Old 03-08-2017, 01:38 PM   #1
matthewjumpsoffbuildings
Human being with feelings
 
Join Date: Mar 2013
Posts: 463
Default Phase inversion noise floor trickery?

Ok this is going to be long winded...

Im using amp sims a lot, and while they sound pretty good, one thing thats always bugged me is the higain sounds amplifying my digital preamps noise floor.

While my preamps are not bad (with the volume turned up to adequate levels on my Presonus Digimax HiZ input, the noise floor is around -80db), once I add an amp head sim with a distortion pedal, you can really hear that harsh digital noise being amplified. Even if I dont have a guitar plugged in (or the guitar volume turned off), the actual white-noise-ish sound of my preamps noise floor is still super audible.

Im aware you can gate the signal chain going into the amp sim, but theres 3 problems with that:
1. it affects the attack of notes, especially the first note after a pause
2. it messes with sustain, and often when on the tail end of the gate the noise becomes really audible again
3. a gate is really just a fancy on/off switch, and when its on, that white noise is going to be part of the signal and adding its nasty brittle sound to the tone

So - i was playing around a little with phase reversal, and I discovered, no surprise, that if you feed the amp sim with 2 copies of the input track, one with the phase reversed, you get no digital noise floor. Of course, you also get no sound at all :P

But it got me thinking, what if you could feed the amp with a normal input, and then a phase inverted noise floor from a second input channel on the preamp. So I tried that, and of course that didnt work, because the noise on each input is random/unique, so the phase inversion didnt cancel anything out

So then I started thinking, what if i feed the amp sim with 2 copies of the SAME input, one of them phase inverted, but on the phase inverted one, do some kind of trickery so that when the volume of any frequency crosses a certain threshold (eg just above the natural digital noise floors level), it doesnt get any louder. Which would mean that it WOULDNT cancel out the normal (in phase) input feeding the amp sim. So then only the noise floor would get cancelled out, at all times, which would be amazing.

The big question is what kind (if any) of dynamic processor could I use to do this? basically what I need is something that can have a threshold, and stop any frequency from getting any louder than that threshold BUT not turn down the volume of the other frequencies in the signal AND not add any extra harmonics/distortion in doing so. Oh and it would have to be as close to realtime as possible, a buffer of 32-64 samples is ok but not much more... is that even possible? It sounds impossible but doesnt hurt to ask.

The other way I can think to do this is somehow in the hardware realm of digital preamps get access to the exact noise floor signal of a specific channel BEFORE the actual input signal is added to it, so then you can phase invert that and remove it, but Im guessing thats even more impossible or all preamps would be doing that already for completely noiseless gain :/

I hope this makes sense to somebody
__________________
[sun sprites] spotify facebook
matthewjumpsoffbuildings is offline   Reply With Quote
Old 03-08-2017, 02:16 PM   #2
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 29,260
Default

ReaFir may work since it can create a noise profile and filter that out.

Just a tiny nitpick... the actual preamps aren't digital, they are purely analog but sitting near digital devices and potentially picking up noise from nearby components or a noisy ground plane since combining digital and analog and similar grounds can cause digital noise to find it's way into the analog signal.

Of course the real solution is a sound card that exhibits less of this effect or a dedicated preamp etc.
__________________
Music is what feelings sound like.
karbomusic is offline   Reply With Quote
Old 03-08-2017, 02:34 PM   #3
matthewjumpsoffbuildings
Human being with feelings
 
Join Date: Mar 2013
Posts: 463
Default

Oh no this isnt interference. It is the actual pure noise floor that all digital preamps have, and its pretty much just white (or brown or pink, i can never remember which is which) noise.

Perhaps I worded it wrong - the noise is coming from the Analogue to Digital converters in my preamps, and all AD converters have a noise floor, some better than others. For example I have 3 preamps in my rack, and after level matching them, with nothing plugged into them, simply arming/monitoring them, they all exhibit a different db level noise floor. Its definitely to do with the AD process

ReaFir is too slow from what i recall.
__________________
[sun sprites] spotify facebook
matthewjumpsoffbuildings is offline   Reply With Quote
Old 03-08-2017, 02:40 PM   #4
Sju
Human being with feelings
 
Join Date: Jun 2015
Posts: 685
Default

If I understand correctly you want to hard-limit the delta signal to the noise floor's amplitude? Unfortunately it'll probably be very hard to do without introducing a lot of artifacts, which will then feed into the main signal.

Basically a well set-up gate will do what you ask, but it'll of course have that buzz under the played notes. You just gotta gain stage it so that the notes are masking the noise.

EDIT: sry, preamp noise, not pickup hum... how damn hard are you driving the amp anyway?

PS. the only way you can phase-cancel noise is if you have (near) identical signal paths for the main and delta signal, like in humbuckers or balanced cables. So not possible for different preamp channels for example as you noticed. The only options left are reducing the noise either linearly (EQ with learn noise profile function) or dynamically (gate). But if you come up with another solution, definitely apply a patent for it ASAP!

