...that one could make a highpass filter for guitar, to filter out hum, without affecting "normal" tone? Let's say one had a form of "brickwall" filter - steep slope, maybe 6-pole, with no resonant bump - to remove most hum from a guitar signal that carries too much of it. I'm trying to think of something that would not engage in any level-based electronic "decisions", like a gate, but something that could simply work all the time, and eliminate that part of what accompanies the guitar signal that you don't need or want.
I'm guessing that this would NOT work for bass, since near-complete eliination of hum would interfere with the lowest content of a bass.
Something tells me that if it could be done, someone would have probably done it by now. On the other hand, most noise-control devices produced to date have usually attempted to address both hum and hiss in the same monolithic unit, so maybe it just slipped under the radar.
When you say hum, do you mean the ubiquitous 60hz hum?
Probably a narrow notch filter is the tool for the job. If you have a pot to tweak the notch frequency you can use an narrow bandwidth. For a fixed filter the bandwidth you need to trade bandwidth and the amount of hum removed when tolerances are considered. A very common hum filter for "HiFi" was a narrow twin-T.
The advantage of a notch over a high-pass is the notch can be deep and you don't have to trade against removing the wanted signal as much.
QuoteSomething tells me that if it could be done, someone would have probably done it by now.
As for the high-pass there's degrees of audible. If you do palm muting and chunking it's surprising what you can hear. Same goes for slapping on the bass. Whether you want to keep those low frequencies is another matter. In practice filters have a finite slope. I suspect you could hear the effect of a 1dB drop at the fundamental with careful A/B testing. So 11Hz filter for a first order . For second order order maybe 60Hz. Obviously not much good for hum removal. You might get away with a 70Hz cutoff for a second order and above and get a small amount of attenuation. I don't know the answer for a brick wall filter.
The low E on a guitar tuned at 440Hz concert pitch is around 81Hz, so, theoretically, you could cut anything below 80Hz and be fine...however, when I was modding my amps and when I make some recording, I noticed that "sub-harmonics" are useful sometimes, and the "full picture" of a guitar tone includes them: an HPF at 80Hz sounds tinny sometimes, or very focused (depending on the mix) but you can definitely hear a difference...but you can still hear some hum if it was present in the beginning, so it's not an ideal solution IMO...
As a reference, when mixing, I generally cut anything below 20-150Hz and above 8-15kHz on electric guitars, depending on the mix and the tone I'm looking for, but if you don't want to alter the tone stay below 40Hz and above 10-12kHz I'd say...
If you are using fullwave rectifiers, the 60 Hz component cancels out and what you have left is 120 Hz and all other harmonics of 60 Hz. The lead guitar output starts at 82 Hz, so there is no way you could filter without cutting some of the signal in-band. Unless you have something like tube heater wiring interfering with the signal, 60 Hz is not worth cutting and if you do have 60 Hz from a heater, twisted pair heater wire is a better solution than a filter. Many preamps use DC heater supplies just to avoid this entirely.
QuoteIf you are using fullwave rectifiers, the 60 Hz component cancels out and what you have left is 120 Hz and all other harmonics of 60 Hz.
Yes, there's many different sources of "hum". It's almost always best to get rid of the mechanism causing it than to fix it up stream. Like filtering DC supplies. Audio filtering should be a last resort.
Hum filters would probably work best for ground loops and magnetic coupling. It would be least effective capacitively couple noise like you get from a guitar pickup. The higher frequencies are boosted and fundamental isn't the main problem - that makes it go from hum to buzz.
Quote from: amptramp on May 04, 2018, 07:00:54 PM
If you are using fullwave rectifiers...
I was wondering if somebody would bring up the harmonic content thing. It's not only full-wave rectifiers, but any kind of nonlinear power conditioning.
For example, a 1/2-wave rectifier usually produces short-narrow high-current pulses at the peak of the waveform. It's the commutation of the diode and rapid current pulses that are likely to couple into the audio path at periodic intervals of 60 Hz and related harmonics.
A thought experiment: Imagine 1us, high-energy pulses coupling into the audio path on a periodic interval of 60 Hz. Do you think you could remove that with a high pass filter? Do you think you could better attenuate it with a low pass filter? What about bursts of 900 MHz energy from a cell phone? Is it effective to filter at the burst rate?
