Mysterious 3rd note appears playing intervals on electric guitars, why????

[bencuri]

The background of the problem:

I always had a problem with guitar pedals especially compressors, that connecting my guitar to them I got a distorted sound. After asking about it at various places, I figured out it is caused by that certain humbucker guitars, regardless of quality, project such a high voltage signal that it makes pedals distort, no matter if they are compressors, delays or something else. The only quality guitar I have at hand is a Schecter Omen Extreme guitar with Schecter Diamond Plus humbucker pickups. I measured the output voltage Vpp on it, and based on my measurements I built a small device to tackle this problem without ruining the tone of the guitar signal and loose trebles.

For a while the problem seemed to be solved, but then I made an interesting discovery.

I noticed, that even if the voltage of the signal of the guitar fell below the limitations of the compressor input, there was still a kind of distortion present. It appeared when you played any interval at around the 20-24th fret positions, on the E-B-G strings. No matter how low the input voltage was (I tested it with as low as 5mV), that distortion on the interval was there. It was unbelieveable. When I switched the compressor off it disappeared, when I switched it on it came back. And it was not caused by my device, because this was the situation when I plugged the guitar directly into the compressor, and I used the guitar volume knob to turn down the guitar volume to avoid distortion on the input.

I had no idea what caused it, but then -listening to the problem intervals- I noticed the distortion is caused by a 3rd, mysterious note present besides the interval notes. But still I had no idea what causes it to be there.

It stayed like that until I borrowed the Schecter Omen guitar again to check if the 3rd note is present with it as well, and suddely I noticed that if I play the problem intervals on that guitar, I can hear the mysterious 3rd notes already present in the guitar signal.

I recorded a video about it. You can check it. But be careful!!!!! The video is loud. Turn down volume before you listen. First listen to it, then start raising the volume, and the 3rd note will appear when the volume is loud enough. You need to raise the volume quite much to hear it when I play the problem interval.

LISTEN ON QUALITY HEADPHONE OR SPEAKERS!!! Else you won't hear the problem note!!!

https://drive.google.com/file/d/1j-5zVS ... oQfIn/view

So for a while I thought it is the guitar that causes the problem, and very likely the compressor amplifies the silent parts in the signal as a byproduct of compression, that's why it is more audible after the compressor. You can listen to that here (again listen on good quality stuff, not built in speakers, please!!!!)

https://drive.google.com/file/d/1Hvt5U2 ... sp=sharing

So I asked about it from various sources what may cause this. Most of the time I got the reply that this is called "beating", an acoustic phenomemon. I knew that term from piano tuning, and I had a suspicion it can't be that. Thank good I searched further. Because this led me to an interesting discovery.

When you consider beating, the formula to calculate it is f1-f2=beating note.

I listened to my video again where you can hear the 3rd note in the problem interval in the guitar signal. I listened carefully, and I identified the mysterious note as 540hz.

The interval I am playing in the video is: A#5+E6. Those are 932Hz and 1318Hz. Lets calculate the beating: 1318-932=!!!! 386!!! That is not 540. So it is not beating we are dealing with here! I couldn't believe noone checked this before writing it is beating.

So, I figured out it is not beating, but what is it then? I came up with the idea to open the video where I demonstrate the 3rd note in the guitar signal in Spectralayers Pro. That makes it available to check notes present in the audio. But surprisingly, I found that there is no note at 540Hz, I saw a gap there:



I did not understand how it's possible that I hear the 3rd note, but it is not there. I concluded the volume of that note is so low relative to the fundamentals of the interval notes that Spectralayers cannot detect it. So, I decided to check a sample from the compressor instead. If I add the compressor after the guitar, it is much easier to catch the 3rd note, the compressor makes it much more audible. So I recorded a sample like that, and I opened it in Spectralayers. And TADAAAAM!!!! The problem note was there:



I checked with the cursor, it was exatly 540Hz!!! So it is proved: the problem note exsists, it is there, it is a real problem.

However, I still couldn't understand why it can't be seen in the other sample, in the signal of the guitar, when I can hear it. So I opened that signal again, and decided to start deleting the interval fundamentals and its harmonics, hoping maybe that way the 3rd note can be heard alone at least. And then a very surprising thing happened. As soon as I deleted 1 fundamental, the 3rd note disappered. I couldn't hear it any more. I was shocked. How is that possible?