Last edited by Sju; 03-08-2017 at 02:46 PM.
Sju is offline   Reply With Quote
Old 03-08-2017, 02:43 PM   #5
Lokasenna
Human being with feelings
 
Lokasenna's Avatar
 
Join Date: Sep 2008
Location: Calgary, AB, Canada
Posts: 6,551
Default

Are you sure it isn't the guitar + pickups + cable? That's a far more likely source of noise than your A/D converters.
__________________
I'm no longer using Reaper or working on scripts for it. Sorry. :(
Default 5.0 Nitpicky Edition / GUI library for Lua scripts / Theory Helper / Radial Menu / Donate
Lokasenna is offline   Reply With Quote
Old 03-08-2017, 03:00 PM   #6
matthewjumpsoffbuildings
Human being with feelings
 
Join Date: Mar 2013
Posts: 463
Default

No - as im trying to explain - even if I have no cable plugged in to the interface - simply arming the channel there is a bed of white noise around -80db. This is most definitely the noise floor of the AD converters. Its this bed of white noise I want to remove, in my opinion it is one of the main reasons higain amp sim sounds 'brittle' and 'wrong'

Humbuckers/interference/cables have absolutely nothing to do with it.
__________________
[sun sprites] spotify facebook
matthewjumpsoffbuildings is offline   Reply With Quote
Old 03-08-2017, 03:05 PM   #7
matthewjumpsoffbuildings
Human being with feelings
 
Join Date: Mar 2013
Posts: 463
Default

Quote:
Originally Posted by Sju View Post
If I understand correctly you want to hard-limit the delta signal to the noise floor's amplitude? Unfortunately it'll probably be very hard to do without introducing a lot of artifacts, which will then feed into the main signal.
ye exactly what i want, i assumed it would be hard without artifacts hey. maybe some kind of dynamic multiband compressor??

Quote:
Originally Posted by Sju View Post
EDIT: sry, preamp noise, not pickup hum... how damn hard are you driving the amp anyway?
ridiculously hard sometimes :P

Quote:
Originally Posted by Sju View Post
PS. the only way you can phase-cancel noise is if you have (near) identical signal paths for the main and delta signal, like in humbuckers or balanced cables. So not possible for different preamp channels for example as you noticed. The only options left are reducing the noise either linearly (EQ with learn noise profile function) or dynamically (gate). But if you come up with another solution, definitely apply a patent for it ASAP!
ye i figured. im guessing on the hardware side, in the AD converters the noise floor is generated on the fly with the incoming analogue signal, its not something that exists separately and can be piped out as a separate delta signal. Im guessing engineers a lot smarter than me have had the thought to try this and its probably really hard to create/separate the delta with exactly the same noise profile as the main signal

ps now i know its called a delta signal, thanks
__________________
[sun sprites] spotify facebook
matthewjumpsoffbuildings is offline   Reply With Quote
Old 03-08-2017, 03:24 PM   #8
Lokasenna
Human being with feelings
 
Lokasenna's Avatar
 
Join Date: Sep 2008
Location: Calgary, AB, Canada
Posts: 6,551
Default

You could try using a white/pink/brown/whatever noise generator and inverting that, then bringing the level up slowly - it might remove some of the noise.
__________________
I'm no longer using Reaper or working on scripts for it. Sorry. :(
Default 5.0 Nitpicky Edition / GUI library for Lua scripts / Theory Helper / Radial Menu / Donate
Lokasenna is offline   Reply With Quote
Old 03-08-2017, 03:26 PM   #9
matthewjumpsoffbuildings
Human being with feelings
 
Join Date: Mar 2013
Posts: 463
Default

Quote:
Originally Posted by Lokasenna View Post
You could try using a white/pink/brown/whatever noise generator and inverting that, then bringing the level up slowly - it might remove some of the noise.
interesting suggestion, ill give it a shot and report back

tho my gut feeling is that seeing as using a second channel as a phase inverted noise source didnt work, the noise floor is so random that a standard noise generator may not help either... we shall see
__________________
[sun sprites] spotify facebook
matthewjumpsoffbuildings is offline   Reply With Quote
Old 03-08-2017, 04:11 PM   #10
Sju
Human being with feelings
 
Join Date: Jun 2015
Posts: 685
Default

Quote:
Originally Posted by matthewjumpsoffbuildings View Post
ps now i know its called a delta signal, thanks
Don't believe me I just pulled that out of my ass well, in this case the signal that is substracted from the main signal, that is the polarity-reversed noise, could be called delta, as in difference, probably. Some AE will soon lecture me I'm pretty sure.

Hmm, actually the signal that is the difference of the main signal and the reversed noise is the delta in this case.

PS. using random noise for generating the negative signal won't work, because the phase of the wave is, well, random. It needs to have (at least some) coherence with the part that is being removed. Even if they sound alike, it doesn't mean that the waves' phases will be similiar.