Maybe a flanger tuned perfectly to 16.67ms delay would notch all the 60 Hz harmonics and remove hum. I think we all have a good idea what that would sound like, not to mention the 16.67ms delay is discernable for some people. Whatever the case, it sounds pretty weird when you do the filter matrix set to more than 10ms.
Ron & Rob pretty well sum it up that the best solution is to prevent it from getting into the audio path.
I don't know the frequency, but i rebuild a wah sometime ago, and it had a ton of noise, and the lowest cap from power to ground that sorted things out was 4700uf, and i bet i could go up to 10000uf with even better results, since a tiny bit of noise is still there, but it is manageble. Not really the answer, but might be useful.
QuoteI don't know the frequency, but i rebuild a wah sometime ago, and it had a ton of noise, and the lowest cap from power to ground that sorted things out was 4700uf, and i bet i could go up to 10000uf with even better results, since a tiny bit of noise is still there, but it is manageble. Not really the answer, but might be useful.
If you put a small resistance in series with the +V power rail, like 100 ohms, then put the cap after the resistor, you will find the hum is dramatically reduced and you don't need such large caps.
I have too tried filtering out the LF stuff with very steep filters and it isn't a good option.
Gave up and spent more time shielding guitars, etc. Only so far this will work with single coils though.
Perhaps instead we should be thinking noise cancelling - i.e. the way Bose et al. do it.
Quote from: Transmogrifox on May 04, 2018, 10:12:29 PM
Quote from: amptramp on May 04, 2018, 07:00:54 PM
If you are using fullwave rectifiers...
...
Ron & Rob pretty well sum it up that the best solution is to prevent it from getting into the audio path.
Which is what I got from Mark's original question. Isn't he asking about single-coil mains hum, not power supply hum? That would always be 50 or 60 Hz wouldn't it?
This isn't an attempt to answer the question. It's my version of "How I Learned to Stop Worrying and Love the Hum".
When I hear recordings which have hum or some cable noise it almost never bothers me. Rather it imbues a sense of honesty, "you were there" or however you would describe it. I'm adjusting my electric guitar collection so that I have more single coil guitars (hum cancelling in middle position) than humbuckers as I'm exploring surf and jazz. This doesn't mean I never kick on the high gain channel, but I only hear the hum and buzz when I stop playing. Lesson #2. Never stop playing!
Glad I could help!
:icon_rolleyes:
Quote from: Digital Larry on May 05, 2018, 09:33:14 AM
This isn't an attempt to answer the question. It's my version of "How I Learned to Stop Worrying and Love the Hum".
When I hear recordings which have hum or some cable noise it almost never bothers me. Rather it imbues a sense of honesty, "you were there" or however you would describe it. I'm adjusting my electric guitar collection so that I have more single coil guitars (hum cancelling in middle position) than humbuckers as I'm exploring surf and jazz. This doesn't mean I never kick on the high gain channel, but I only hear the hum and buzz when I stop playing. Lesson #2. Never stop playing!
Glad I could help!
:icon_rolleyes:
It doesn't bother me either, but there are some situations where it needs to be avoided if possible. I play in a worship band at my church, and coil hum is actually distracting in that setting.
Quote
Which is what I got from Mark's original question. Isn't he asking about single-coil mains hum, not power supply hum? That would always be 50 or 60 Hz wouldn't it?
That word "always" again. There's nothing to guarantee the E&M fields correlated with mains power frequency come from a nearly pure 60 Hz source. The wiring surrounding your stage is likely carrying harmonic currents, garbage from different equipment rectifiers, even stuff from variable frequency motors, pulsed current loads, light-load burst-mode switching power supplies, etc..
What you pick up in your single coils isn't *always* dominantly the fundamental 60Hz such that notching 60 Hz would always take out the majority of it. Your pickups are getting the magnetic fields, which are generated by the currents, so even if the power in the outlet looks mostly pure doesn't *always* mean the magnetic fields are nearly pure 50/60Hz.