I realized that there is gap in the dry guitar signal sample at 540Hz in the editor, because the 3rd note is not there in the signal in reality, but when you play the sample, you hear it. And!!! The compressor also adds it, because as I mentioned above, in the compressor signal it becomes visible in the editor.

I am running a discussion about it at the Steinberg forum as well, and I haven't received replies yet to my latest findings, but it was already mentioned that a phenomenon called intermodulation distortion exists in non linear systems. And as the human ear is a non linear system, maybe that's why we hear that 3rd note, even if it's not there in reality, only the interval. Or, it gets into the signal when we play back the sample when it passes some problem circuit until it reaches our ear (headphone, speaker, computer).

AND! Last but not least, the compressor also behaves as a non linear system, and it adds that 3rd note to the signal. That's why it appears in compressor signals when you check it in the editor.

The question is: what causes it in compressors, and how to modify them to stop this? You can say: OK, what's the use when even so by the time the signal reaches your brain, you will hear the 3rd note. Yes and no. If you check my video you can hear it is there in the dry signal, but actually it is very silent, You need to raise the volume quite much, you start hearing it afterwards. So there certainly is a space or "headroom" until you reach the volume so that it might become audible. In case the signal can leave the compressor without that 3rd note, at normal volume levels you won't notice it in recorded stuff.

So any ideas how to modify the compressors to get rid of this?

Let me emphasize something again before anyone would ask: this is present in all intervals at the problem fret locations in both of the guitars I checked. It is always a note lower than the fundamentals of the intervals. And all my compressors add it: Keeley Compressor Plus, Gurus Optivalve, Demeter Compulator, Carl Martin Compressor/Limiter. They all add it. However the Carl Martin is interesting, because once you lower the guitar signal voltage before it enters the pedal, in the compressed wet signal it is no more present. OR at least I didn't manage to spot it so far.

Thanks for reading!

[Folderol]

It's called intermodulation and is caused by non-linearity/distortion in the kit somewhere

[bencuri]

Any ideas how to track it down in the compressor circuit? I need to report an update: I misjudged the behavior of the Carl Martin Limiter, that does not add this phantom note, neither on the input of default gain of the guitar signal (over the compressor limit level), and nor on compression. When you feed the guitars signal to it at the default gain, as the guitar signal is around 3.2V (my guitars), you can hear the clipping. On the other compressors, besides the clipping, the 3rd note appears at the mentioned high intervals. On the Carl Martin, those intervals are clean. Only the clipping occurs. Once you reduce the gain of the input signal, the clipping stops, and the high interval is still clean.

Based on the behavior of the Carl Martin, any idea what to check? The two device differs, but which element is the crucial stage here?

Also, in connection with the dry signal: what do you think is this the human ear adding the phantom note, or it is also in the circuits between playing the file and arriving to your ear?

[Jimmy B]

It's called intermodulation and is caused by non-linearity/distortion in the kit somewhere

To be specific:
A non linear device generates harmonics at multiples of the note frequencies and there can be sum and difference tones between all of the harmonics of the fundamental notes.
In this case 2f2-f1 = 1864-1318 = 546 Hz. Could that be the frequency?

[bencuri]

That is a possible solution. One thing is sure, the phantom note sounds as 540Hz, and that is the note missing before the compressor and appears in the spectrum leaving the compressor.

But what does it mean non-linear system? Maybe the pots are logarithmic, and that causes it?

[Folderol]

That is a possible solution. One thing is sure, the phantom note sounds as 540Hz, and that is the note missing before the compressor and appears in the spectrum leaving the compressor.

But what does it mean non-linear system? Maybe the pots are logarithmic, and that causes it?

It's definitely the compressor that's doing it then - always a risk with these. You've probably made it too aggressive.

[bencuri]

That is a possible solution. One thing is sure, the phantom note sounds as 540Hz, and that is the note missing before the compressor and appears in the spectrum leaving the compressor.

But what does it mean non-linear system? Maybe the pots are logarithmic, and that causes it?

