in total
Joined: Feb 20, 2008 Posts: 414 Location: Stockholm
Posted: Mon Sep 07, 2009 7:26 am Post subject:
Nice sounds!
I've kinda neglected my boards for very long time now. But I've got a new project going, and I think they'll fit right in! _________________ http://www.thehumancomparator.net/
Joined: Jun 29, 2008 Posts: 116 Location: Isle De Mort
Posted: Mon Sep 07, 2009 5:47 pm Post subject:
Thanks!
It's a straightforward build, everything is well explained. The websites of Dave Brown & Richard Brewster were very helpful, & i often referred to photos of their builds to help out with my own.
Joined: Aug 05, 2007 Posts: 223 Location: Elektron City, Sweden
Audio files: 1
Posted: Tue Sep 08, 2009 6:00 am Post subject:
Example of linear FM with a Non-TZ VCO.
(CEM3340-based Doepfer A-111)
I patched together a little mp3 demonstration of what happens in a Non-TZ VCO when FM is applied in three increasing amounts, and what happens when the 0 Hz limit is reached.
I hope the attachment texts are clear enough to describe what happens, and to give some new insight to those of you who are still confused by the "Through-Zero" concept.
I haven't built my TZ VCOs yet, but some of you can try the same thing out. If all is well, the bending behaviour in the 2nd and 3rd part should then not occur. Much stronger modulation should be possible without considerable side-effects.
fm_a111_spectrum.png
Description:
Spectrum of the linear FM example. In the 1st part, the FM is within the linear range and the carrier pitch stays constant even when the modulation amount is changing. In the 2nd and 3rd part, the pitch bends upwards when the modulation is stronger than
Filesize:
358.92 KB
Viewed:
444 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.
fm_a111_waveform.png
Description:
Zoomed-in waveform portions from the three parts. 1st part; the triangle wave runs faster (steeper) and slower (less steep) as it is modulated by the square wave, but it's keeping its total pitch constant since the modulation is linear. The square is one
Filesize:
15.26 KB
Viewed:
422 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.
fm_a111.mp3
Description:
Example of linear FM with a Non-TZ VCO (CEM3340-based Doepfer A-111). Triangle wave with linear FM from a square wave one octave higher. The modulation amount is slowly varied up and down. The recording has three parts with increasing modulation amount:
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
Posted: Wed Sep 09, 2009 8:39 am Post subject:
/mr wrote:
I haven't built my TZ VCOs yet, but some of you can try the same thing out. If all is well, the bending behaviour in the 2nd and 3rd part should then not occur. Much stronger modulation should be possible without considerable side-effects.
Whew, that's a tough test. I'm glad you said "considerable". I see a freq shift of 6 Hz or so using a setup similar to yours. To get to the deep TZ regime, the bias has to be set fairly low, which means the freq control must be turned up fairly high. This puts the system into the regime where the HF tracking is degraded a bit.
What is interesting to me is that the modulating waveform makes a significant difference in the freq shift. A narrow pulse gives a bigger shift than a square wave. Apparently this is related to the (ac-coupled) modulation giving asymmetric excursions.
Fortunately, with just a touch of soft sync (sync amount turned all the way down, in fact) the signal locks right into place. I have it doing more than 200% modulation with not a bit of beating. The sync signal puts only a tiny blip on the output waveform, so the timbre isn't much affected.
I'm really glad now that I put so much effort into getting a good sync circuit working.
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
Posted: Wed Sep 09, 2009 9:50 am Post subject:
/mr wrote:
I haven't built my TZ VCOs yet, but some of you can try the same thing out. If all is well, the bending behaviour in the 2nd and 3rd part should then not occur. Much stronger modulation should be possible without considerable side-effects.
Here is an mp3 and a spectrogram of the same setup using the Teezer and a touch of soft sync. There is only a miniscule waveform alteration due to the sync. No audible pitch shift at all.
Now I'm starting to think that multi-operator patches could possibly work!
Ian
sono3.jpg
Description:
Filesize:
27.86 KB
Viewed:
339 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.
Joined: Aug 05, 2007 Posts: 223 Location: Elektron City, Sweden
Audio files: 1
Posted: Wed Sep 09, 2009 9:57 am Post subject:
frijitz wrote:
I'm really glad now that I put so much effort into getting a good sync circuit working.