Whether "always" can be substituted with "usually" is something I don't know. Somebody who develops hum-canceling technology and studies the stuff that couples into guitars would probably know what is *usually* encountered.
I think my comment applies regardless of the source of the power frequency hum -- no matter where it's coming from it's not guaranteed to be a pure 60Hz sine wave.
The difficulty with narrow-band or steep filters is they have a long transient response. At a cut-off in the low end the effect is likely to be audible even if (especially if) it has negligible attenuation at the lowest guitar fundamental.
That said, my intuition tells me a twin-T passive notch filter might be the best for getting the 50 or 60Hz fundamental and minimal effect on the pass-band.. It might be some of the attenuation on the low end can be corrected with an EQ without significantly pulling hum back up, since the T notch filter when matched theoretically has perfect attenuation at the notch. A real filter is probably better because it's a little bit wider to capture the inevitable variation from true 60.0000Hz, but still very good attenuation near the notch frequency.
Make yourself a guitar booth lined with foil and ferrite plates. I guess you have to keep eye contact with your band mates through low latency video...
^ TL;DR
I'm reasonably sure Mark is asking about 60Hz hum picked up by single coils. My main point was that he's not asking how to filter out rectifier or other power conditioning hum, which seems to have dominated the discussion so far. That info is useful, but most of the time what we predominantly hear is 60Hz (or 50) hum, so tackling that if possible - "at the source" is going to probably help the most.
Mark - are you out there? Is that what you are asking about?
QuoteThat info is useful, but most of the time what we predominantly hear is 60Hz (or 50) hum, so tackling that if possible - "at the source" is going to probably help the most.
I mentioned above about the pickup noise case. It's not really a hum problem but a buzz problem. (The harmonics get boosted at 20dB/decade so it doesn't take long before they become bigger than the fundamental even though they are smaller than the fundamental on the mains signal.)
Quote from: Rob Strand on May 05, 2018, 08:42:13 PM
I mentioned above about the pickup noise case. It's not really a hum problem but a buzz problem. (The harmonics get boosted at 20dB/decade so it doesn't take long before they become bigger than the fundamental even though they are smaller than the fundamental on the mains signal.)
So how does the notch filter suggestion help with that?
QuoteSo how does the notch filter suggestion help with that?
It doesn't! My first post was a general answer. Amptramp posted about the PSU. Then I posted about the pickup. Then you guys worked out Mark might have been talking about pickups.
BTW: the DSP/noise cancelling stuff samhay mentioned effectively works out what to filter. If you have a number if harmonic noise frequencies it will actually generate a filter which has a number of narrow notches. For biomedical stuff you might come up with tougher algorithms that focus on the removing harmonic interference not noise in general. If you have ever played around with noise removal software its pretty clear it can leave some unwanted artefacts - sometimes more objectionable than the original problem! This is stuff largely impractical for analog circuits.
One way to avoid hum pickup is to use a buffer at the guitar. It can be simple - a Tillman will do and you can put the gate resistor, FET and source resistor in the guitar or in the cable plug at the guitar end and have the drain resistor, coupling capacitor and resistor to ground in a stompbox at the entry to your pedalboard. Your cable impedance goes from a variable (with the guitar level controls) number that may go as high as 250 Kohms to a constant 6800 ohms, meaning it will be far more immune to electrostatic fields. If you have a magnetic field problem in the guitar, electrostatic shielding will help overcome it because the shielding acts like a mirror for magnetic fields.
In the modern world of fluorescent and LED lighting with harmonics from their internal inverters, shielding of the guitar is essential. You have no hope of overcoming buzz in any other way. Even old fluorescent fixtures have a hitch in the sine wave when the mercury starts conducting and they can interfere without needing an inverter to do so.
QuoteOne way to avoid hum pickup is to use a buffer at the guitar. It can be simple - a Tillman will do and you can put the gate resistor, FET and source resistor in the guitar or in the cable plug at the guitar end and have the
That only really solves coupling around to the cable.
If the pickup shielding is poor it's virtually impossible to remove electronically. Some basses use a diff-amp to the pickup. While that might seem to solve all the problem it doesn't since the pickup isn't physically symmetrical. Noise couples differently to the winding start and the winding finish so the diff-amp doesn't cancel perfectly. If you shield the windings it evens things out and the diff-amp takes up the remainder.