It's definitely the compressor that's doing it then - always a risk with these. You've probably made it too aggressive.

No, it is on light settings. On a scale 1-100 the compression intensity is at 35%. The other two settings available are WET/DRY mix and Tone (EQ).

I have 2 other pedals that are doing the same: Gurus Optivalve and Demeter Compulator. However the Demeter differs in that it is adding the distortion on the decay. The Gurus is optical, cannot set it to compress heavily.

[Jimmy B]

But what does it mean non-linear system? Maybe the pots are logarithmic, and that causes it?

This is a very interesting discovery. I congratulate you on your thorough investigation.
A linear system, simply put, is one where the output is proportial to the input. A non-linear system is something that isn't quite a linear system. The amount of non-linearity determines the amount of distortion that it generates.
A potentiometer, even a logarithmic one, comes close to being a linear system, so it doesn't usually generate audible distortion.
There is unlikely to be one solution to making all of your compressors less non-linear. You will need a qualified electronic systems designer (or a student of electronics- they tend to be cheaper ) to look at each compressor, and its schematic, to find out where the non-linearities are and see if they can be fixed. It is likely to be much cheaper to buy a compressor that does not have this issue. But you seem to have one such compressor already?

[bencuri]

This is a very interesting discovery. I congratulate you on your thorough investigation.

Thank you very much! It was not easy especially that I am not a pro, but I am glad that the source of problems in the chain could be identified at last, after years of being puzzled.

There is unlikely to be one solution to making all of your compressors less non-linear. You will need a qualified electronic systems designer (or a student of electronics- they tend to be cheaper ) to look at each compressor, and its schematic, to find out where the non-linearities are and see if they can be fixed. It is likely to be much cheaper to buy a compressor that does not have this issue. But you seem to have one such compressor already?

Actually the compressor I found that fits my needs the best is the Keeley. Apart from the intermodulation, it is by far the most intelligent from all the ones I own, and for solo guitar I could never set plugin compressors as good as the Keeley is set. It almost never does breathing, lacks the annoying "knock" effect on transients, and it is straightforawrd. All the other compressors are not important for me by now. All of them does much more artifacts than the Keeley, and almost all of them have annoying settings. That's why it would be useful to look into the schematics of the Keeley. It is available as I know. If I manage to find someone who has experience, I will try to have it checked.

And yes, I have the Carl Martin Limiter/Compressor, it does not suffer from this intermodulation problem. All my other compressors do. (The Demeter might be an exception as it is distorting on the decay, I have never experienced anything like that so far.) The signal of the guitars I have at hand, including the Schecter, has higher voltage than the limit of the Carl Martin. So clipping occurs when the guitar is directly rooted into it, and you want to keep the volume knob at max to keep the trebles. But after about 50% reduction with my op-amp's trimmer, that clipping disappears, and the signal is clean from that moment on in every fret positions, even with the problem intervals. The only problem is that compressor has very odd settings, the Comp knob, that is referred to as Ratio in the manual can overlook the Threshold settings and can clamp down on parts that are below
the Threshold, even when the threshold is at 0 and no compression should happen. And the attack and release has a common Response knob, that makes it impossible to finetune the attack-release times for the tempo of what you play. So I can't really use it to get a pleasing sound in the end.

[ef37a]

Just my 2 penn'o'th, Compressor pedals all most likely use a FET as the gain control element. These are inherently non linear. There is a common circuit 'trick' used to reduce 2nd harmonic distortion but in the guitar world they might not bother and leave the distortion as some kind of "magic mojo"!

Even when the compressor is not working, i.e. no gain reduction, the signal is still impressed across the FET.

Some pedals might use an optical gain control device such as a CdS cell but again, they are far from linear! Even some of the old carbon composition resistors did not fully obey Ohm's Law but you need a voltage swing of 200V or so to measure any effect.
.
Dave.

[Jimmy B]

Just my 2 penn'o'th, Compressor pedals all most likely use a FET as the gain control element. These are inherently non linear. There is a common circuit 'trick' used to reduce 2nd harmonic distortion but in the guitar world they might not bother and leave the distortion as some kind of "magic mojo"!
.....................................................
Dave.