Yes, that's great to hear! In analog FM applications, sync is probably more useful than what one would normally expect, and maybe even necessary to get it really close to "perfect"... Sync is not a core functionality of FM (it's rather something you'd prefer not to use at all, since it adds an unwanted DC offset to the signal) but in the end it could be part of the best solution. I wouldn't expect the TZ mechanism in itself to get 100% perfect, so it's good to hear that the sync facility is a welcome experience!
But in the end, if we ever wanted it "perfect", we'd just sit here with a DX7, wouldn't we.
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
Posted: Wed Sep 09, 2009 3:25 pm Post subject:
To extend the above demo a bit, here are results using .25V bias and 0-5.5V modulation depth, i.e. a modulation index of ~22 (or 2200%). Same carrier and modulation frequencies (180/360Hz), so the main sideband frequency is swept up to ~7.8kHz. As before, a small amount of sync. Again, no pitch shift and only a tiny waveform glitch from the sync.
Ian
sono4.jpg
Description:
Filesize:
38.97 KB
Viewed:
288 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.
Joined: Aug 05, 2007 Posts: 223 Location: Elektron City, Sweden
Audio files: 1
Posted: Wed Sep 09, 2009 4:19 pm Post subject:
frijitz wrote:
To extend the above demo a bit, here are results using .25V bias and 0-5.5V modulation depth, i.e. a modulation index of ~22 (or 2200%).
Wow! I actually didn't have the guts to ask you for such a high modulation index, haha. It seems very stable!
Btw, just after the last post I realized an obvious drawback of using sync: carrier and modulator cannot be detuned. (Unless one uses another VCO as sync source, of course.) I've never worked with programming FM synths myself, but I'm sure there are situations when all operators need to be detuned away from each other.
But whatever - combinations are infinite and there is still enough to explore right here!
("The ultimate TZ FM VCO" would perhaps be a twin VCO, with a silent one for just supplying a steady frequency for the other [modulated] one to sync to. Or to PLL to... I guess the DC offsets & waveform glitches would be minimized better with a PLL, for further use in a multi-operator system. (The project would grow to become a humongous physical evidence of why FM synths provided such a new world of timbres previously unreachable...))
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
Posted: Wed Sep 09, 2009 4:36 pm Post subject:
Actual tuned FM in the analog realm! Those timbre shifts are great.
Quote:
Btw, just after the last post I realized an obvious drawback of using sync: carrier and modulator cannot be detuned. (Unless one uses another VCO as sync source, of course.)
Interesting thought. Have you experimented with a detuned VCO supplying the FM/sync yet, Ian? _________________ My Site
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
Posted: Thu Sep 10, 2009 7:00 am Post subject:
[quote="/mr"]
frijitz wrote:
Btw, just after the last post I realized an obvious drawback of using sync: carrier and modulator cannot be detuned. (Unless one uses another VCO as sync source, of course.) I've never worked with programming FM synths myself, but I'm sure there are situations when all operators need to be detuned away from each other.
So we've come full circle? Originally you were concerned with the VCOs tracking each other. If you want them mistuned then it seems to me you just wouldn't use the sync.
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
Posted: Thu Sep 10, 2009 7:07 am Post subject:
I betcha I'm just not forming my question right.
Usually, when I'm modulating one VCO with another, I will play around with the frequencies of the two VCOs to get different harmonics. In this case, say I did that and had the two VCOs at different frequencies (slave higher than master), would the sync application still work (no pitch shift as you increased and decreased the mod index)? Or is that even a valid question concerning this application?
In any event, the samples are marvelous! _________________ My Site
Joined: Aug 05, 2007 Posts: 223 Location: Elektron City, Sweden
Audio files: 1
Posted: Thu Sep 10, 2009 8:17 am Post subject:
frijitz wrote:
So we've come full circle? Originally you were concerned with the VCOs tracking each other. If you want them mistuned then it seems to me you just wouldn't use the sync.
No. Apparently I didn't make myself clear enough:
I was only concerned about the fact that pitch should be unaffected by the amount of FM modulation. When you said
"It takes a bit of synching of the two oscillators, but constant waveforms at high modulation indices over many octaves can easily be attained"
I (mis)interpreted it as if you only tested a fixed modulation amount to recreate a certain waveform in different pitches, and that you needed sync in order to keep the linearity.