Some noisy pickups can be quietened by making sure the winding finish is the ground but then you might find it can get noisy touching the pole pieces.
Quote from: GGBB on May 05, 2018, 08:02:19 PM
^ TL;DR
I'm reasonably sure Mark is asking about 60Hz hum picked up by single coils.
The point being made is whether coming in through directly connected PSU's, or through the air into the coil, this isn't likely to be a pure 60 Hz sine wave. It's most likely to be jagged pulses that repeat on 16ms or 8ms intervals. Sounds low frequency, but filtering off the low frequency (fundamental) takes away only a small portion of what you hear.
I'm pretty sure Mark lit the fuse and now he's just sitting back watching the show with an amused grin on his face :)
Popcorn anybody?
FYI: (many examples over the past 40 years or so.)
https://www.researchgate.net/publication/235388107_Power_line_interference_cancellation_in_in-vivo_neural_recording
For guitars it's much easier to get out the copper tape. For biomed stuff you *have to* connect to the patient's body.
QuoteI'm pretty sure Mark lit the fuse and now he's just sitting back watching the show with an amused grin on his face :)
Popcorn anybody?
:icon_mrgreen:
(http://2.bp.blogspot.com/-haa5o5YXFvE/UJc2Ax_BYQI/AAAAAAAADeU/09GzSTzsey0/s1600/Ballot+Box+Bunny+melon+bomb.jpg)
Definitely not sitting back like Bugs. I got busy the rest of the day yesterday and couldn't check back to see any responses until just now. The discussion has been thoughtful and thought-provoking, so thanks to all.
Yes, my query was in regard to hum in single-coil pickups. Some folks find their personal cure lies in any of the varieties of hum-cancelling pickups that approximate SC tone; maybe supplemented by forms of shielding. Others find the tone of their SC pickup something they wish to hang onto, so they seek other sorts of cures.
Like others here, a lttle bit of hum is not always a deal-breaker. The problem lies in the extent to which a great many players take that liveable hum and plug it into a pedal, or even succession of pedals, that amplify that hum manyfold. And THAT's where noise-control products can start to become objectionable for many, perverting the attack and decay characteristics of the signal as the threshold gets turned up to keep out the amplified hum.
I like the idea of a steep notch filter. The question this prompts, however, is whether what it takes to provide a serviceable degree of attenuation introduces any unwanted "humps" around the notch frequency. The other question it prompts is whether going at 60hz in surgical fashion still leaves us with problematic content at multiples of 60hz.
QuoteI like the idea of a steep notch filter. The question this prompts, however, is whether what it takes to provide a serviceable degree of attenuation introduces any unwanted "humps" around the notch frequency.
If the bandwidth of the notch too wide (ie. the Q too low) then you will get a wide "suck out" region either side of the notch which is unwanted. So a high Q notch with a pot to tweak the frequency is the best bet. You can design high-order notch filters which produce humps each side of the notch like you see with low-pass and high-pass filters. These Notch filters are fraught with tolerance problems.
QuoteThe other question it prompts is whether going at 60hz in surgical fashion still leaves us with problematic content at multiples of 60hz.
The annoyance of the harmonics, the buzz, will still be present I'm sure. I have my doubts the notch will be good enough for a single coil pickup. It might stop the massive hum you get when you place the guitar close to the amp, or near a plug-pack, but that's about it.
If you think about guitar speakers and the general tendency to cut bass on high-gain pedals (eg. the 47nF cap on the tube screamer) there's already some low frequency cut going on and it's not good enough.
[Edit: The notch with do *something* but it's probably not enough. It seems to me that filtering at 150Hz/180Hz is far more noticeable.]
Quote from: Transmogrifox on May 05, 2018, 10:17:24 PM
The point being made is whether coming in through directly connected PSU's, or through the air into the coil, this isn't likely to be a pure 60 Hz sine wave.
Right, but no one has said otherwise, and you're the only one who has used the word "pure."
Quote from: Rob Strand on May 06, 2018, 06:53:08 PM
So a high Q notch with a pot to tweak the frequency is the best bet.