Yes, I can hear in the example that some people may prefer the sound with the added frequency. It isn't a studio compressor, after all.
What is the "trick" that you are referring to?

[ef37a]

Just my 2 penn'o'th, Compressor pedals all most likely use a FET as the gain control element. These are inherently non linear. There is a common circuit 'trick' used to reduce 2nd harmonic distortion but in the guitar world they might not bother and leave the distortion as some kind of "magic mojo"!
.....................................................
Dave.

Yes, I can hear in the example that some people may prefer the sound with the added frequency. It isn't a studio compressor, after all.
What is the "trick" that you are referring to?

Oh! Beer into water. Two capacitors, almost always 100nF, in series from drain to ground and the junction connected to the FET's gate. That cancels even harmonic distortion to a great extent apparently. I bet you would find a schematic by JLHood pretty easily.

Dave.

[Hugh Robjohns]

While it is possible that the phantom note is the result of distortion products when compressor the gain reduction is being applied, it is just as likely to be due to input overload.

If I recall correctly, the OPs guitars (and/or playing style) generate an unusually high output level.

I also note that the Carl Martin compressor, which doesn't exhibit the phantom note, operates internally with a 12V power rail, so has a greater input headroom than standard 9V pedals. (Roughly +14dBu instead of +11dBu).

[bencuri]

While it is possible that the phantom note is the result of distortion products when compressor the gain reduction is being applied, it is just as likely to be due to input overload.

If I recall correctly, the OPs guitars (and/or playing style) generate an unusually high output level.

I also note that the Carl Martin compressor, which doesn't exhibit the phantom note, operates internally with a 12V power rail, so has a significantly greater input headroom than standard 9V pedals.

I mentioned that I have the Gurus Optivalve compressor, purchased brand new with its factory PSU. That is 12V. Does the same intermodulation. I have contacted Keeley before for input limits, I got the reply that I can use it even with 12V PSU. Same intermodulation.

If you read my first post, it revealed, that if I record the guitar plugged DIRECTLY!!! into the USB interface, the singnal has no 3rd phantom note. But when you listen, you can hear it. I also mentioned, that the compressors do the intermodulation even when the input gain is okay. I also already mentioned in the first post, that the compressor do the intermodulation even if the input signal is 5mV peak to peak. When all of them has 1Vpp input limit at least.

Please read my fist post more carefully. Most of the questions that arise from your comment are already mentioned there. Thank you!

[Jimmy B]

Oh! Beer into water. Two capacitors, almost always 100nF, in series from drain to ground and the junction connected to the FET's gate. That cancels even harmonic distortion to a great extent apparently. I bet you would find a schematic by JLHood pretty easily.

Dave.

Thanks.
I hadn't seen that before.

[bencuri]

Just my 2 penn'o'th, Compressor pedals all most likely use a FET as the gain control element. These are inherently non linear. There is a common circuit 'trick' used to reduce 2nd harmonic distortion but in the guitar world they might not bother and leave the distortion as some kind of "magic mojo"!
.....................................................
Dave.

Yes, I can hear in the example that some people may prefer the sound with the added frequency. It isn't a studio compressor, after all.
What is the "trick" that you are referring to?

Oh! Beer into water. Two capacitors, almost always 100nF, in series from drain to ground and the junction connected to the FET's gate. That cancels even harmonic distortion to a great extent apparently. I bet you would find a schematic by JLHood pretty easily.

Dave.

Don"t you refer to something similar to what can be seen in this one?

[bencuri]

Just my 2 penn'o'th, Compressor pedals all most likely use a FET as the gain control element. These are inherently non linear. There is a common circuit 'trick' used to reduce 2nd harmonic distortion but in the guitar world they might not bother and leave the distortion as some kind of "magic mojo"!
.....................................................
Dave.

Yes, I can hear in the example that some people may prefer the sound with the added frequency. It isn't a studio compressor, after all.
What is the "trick" that you are referring to?

Oh! Beer into water. Two capacitors, almost always 100nF, in series from drain to ground and the junction connected to the FET's gate. That cancels even harmonic distortion to a great extent apparently. I bet you would find a schematic by JLHood pretty easily.

