Synth Forum

Notifications
Clear all

FM "nasality"

7 Posts
3 Users
0 Likes
260 Views
Posts: 1718
Member Admin
Topic starter
 

Sometimes I get lucky, and make an FM sound that's not nasal-ish. But most of the time I'm battling, in the final phases of an FM sound design, to get rid of the nasality.

What causes this? Is it a known problem, or just in my mind?

 
Posted : 21/05/2022 3:49 am
Antony
Posts: 0
Estimable Member
 

Nasal Timbre is typically caused by an excess or inbalance of Mid Frequencies. Sometimes known as a "Mid Hump".

Now that does not mean FM-X is incorrectly balanced. It could simply mean it is the "natural" resultant Harmonic Spectrum of whatever you have programmed into the FM-X.

Also note that most real musical instruments are "nasal" in nature. Mid frequencies are typically prominent and desirable in most musical instruments, because these frequencies most resemble the Human voice. In a musical mix, Mid frequencies are also more desirable, because our Human ears are more attuned to them. We perceive this as Clarity and Sibilance.

Anyways... if it's too Nasal for your liking, you can use the Part - Effect - EQ to reduce or reshape the Mid Frequencies This is entirely normal music production procedure.

You have a choice of 2 or 3-band EQ.

You can "sweep" the Mid-Band Centre Frequency, adjust the bandwidth ( "Q" ) and level (Gain).

Remember you can reduce the prominence of Mids by lowering the Mids themselves, or by raising the Low and High bands either side (Treble, Mid, Bass are common "Amp" EQ controls).

 
Posted : 21/05/2022 5:12 am
Posts: 1718
Member Admin
Topic starter
 

What your describing is more the focus of sounds on the easily heard spectrums, something that old telephony did very well.

I wouldn't describe this mid strength as necessarily being nasally. It's just a stronger than average mids.

I'm trying to describe a timbre that's nasally, seemingly inherent to FM. Not just this mid strong-ness, which is something else, though often serves to accentuate the nasality of what I'm hearing.

Other than this mid-ness, is there something else inherent to FM that might be causing this nasality?

 
Posted : 21/05/2022 7:29 am
Jason
Posts: 7896
Illustrious Member
 

From Yamaha (wrt pulse waveforms):

Source: https://www.yamahasynth.com/learn/synth-programming/synth-basics-all-squares-pulse

Pulse Wave
Pulse waves would be like a simple ordinary On/Off switch, what varies is the length of time On versus the length of time Off. When it's On 50% and Off 50% this relationship being equal is called a "Square" wave. All other percentage On to Off create narrower pulse waves. You might, for example, have a pulse that is On just 25% of the time (Off the other 75%) or one that is On just 10% of the time, (Off the other 90%) each will have a different tone to our ears.
SquareA Square wave is described as sounding hollow and woody. Any pulse wave that is not a 50/50 "square" will sound more and more nasal in nature. A Square wave sounds closest to a Clarinet. An acoustic clarinet has a smooth woody hollowness to its sound. The narrower the pulse the more it approaches the sound of an oboe. Hollow-to-Nasal... As the ratio between the On/Off time changes the tone has a more pinched, tight or nasal tone. Think of a clavinet, it has a nasal, pinched tone and is certainly a candidate for the Pulse wave family.

As we go deeper into the Synth Basic series we'll hear more about the harmonic series. Briefly, harmonics are overtones - related pitches to the fundamental tone. The sawtooth wave family is rich and bright - it contains all harmonics in the series. A Square wave, by contrast has only the odd numbered harmonics. In a Square wave every other harmonic is missing. As the width of the pulse narrows the spacing of the missing harmonic will change accordingly. Without over analyzing this, you will notice a more nasal, thinner tone as you move farther away from the "square", 50 on/50 off ratio.

BTW - the series that focuses on synth fundamentals taking the Motif XF as the vehicle (AWM2 only) is an interesting ride as with MODX/Montage one has all of the AWM2 waveforms and possibilities discussed along with some new tricks like motion sequence. This is somewhat of a big tangent off the original request - but material worth exploring nonetheless.

 
Posted : 21/05/2022 6:23 pm
Antony
Posts: 0
Estimable Member
 

It seems what Andrew is angling at here is that all FM-X is Nasal.

(Quote "I'm trying to describe a timbre that's nasally, seemingly inherent to FM).

One might read into the query, a precursory suggestion that "Yamaha have done it wrong again". Only time will tell.

But, as the query stands, for comparison, I don't hear a generalised Nasal quality to "everything FM-X".

That does not mean Andrew does not experience this phenomena.

Some food for thought. Roland, for decades, produced a proprietary Digital Modelling Engine known as C.O.S.M. This engine appeared in nearly every Guitar FX Processor since the 90's and was only recently ditched (2019-2020 timeframe) for a brand new, ground up design. COSM was used to model many different guitar amplifiers and effects.

COSM became infamous for what musicians described as the "Cocked Wah Syndrome". It was a noticeable, narrow band, Mid frequency hump, that superimposed on every sound. This translated to everything sounding "Nasal".

Roland either couldn't or wouldn't "fix" this. Their get out of jail free card was that this was a necessary element (bias) for "cut" in a musical mix, and vehemently stood by their science. They argued that the resultant sound was in fact a very accurate depiction of what you would actually hear, in a Mix, if all the other sounds were removed.

To some extent, they were correct, regarding the "Cut" that a Studio Engineer might employ and manipulate during production.

However, they perhaps failed to realise that most of their customers were home hobbyists and didn't play in bands or record.

The forums were abound with EQ "Solutions" to diminish the effect of the baked in "Cocked Wah Syndrome".

But, as many correctly pointed out, there could be no single "fix all" band-aid solution, because what customers were actually hearing depended on the guitars they were playing and the "speakers" through which they were listening, including Headphones.