Maybe. You still have the transient response of the filter to deal with. There's a chance a high-Q notch will muddle up palm mutes or other forms of percussive playing. This might be the optimization:
>Too high Q, long transient response, audible.
> Too low Q, too much suck out of band.
QuoteThe other question it prompts is whether going at 60hz in surgical fashion still leaves us with problematic content at multiples of 60hz.
That is the point I was trying to make however it may have come off. I think the answer is what Rob has been saying.
Quote from: GGBB on May 06, 2018, 07:29:04 PM
Right, but no one has said otherwise, and you're the only one who has used the word "pure."
Yes, with emphasis on *not* pure. Which means everything not 60 Hz is still audible after you notch out 60 Hz, whether it's coming from the pickup or from somewhere else.
QuoteMaybe. You still have the transient response of the filter to deal with. There's a chance a high-Q notch will muddle up palm mutes or other forms of percussive playing. This might be the optimization:
>Too high Q, long transient response, audible.
> Too low Q, too much suck out of band
Yes it crossed my mind you wouldn't want to build a drum emulator. I suspect it would end up with a Q between 5 and 10 (roughly 1/3 to 1/6 octave equalizer). If you consider mains frequency wander and temperature effects (on the caps) you probably wouldn't want to go much higher than Q=10.
Quote from: Transmogrifox on May 06, 2018, 07:45:06 PM
Yes, with emphasis on *not* pure. Which means everything not 60 Hz is still audible after you notch out 60 Hz, whether it's coming from the pickup or from somewhere else.
Naturally - that's how notch filters work.
Quote from: GGBB on May 07, 2018, 07:45:26 PM
Quote from: Transmogrifox on May 06, 2018, 07:45:06 PM
Yes, with emphasis on *not* pure. Which means everything not 60 Hz is still audible after you notch out 60 Hz, whether it's coming from the pickup or from somewhere else.
Naturally - that's how notch filters work.
Sorry to detract from the thread. I think we're both saying the same thing and we don't get each other's language.
If we're saying the same thing then we all agree:
1) The original intent of the thread was a question about whether a high-pass filter can be designed to sufficiently reduce 60 Hz hum and not significantly affect the guitar tone.
2) The overwhelming response is to the question expressed by reading between the lines. Yes you can probably make a filter to eliminate a 60Hz sine wave and still keep it out of the way of your guitar tone, but even so this probably doesn't solve the problem implied by the original question.
3) It doesn't matter where it comes from, whether currents in surrounding wiring couples into the guitar pickups or whether it's coming in to/from the amp, rack, lighting, or stompboxes, etc, or all of the above at the same time in different parts of the circuit. It all has harmonic components that remain even if 60 Hz is removed, and we still call it "60 Hz hum" and the more discerning call it something like "buzz".
Something I may be unwittingly assuming to be obvious to everybody is the physics behind power line noise getting into circuits. A lot of people who frequent this forum pick up (pun intended) a lot of this stuff and show evidence of a pretty good foundation in physics and engineering (Gordon included), so it's natural to assume the general audience is on the same page regarding things left unstated. Let that be a compliment to all of you whether it's true or not)
For those of whom this assumption is wrong, here's my attempt at outlining it. Note, this still requires a certain level of background knowledge, but there are enough buzz words (pun intended again) to google these concepts and maybe explore a bit.
1) Direct conduction: The circuit generating the noise is in some way connected directly to your amplifier or stompbox (PSU rectifier, wall-wart, "suicide" capacitor, etc). In rare cases it may be a somewhat conductive floor connected to your somewhat conductive body, connected to your very conductive guitar strings which are capacitive plates coupled to your pickup coils.
2) Electric fields: The tendency everything attached to your guitar has for acting like the opposite plate of a capacitor. The other terminal of this capacitor is found in the venue wiring, lights, racks, anything not grounded with some coupling to something energized. Shielding takes care of this part of the problem.