Dave.

Unfortunately the Keeley Compressor uses BJT, not FET:

[Jimmy B]

While it is possible that the phantom note is the result of distortion products when compressor the gain reduction is being applied, it is just as likely to be due to input overload.

If I recall correctly, the OPs guitars (and/or playing style) generate an unusually high output level.

I also note that the Carl Martin compressor, which doesn't exhibit the phantom note, operates internally with a 12V power rail, so has a significantly greater input headroom than standard 9V pedals.

I mentioned that I have the Gurus Optivalve compressor, purchased brand new with its factory PSU. That is 12V. Does the same intermodulation. I have contacted Keeley before for input limits, I got the reply that I can use it even with 12V PSU. Same intermodulation.

If you read my first post, it revealed, that if I record the guitar plugged DIRECTLY!!! into the USB interface, the singnal has no 3rd phantom note. But when you listen, you can hear it. I also mentioned, that the compressors do the intermodulation even when the input gain is okay. I also already mentioned in the first post, that the compressor do the intermodulation even if the input signal is 5mV peak to peak. When all of them has 1Vpp input limit at least.

Please read my fist post more carefully. Most of the questions that arise from your comment are already mentioned there. Thank you!

This seems a bit harsh. Did you mean it to look like that?

[ef37a]

Just my 2 penn'o'th, Compressor pedals all most likely use a FET as the gain control element. These are inherently non linear. There is a common circuit 'trick' used to reduce 2nd harmonic distortion but in the guitar world they might not bother and leave the distortion as some kind of "magic mojo"!
.....................................................
Dave.

Yes, I can hear in the example that some people may prefer the sound with the added frequency. It isn't a studio compressor, after all.
What is the "trick" that you are referring to?

Oh! Beer into water. Two capacitors, almost always 100nF, in series from drain to ground and the junction connected to the FET's gate. That cancels even harmonic distortion to a great extent apparently. I bet you would find a schematic by JLHood pretty easily.

Dave.

Unfortunately the Keeley Compressor uses BJT, not FET:

Ah, Sod's law innit! But anyway, bipolar transistors are not noted for great linearity unless you use a lot of them and oodles of NFB. Nine volts also does not give a lot of headroom.

Dave.

[bencuri]

This problem has nothing to do with headroom. If you read my previous replies, I already mentioned that the Keeley support confirmed I can use the pedal with 12V, for more headroom. Changes nothing. Headroom only affect the input clipping. That also occurs with my guitar, but I have tackled that problem already with a device I built. Even if you increase headroom, the intermodulation is there, and there is no improvement in that regard. It has been confirmed elsewhere too in a comment, that the element that causes the problem does this no matter the headroom or the input signal strength.

[bencuri]

While it is possible that the phantom note is the result of distortion products when compressor the gain reduction is being applied, it is just as likely to be due to input overload.

If I recall correctly, the OPs guitars (and/or playing style) generate an unusually high output level.

I also note that the Carl Martin compressor, which doesn't exhibit the phantom note, operates internally with a 12V power rail, so has a significantly greater input headroom than standard 9V pedals.

I mentioned that I have the Gurus Optivalve compressor, purchased brand new with its factory PSU. That is 12V. Does the same intermodulation. I have contacted Keeley before for input limits, I got the reply that I can use it even with 12V PSU. Same intermodulation.

If you read my first post, it revealed, that if I record the guitar plugged DIRECTLY!!! into the USB interface, the singnal has no 3rd phantom note. But when you listen, you can hear it. I also mentioned, that the compressors do the intermodulation even when the input gain is okay. I also already mentioned in the first post, that the compressor do the intermodulation even if the input signal is 5mV peak to peak. When all of them has 1Vpp input limit at least.

Please read my fist post more carefully. Most of the questions that arise from your comment are already mentioned there. Thank you!

This seems a bit harsh. Did you mean it to look like that?

