eliminating noise!?

S

Soultrash

Guest
i was just thinking if it would be possible to give us an option/switch that removes/dampens amp noise!?

i am not talking about a noise gate but more something like cliff "coding" the noise out.

as the axe-fx is a digital unit so many things can be programmed, distortion, saturation etc. logically it should be possible to remove amp noise with out altering the amp sound in general, right?

if so, i'd be really glad for such an option, it would make
recording much easier. i mean noise gates work but they surely have their limits...

what do you guys think?
 
I think once someone figures out how to define what parts of a complex digital audio stream are 'good sound' and what parts are 'noise' then this could be done. Maybe when quantum computing DSP's become available they will have the ability and computational horsepower to do this on the fly with negligible latency. The reason noise gates are easy it's it's purely looking at levels. More advanced ones are possibly applying greater weights to different frequency bands, but at the end of the day the processing load is fairly low.

Distortion/saturation, effects, etc. are all adding something to a signal or modifying a signal. What you are looking for is something to remove very specific pieces of an audio stream and since that can be situational it makes that very hard to do.
 
No way your going to be able to remove just "noise" and not take some desirable part of the signal with it. Not a trade off I think most users would want to make.
 
@Hellbat, @Iqdsnddist
hmm, i am actually not sure about that, it does not necessarily have to be something
that requires intense DSP power, though it would be great to read a few words about this from the master himself,
that's why i made this thread.

i am also not entirely thinking of removing noise from a existing source, it's more like
determing what is actually causing the noise and completely leave it out in the "code"
so that there is no noise to begin with, hard to explain, but if we can have different tubes,
transformer matching and all these other things that can't be done just so easily on a real amp
maybe my idea would work too...

@FreeMind
yeah, i know...but that's not what i was thinking of...
 
You can reduce noise a LOT by wiring your guitar without ground loops and proper shielding. For me the difference was night and day. Remember that stock wiring diagrams need to be modified when considering electronic shielding, as shielding adds paths to ground, and if you solder the pots together, you are going to create extra ground connections, and loops! Anyhow, it's not hard to modify the diagrams, you only need to modify the ground connections to ensure 1 path to ground. After I did this, my guitar is so quiet, my amps are quieter, everything. Start with the easy stuff first.
 
Who's to say what is "noise" though ?

An amplification system can be designed to not produce distortion. In a perfect world there could be enough head room so that nothing ever clips or distorts. It would be, for the lack of a better term "perfect". It would also sound boring and sterile. Guitarist, and essentially modern music, developed out of pushing amplification to the point of distortion. What sounded "bad" actually was sound to be pretty "good".

I would bet most users would actually prefer a "noise" in their signal, as I bet taking it all away would sound super sterile and artificial.

Noise, hum, feedback.... all part of rock n' roll
 
Who's to say what is "noise" though ?

An amplification system can be designed to not produce distortion. In a perfect world there could be enough head room so that nothing ever clips or distorts. It would be, for the lack of a better term "perfect". It would also sound boring and sterile. Guitarist, and essentially modern music, developed out of pushing amplification to the point of distortion. What sounded "bad" actually was sound to be pretty "good".

I would bet most users would actually prefer a "noise" in their signal, as I bet taking it all away would sound super sterile and artificial.

Noise, hum, feedback.... all part of rock n' roll

In terms of tube noise, perhaps, but not in terms of guitar noise! My sound has never been clearer, my high gain tighter. I have full control over when my guitar feeds back at extreme high gain, and very loud conditions, it's easier to manage. There will be noise from the gain stages, but that noise is amplified by the gain, and creates a lot of extra hiss and 'bad' noise. I don't think we should be muckling with noise reduction when emulating a tube amp, because that is part of their characteristic. Better to attack noise from a few different directions.. Try a TC Electronic Sentry gate, or an ISP Decimator 2 g-string as a gate post pre-amp, but with tracking. I use a decimator and a Mesa Mark IV, and and between my guitar wiring and the decimator and my FX8, I have an amazingly low noise levels, even with the decimator off. Clean channel, absolutely no noise noticeable via 4x12 cab at near max volume. Via the gain channel, there is a bit of tube noise, but it's barely noticable. I hardly notice my gate was off on the first trial of my new guitar wiring.

I'm certain this is the same for modeled amps.
 
@brads4
good point, though i think that my guitar is shielded well, i migh check it out, maybe there are still things that can be improved.
in general i feel that noise got worse in the last few FW, updates, even i am using less gain than before the noise has almost doubled up.

noise gates ar ekind of bad in general, as soon as i find a working setting it absolutely kills the sound/tone/sustain, it's not really a solution...

@lqdsnddist
talking noise i am mainly thinking of hum and hissing, though sometimes my axe generates some kind of weird alien digital noise,
but that's not as problematic as the hum and i really need EXTREME noisegate settings to get rid of it, which absolutely kills the sound.
 
@Soultrash in order to be accurate, the algorithms need to amplify the audio just like a real amplifier.

So you get the amplified noise floor just like the real deal. It's all part of basic amplification.
 
Depends on your definition of "noise". The correction definition of noise is a random signal caused by thermal agitation of the atoms in a conductor. Any atom higher than zero degrees Kelvin will experience electrons randomly switching in and out of valence states (IIRC). This causes random electron flow in the conductor. The higher the resistance of the conductor the more this random electron flow will induce voltage since V = I*R. This is known as Johnson noise. Johnson noise is proportional to resistance and temperature ( = sqrt(4*K*T*B*R) where K is Boltzman's constant, T is temperature in Kelvins, B is bandwidth in Hz and R is resistance).

Since Johnson noise is random and the signal from your guitar is also random then there is no way to separate the good from the bad.