3) Magnetic fields: A guitar pickup is an electric generator, working from a change in a magnetic field. A good primer on the physics of this: http://zerocapcable.com/?page_id=219 (http://zerocapcable.com/?page_id=219)
Further, a current in a wire generates a magnetic field https://web.pa.msu.edu/courses/1997spring/PHY232/lectures/ampereslaw/wire.html (https://web.pa.msu.edu/courses/1997spring/PHY232/lectures/ampereslaw/wire.html)
Therefore, a wire carrying current at power line frequency generates a proportional magnetic field. This magnetic field adds to the magnetic field within the coils of your pickup and is injected using the same principle in physics that makes your pickups work, so you can't exactly shield this with foil and good grounding. Magnetic fields blow through grounded conductive materials as if they are air unless they are a magnetic material (like iron).
Your guitar pickup combined with surrounding powerlines and power wiring form a loosely-coupled transformer. The primary is connected to noise you don't want, and the secondary is connected to the place you don't want the noise.
In comes the humbucking pickup and noise-cancelling wiring configurations with multiple single coil pickups. This essentially adds another primary winding that is wound in the opposite polarity as the original, so the current in both primary windings sums to something very small in the secondary.
https://en.wikipedia.org/wiki/Humbucker#How_humbuckers_work (https://en.wikipedia.org/wiki/Humbucker#How_humbuckers_work)
If humbuckers or other hum-cancelling technology is objectionable, magnetic fields can be shielded by ferrous material. You need something that will generate eddy currents when exposed to a magnetic field, then you need to ground this thing so if those eddy currents develop a voltage across the ferrous material, it will be shielded so it cannot subsequently couple capacitively (electric fields) to your guitar pickup.
The complete shield would be an iron plate covering the back of your guitar, and an iron plate in front of your guitar, spaced far enough to give you room to strum the strings and work the knobs. The assumption is magnetic fields coming in between the plates perpendicular to the coils is not correctly polarized to generate much noise. Both iron plates are lined inside with foil (electrostatic shield), and both sides are grounded.
> Magnetic fields generate eddy currents in the iron plate and energy is dissipated as heat.
> Electric fields are shorted to ground by foil lining.
> Any further electric fields generated by potential across the iron (ferrous) plate due to eddy currents gets shorted to the foil, to ground.
> Other magnetic materials could be substituted for iron although iron is generally cheapest and most common.
* It is obvious this type of shielding is impractical unless you play in a really weird band where people expect you to be wearing strange stuff and have strange things on your guitar.
This is all the background to my statement about playing in a shielded booth lined with ferrite plates. The metal shield attenuates electric fields and the ferrite plates attenuate magnetic fields. This arrangement is the composition of an anechoic chamber used for measuring RF emissions for the reasons I have stated.
*************************************************************************
So...
Whether injected by conduction or electric fields or magnetic fields it is likely to couple strong in harmonics. So the sum of it all is that even if you perfectly notch 60 Hz, you don't get rid of power line noise.
The question of whether notching 60 Hz is effective at removing the objectionable hum all depends on how much of the total induced noise is 60 Hz fundamental and how much is "buzz".
Now finally the crux of it all is that all means of coupling are high-pass in nature. The higher harmonics are going to couple into the wrong places at greater magnitude than low frequency stuff. If you have harmonics, that's the part that gets coupled the strongest.
QuoteNow finally the crux of it all is that all means of coupling are high-pass in nature. The higher harmonics are going to couple into the wrong places at greater magnitude than low frequency stuff. If you have harmonics, that's the part that gets coupled the strongest.
That's about it.
FYI, I did the experiment. Took a file with junk in it and did the filtering using various equalizers/parametrics/notches available in audio software and vst plugins. (BTW a considerable percentage of these softwares produce terrible notches, nothing like the analog model that conjures in your mind when you set the parameters. Check the frequencies of the hum and harmonics before filtering.)
Removing 50/60Hz is audible but nowhere near as effective as removing 150/180Hz. In my mind removing 150/180Hz takes much of the perceived hum out of the signal. Even if you take out 50/60Hz and 150/180Hz, as expected, the annoying buzz is still present. Moreover you start to detect the whine/whistle of the harmonics between say 200 and 800Hz.
I'd say if you wanted to remove harmonics without affecting the signal (much) you need a harmonic canceller like that paper I posted above - it has source code as well BTW. There's apparently commercial products around that do this for audio now.