Definately! And Hugh knows exactly why. I posted this to get help, not troublemakers whose all time practice is not reading my posts just trolling them, repeating aspects already mentioned and sorted out, making it seem as if the one who opens the thread is stupid and they are the professors. Very likely Hugh doesn't like the situation when it turns out "Good" quality devices may have glitches. He behaved the same when I opened a related thread in the past. He closed that in rush before I could post my measurements about the output of the Schecter, to silence trolls. Luckily this time they didn't show up in the thread, only Hugh. But like that this is already a better situation than before.

To tell the truth I was surprised on your reaction to my opening post. This is the attitude that I rarely see here. Not reading details, repeating things sorted out then blaming the thread opener if he points it out is more common.

No need to comment Hugh, I know it is not an obligation to be here for me!

[Wonks]

I will point out that guitar compressor pedals use a lot of gain reduction and then a lot of make up gain in order to get a note to sustain.

And what to compressors do? Make quiet sounds louder.

I haven’t thought of any clear reason for the third note occurring, but I think it’s probably there normally but so quiet you can’t hear it above the played notes. Putting it through the compressor simply raises it to an audible level.

[Murray B]

Although it might be a rare phenomenon with a humbucker - given that the problem occurs on the lowest mass strings when they are closest to the pickups I'm wondering if it's wolf tones. Try lowering the pickups and see if they go away.

[Studio Support Gnome]

It's called intermodulation and is caused by non-linearity/distortion in the kit somewhere

To be specific:
A non linear device generates harmonics at multiples of the note frequencies and there can be sum and difference tones between all of the harmonics of the fundamental notes.
In this case 2f2-f1 = 1864-1318 = 546 Hz. Could that be the frequency?

Let’s take this as an interesting starting point for consideration.

Ignoring for a moment the non linear device red herring .

The sum and difference tone angle is of potential relevance ,

It should be borne in mind that the guitar is not a pure sine fundamental signal but a composite of harmonics.

The 546Hz point being close enough to 540 for uncertainty limits of measurement, inaccuracies of tuning and intonation , to combine to make it a much more interesting avenue of thought .

For example , It’s a common playing effect to Mute the fundamental leaving the harmonic series still ringing . ( tap harmonics , pinch harmonics and other related techniques, especially when partial muting is applied )

These can involve both left hand technique and right hand technique, .

Now I’m not saying that the OP is simply creating these and not realising it . I am simply illustrating that focussing attention solely on fretted note fundamental frequencies is potentially blinding people to other possible explanations or at least avenues of investigation.

Play a solid body acoustically and listen to notes around the 12th and 24th , you’ll also hear the dead end of the string .

Very high output pickups ,especially own brand ones , are often microphonic to some degree .

Potentially that’s another avenue to investigate .

Compressors reduce the difference between loudest and quietest components of a signal ,

Different compressors do it differently, with resulting differing effects , although I personally dislike any you can actively hear operating overtly .. I tend towards transparency in processing , be it eq or dynamics . So mucking about with loads of compressor pedals isn’t my thing ..,

I however do have issue with OP responses to several contributions , calling Hugh , the permanent technical editor emeritus and most able technical resource for forum and magazine , a trouble maker , not least of them ,

, also the implied expectation that anyone has to go and look up the individual circuits and be familiar with all the Mentioned pedals in a post almost as long as one of mine in my wordier heyday .., To be able to contribute with validity ensured that I won’t be bothering in any detail

Back to the matter at hand .
It occurs to me that , one effect that might also be good food for thought is the use of multi band techniques, where your fundamental might be squashed , but gain applied to remaining harmonic series .

Someone might band limit the gain reduction element but leave make up gain wide band , to give a bright sparkle tonality with focused low and mid range dynamic control .

( something I have routinely done with multiband dynamics processors for decades )

My point essentially being that there’s lots of possibilities that bear exploring . Just from resulting dialogue .

Without necessarily involving a line by line analysis and riposte to original post . And in the process you might learn a lot more than if someone delivers the specific answer on one line .

But the tone of some of the responses cleanly illustrates why I stopped routinely supporting the forum on the level Hugh has kept up for decades . I’m the lucky one , I can walk away anytime I feel like it and offer digitus impudicus by way of parting ,

[Studio Support Gnome]

Oh and fret clearance at 20-24th fret is often a bit low , it could easily be fret interference being exaggerated by compression , in addition to all previously mentioned