In other words: You were pointing out the necessity of a waveform unaffected by pitch; I was pointing out the necessity of a pitch unaffected by modulation amount. Each of these would be a result of - but not a proof of! - high linear FM precision.
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
Posted: Thu Sep 10, 2009 10:41 am Post subject:
Scott Stites wrote:
Usually, when I'm modulating one VCO with another, I will play around with the frequencies of the two VCOs to get different harmonics. In this case, say I did that and had the two VCOs at different frequencies (slave higher than master), would the sync application still work (no pitch shift as you increased and decreased the mod index)?
OK, so you turn the knobs till you get the perfect sound and then you change it by modulation?
This will take a bit more work. I know that the soft sync will capture over a fairly large range of frequency ratios. Things like 7:5, etc. But I'll need to look closer to see if it will hold its lock vs modulation depth as well as it does in the 2:1 case.
But if you are making strongly anharmonic sounds would a pitch shift of a few Hz be objectionable?
Last edited by frijitz on Thu Sep 10, 2009 10:53 am; edited 1 time in total
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
Posted: Thu Sep 10, 2009 10:52 am Post subject:
/mr wrote:
I was only concerned about the fact that pitch should be unaffected by the amount of FM modulation.
Yes, I understand now what you were saying. The sync will hold the pitch steady for a variety of frequency ratios. I'll have to look deeper to see if still holds lock under changing index for the less consonant intervals.
But my question is whether a few Hz change in the base frequency will be objectional. The notion of pitch can get pretty fuzzy for anharmonic sounds. I remember listening to a clock chime and perceiving a low -> high pitch interval, whereas my wife insisted it was the other way around!
I'll try to get some more tests done today. This is good -- figuring out the limits of what analog can do.
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
Posted: Thu Sep 10, 2009 7:43 pm Post subject:
So today I spent a little time looking at the case of carrier = 180Hz, modulation = 270Hz (2:3). I couldn't sync over as large a range, "only" up to m = 20. This is about a 4.9 octave sweep. Without sync, the 180 Hz component sags to ~170 Hz at the deepest modulation. This is audible if you do fairly rapid sweeps (e.g., 2 Hz). Slow sweeps with and without sync are given below.
Ian
sono4a.jpg
Description:
With sync.
Filesize:
48.77 KB
Viewed:
237 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.
sono5.jpg
Description:
No sync.
Filesize:
48.06 KB
Viewed:
222 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.
Joined: Aug 05, 2007 Posts: 223 Location: Elektron City, Sweden
Audio files: 1
Posted: Sat Sep 12, 2009 8:07 am Post subject:
frijitz wrote:
Slow sweeps with and without sync are given below.
Nice!
Both of them sound very useful, and any detuning problem is not easy to hear. But they do end up very different - the unsynced version is so much more hollow/metallic. I imagine this would be great to have as the start portion of percussive sounds, giving a metallic quality instead of the more conventional decaying filter or waveshaper/dist that we've heard for ages. So freshhh!
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
Posted: Sat Sep 12, 2009 9:46 am Post subject:
/mr wrote:
Both of them sound very useful, and any detuning problem is not easy to hear. But they do end up very different - the unsynced version is so much more hollow/metallic. I imagine this would be great to have as the start portion of percussive sounds, giving a metallic quality instead of the more conventional decaying filter or waveshaper/dist that we've heard for ages. So freshhh!
Thanks for the comments. Yes, they are different. Small changes in the relative initial frequency are interesting, too. Also, I had the sync level turned up a bit, which gives the initial waveform some buzz. I wonder if this would also be interesting to make striker transients, as for bells, marimbas, etc.
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
Posted: Sun Sep 13, 2009 5:42 pm Post subject:
One more example. TZ synced to modulation source at 4:5 ratio (176/220Hz). Lots of good anharmonic sounds. Sync was maintained up to 2.0 kHz (3.5 oct) or an index of m = 9 (900% modulation).
Ian
sono6.jpg
Description:
Filesize:
88.28 KB
Viewed:
239 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum You cannot attach files in this forum You can download files in this forum
Forum index » DIY Hardware and Software
FAQ
Calendar
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Chat Room
But I'm trying to sit still.
Actual tuned FM in the analog realm! Those timbre shifts are great.