To wit, most Guitar Amplifiers with their Cabs and Speakers are basically giant Mid Frequency "Boosters", or "Mid-Band-Pass Filters".

Equally, most off-the-shelf Headphones were of the "Home HiFi" variety, and intentionally "Mid Heavy".

So, you put a Mid Heavy Signal through a Mid Boosted Amplifier, you end up with everything sounding really "Nasal".

This was the fall down of COSM (and some other Modelling Systems with different "baked in" grievances... Line6, Digitech etc).

Still, In order to benefit from the power and convenience that "FX Processors" offered, Guitarists were falling over themselves to buy FRFR Speakers (Full frequency Range, Flat Response) and powerful EQ solutions (Graphic and Parametric Equalisers). This was all done in order to eliminate any extraneous (and hard to troubleshoot) errant frequencies.

So, maybe, and this is a big maybe, the Nasal quality you are persistently hearing, could be derived from a "bad match" between your MODX and whatever Speaker/Headphone system you are using.

Roland certainly used the "bad speaker" disclaimer to avoid addressing, nee denying, the Cocked Wah Syndrome. All said and done though, history has proved there was "something fundamentally wrong" with COSM.

I'm not sure Roland ever admitted COSM was at fault. But it became a moot point when they released their new "range" of Modelling Engines, which they were eager to point out, did not suffer "cocked wah syndrome". Of course they didn't use those exact words, but everyone knew what they meant.

Anyway, if I found everything sounded Nasal from my MODX, and in the absence of any other complaints, the first thing I would look at is my amplification system.

I would also try various EQ'ing options before rushing out and spending $1000's on an FRFR system.

 
Posted : 22/05/2022 2:28 am
Posts: 1718
Member Admin
Topic starter
 

Bill, you've nailed it.

I was looking for confirmation of what I was (more often than not) hearing.

Was going to ask about rounded squares, but thought that would be distracting, so focused on the subjective in the hope it wasn't just me, and that it's a known quality of some (quite a lot) of the nature of FM (as PM).

I've made my own amp modulation and actual FM synth with code and Brian (at a point I was at odds with the FM-X vs what I was doing with actual FM) pointed out that Yamaha FM is actually PM. This opened my ears and eyes, greatly. But am still a rank amateur with it, as I try to get at futuristic basses and somewhat edgier, sometimes harsher sounds than recreating existing instruments. Trying to see where FM-X can go, as there's almost nothing exploratory at the edges of what it might be able to do.

Since then I've been progressing (admittedly in my own idiotic way) towards building up an intuitive insight and mechanical approach to making FM-X sounds, at speed. I fly through ideas and relationships between operators, and add Motion and controls as my mood and fancy strikes. Sometimes this ends in wonders. Other times it's back to what I'm supposed to be doing on computers, head down, furrowed brow.

In answer to your question about why I wait till late to try reduce nasality (when it occurs, as it often doesn't, it's merely more than 50% of the time I wind up with some/lots of nasality) it's because sometimes it'll end up suiting the sound, and because I'm working at things in an exploratory manner I'm ever hoping it will resolve or suit through my experimentation, it happens so often I think it's intrinsic to some aspect of FM-X etc... or (the reason for this question) thought that I'm just getting tired of the sound I'm working on and imagining the nasality, so begin working even more quickly to alleviate chances of fatigue. Yes, I could just turn down the monitors, but that's no fun.

Most of the time, too, am adding lots of other Motion stuff to parameters that will somewhat change the character of the sound, such that nasality drops in and out, but there's often lots of deliberately conflicting and polyrhythmic Motion Sequences by the time I'm wanting to finish up (making it hard/tedious to then reverse and discern what's actually going on) because it's fun to just go where the numbers and motion will.

It's probably, from reading your links, the times when ratios "invert" sufficiently that they're into the nasal zones of relationships with one another that this is happening, and my passion for bass is seeing that happen more often than not.

It's a great comfort to know that this is not in my mind, this is an actual characteristic of one whole half of the nature and kinds of relationships that operators can have with one another.

THANK YOU!!

 
Posted : 22/05/2022 4:08 am
Jason
Posts: 7896
Illustrious Member
 

M:C ratios would refer to the Ratio (not fixed) frequency of the modulator vs the carrier in a 2-op stack.

Say [1]-->(2)

Then M:C ratios would be the frequency ratio of operator 1 (moderator) first, then the colon, then the frequency ratio of operator 2 (carrier) second.

... assuming sine waves:

1:1, most simply stated, would mean the above 2-op structure would have operator 1's freq ratio set to 1.00 and operator 2's freq ratio set to 1.00. This creates a saw wave.

2:1 creates a square wave (pulse with equal 50/50 duty cycle - meaning "high" for 50% of the time and "low" for 50% of the time). A lower level of the modulator (freq ratio 2) can also create a triangle wave.

3:1 creates a pulse wave with a more narrow pulse (more nasal)

4:1 creates a more narrow pulse than 3:1 (even more nasal)

... and so on.

So from a 2-op stack of sine waves and non-fractional multiples of M's vs C's freq ratio - you have 2 possibilities (1:1 and 2:1) creating different waveforms that are not nasal - then you venture into nasal territory with more possibilities on that side.

And then fractional multiples create metallic or bell sounds. And there are lots of these once you introduce fine freq (ratio) control.

So FM is no more inherently "nasaly" than inherently "metallicy" as there seem to be more combinations that lean towards "metallic".

As far as getting rid of this, I'd examine M:C and see what you've got.

 
Posted : 23/05/2022 7:21 am
Share:

© 2024 Yamaha Corporation of America and Yamaha Corporation. All rights reserved.    Terms of Use | Privacy Policy | Contact Us