Now "interference" is a different thing. Interference is not random. In our microcosm interference is typically electromagnetic interference from nearby radiators. In the old days it was predominately 60 cycle in nature but with the advent of computers and binary power supplies this interference now includes high frequency sources. However you can't predict what that interference will look like so removing it is nearly impossible. You can put notch filters in to remove 60 Hz interference but that still damages the desired signal.

The best way to remove interference is to shield against it. There are two possibilities: shield the source or shield the receiver (or both). Since you may have no control over shielding the source that leaves shielding the receiver, e.g. shielding your guitar and using well-shielded cables.
 
I should add there are noise "reduction" techniques but all alter the desired signal in some way. Noise gates are among the most useful for our particular needs. Digital cameras use various techniques based on the statistics of the image (i.e. if an area of the image is monochromatic heavier filtering is applied).

The Intelligent gate in the Axe-Fx uses some crude statistical processing where the statistics are based on the typical stats of a typical guitar.

IOW, noise reduction is destructive. You can't beat the laws of physics. So you can't remove noise without somehow altering the original signal.
 
I don't see how it could work in real time, but theoretically if you were to reprocess the same signal over and over and then average it together, the random parts of "noise" should be reduced, no ?

With regards to photography, there is an average method of noise reduction one can do by stacking many shots taken of the same scene. Noise being random isn't going to appear in exactly the same spots in each image, so it averages out.

If one were to do some sort of reamping in post, and then average all the tracks together would this have the same or similar result in reducing the random aspects of the signal ?

Again though, I would wonder if this might not actually sound desirable, or shall we say "musical". Perhaps a little bit of noise adds to the overall perception of tone, keeping things too sterile sounding ?

Relating back to digital photography, lots of photographers actually add simulated noise/grain back in post processing, because files that are too clean can look sort of "plasticy". Little random grain helps add a little more depth and texture, especially on detailess expanses such as a clear blue sky.

HB's are probably a "better" pickup with regards to noise et al, but just the same, I really love sticking hot p90's into some guitars, noise and all.
 
I don't see how it could work in real time, but theoretically if you were to reprocess the same signal over and over and then average it together, the random parts of "noise" should be reduced, no ?

With regards to photography, there is an average method of noise reduction one can do by stacking many shots taken of the same scene. Noise being random isn't going to appear in exactly the same spots in each image, so it averages out.

If one were to do some sort of reamping in post, and then average all the tracks together would this have the same or similar result in reducing the random aspects of the signal ?

Again though, I would wonder if this might not actually sound desirable, or shall we say "musical". Perhaps a little bit of noise adds to the overall perception of tone, keeping things too sterile sounding ?

Relating back to digital photography, lots of photographers actually add simulated noise/grain back in post processing, because files that are too clean can look sort of "plasticy". Little random grain helps add a little more depth and texture, especially on detailess expanses such as a clear blue sky.

HB's are probably a "better" pickup with regards to noise et al, but just the same, I really love sticking hot p90's into some guitars, noise and all.

No, because the noise in this case has been recorded. The source of the noise isn't the amp model, it's the guitar. The amp model doesn't add noise, it just amplifies it.
 
@brads4
good point, though i think that my guitar is shielded well, i migh check it out, maybe there are still things that can be improved.
in general i feel that noise got worse in the last few FW, updates, even i am using less gain than before the noise has almost doubled up.

You need to check for ground loops. Here's the deal, if there is wires connected to the bottom of your pots, and your guitar is shielded, you have created a ground loop. More than one. Modify your wiring to account for the pots being grounded to the shielding, then take 1 wire off the shielding to merge with the audio grounds, then out of the cable. This fix will make your guitar even quieter, as it will be fully shielded from interference, and ground loops. I can show you my wiring if it would help.



Noise gates work well when they track the guitar signal. My ISP Decimator 2 G-String does NOT cut off sustain, at all. It just ensures when the signal is off, the noise is clamped. This does not reduce noise while playing but it certainly ensures everything is quiet when your guitar is muted, and no sustain is affected by the gate.
 
No, because the noise in this case has been recorded. The source of the noise isn't the amp model, it's the guitar. The amp model doesn't add noise, it just amplifies it.

Yup.. if one has ever re-amped and scrutinized their dry track, it's actually very disgusting the unwanted noise some guitars exhibit. Guitars with Floyd Rose systems for example can be really bad. I also had this one guitar where the wiring rattled and once amplified it was ugly to say the least.... these are just some basic examples.
 
For some of the things I've dealt with in the past, the Axe Fx is noiseless. Grab a few pedals and a tube amp, a bunch of patch cables, an a/c to d/c power supply for the pedals and see what happens?
Only time I've had noise issues is when using single coil pickups. YMMV
 
@brads4
good point, though i think that my guitar is shielded well, i migh check it out, maybe there are still things that can be improved.
in general i feel that noise got worse in the last few FW, updates, even i am using less gain than before the noise has almost doubled up.

noise gates ar ekind of bad in general, as soon as i find a working setting it absolutely kills the sound/tone/sustain, it's not really a solution...

@lqdsnddist
talking noise i am mainly thinking of hum and hissing, though sometimes my axe generates some kind of weird alien digital noise,
but that's not as problematic as the hum and i really need EXTREME noisegate settings to get rid of it, which absolutely kills the sound.

It very well could be radio interference from your guitar! I was just recording bass takes for a song with my buddy, his bass was totally just picking up the radio. LOL I'm gonna have to fix his issue. I had to pump the gain to identify it , but it was the radio. I would wager this is your issue, not 100% sure, but ... this is not the first time for buddy and his bass. We had a recording ruined because of it one time. Not a final thankfully, but ya, annoying.
 
Back
Top Bottom