Last edited by Sju; 03-08-2017 at 04:50 PM.
Sju is offline   Reply With Quote
Old 03-08-2017, 04:14 PM   #11
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 29,260
Default

Quote:
Originally Posted by matthewjumpsoffbuildings View Post
Oh no this isnt interference. It is the actual pure noise floor that all digital preamps have, and its pretty much just white (or brown or pink, i can never remember which is which) noise.

Perhaps I worded it wrong - the noise is coming from the Analogue to Digital converters in my preamps, and all AD converters have a noise floor, some better than others. For example I have 3 preamps in my rack, and after level matching them, with nothing plugged into them, simply arming/monitoring them, they all exhibit a different db level noise floorIts definitely to do with the AD process

ReaFir is too slow from what i recall.
Just mentioning, it still isn't a 'digital preamp'; it's an analog preamp feeding an A/D converter. The converters aren't 'in' your preamps, they are after them but in the same enclosure.
__________________
Music is what feelings sound like.

Last edited by karbomusic; 03-08-2017 at 05:27 PM.
karbomusic is offline   Reply With Quote
Old 03-08-2017, 04:21 PM   #12
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 29,260
Default

Quote:
This is most definitely the noise floor of the AD converters.
I'm not sure exactly why it matters being the converters or preamps (even though it's sort of obvious the preamps job to raise the noise floor). That's all irrelevant anyway because it doesn't matter where it came from... Once you start cranking the gain in a SIM (or a real pedal/amp for that matter) you WILL eventually have to deal with the noise floor regardless of where it came from, that's inevitable.

It's not that SIMs are magically noisier than amps, it's that they tend to offer ridiculous amounts of gain (yes I said ridiculous) on top of the initial pre-amplification that most amps don't have unless you start chaining pedals and channels etc. which is now turning up that noise you picked up during pre-amplification. You can try to filter it but the real solution is using less gain, a noise gate and/or a quieter preamp assuming nothing is broken.

Physics doesn't care which but will absolutely stop us dead in our tracks when we ignore the rules.
__________________
Music is what feelings sound like.

Last edited by karbomusic; 03-09-2017 at 08:28 AM.
karbomusic is offline   Reply With Quote
Old 03-08-2017, 04:45 PM   #13
Tod
Human being with feelings
 
Tod's Avatar
 
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
Default

Quote:
Originally Posted by karbomusic View Post
ReaFir may work since it can create a noise profile and filter that out.
Quote:
Originally Posted by matthewjumpsoffbuildings View Post
ReaFir is too slow from what i recall.
That has not been my experience with it. What it might do for a -80dB noise floor, I don't know, but it's worth a try.
Tod is offline   Reply With Quote
Old 03-08-2017, 04:49 PM   #14
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 29,260
Default

Quote:
Originally Posted by Tod View Post
That has not been my experience with it. What it might do for a -80dB noise floor, I don't know, but it's worth a try.
I agree Tod... Also, -80 is pretty low (and generally insignificant until you amplify it from a noisy guitar perspective) so I'm back to too much gain. Even my UFX when gain is full results in a -62dB noise floor because that's what preamps do.
__________________
Music is what feelings sound like.

Last edited by karbomusic; 03-08-2017 at 05:59 PM.
karbomusic is offline   Reply With Quote
Old 03-09-2017, 04:58 AM   #15
4x4uk
Human being with feelings
 
4x4uk's Avatar
 
Join Date: Jan 2013
Location: Newcastle UK
Posts: 474
Default

i use reafir as an input fx on all tracks where i record electric guitars
4x4uk is offline   Reply With Quote
Old 03-09-2017, 05:07 AM   #16
cyrano
Human being with feelings
 
cyrano's Avatar
 
Join Date: Jun 2011
Location: Belgium
Posts: 5,246
Default

Quote:
Originally Posted by matthewjumpsoffbuildings View Post
No - as im trying to explain - even if I have no cable plugged in to the interface - simply arming the channel there is a bed of white noise around -80db. This is most definitely the noise floor of the AD converters. Its this bed of white noise I want to remove, in my opinion it is one of the main reasons higain amp sim sounds 'brittle' and 'wrong'
It's not your AD's noise floor.

It's your environment's noise floor. Even without a guitar plugged in, a high impedance input will pick up a lot of RFI junk. Cell phone noise, FM/AM noise, your frdige when it switches on or off, the welding gear down the road...

To measure the noise floot of your preamp/AD, you need to plug a 150 or 220 Ohm resistance on the input, to simulate a mic. That's for mic preamps. I've never even bothered to measure an instrument input. As these are supposed to have a 1 MOhm impedance, they will be noisy when nothing is attached. And since guitar pickup impedance also varies wildly, you'd need to graph it.
__________________
In a time of deceit telling the truth is a revolutionary act.
George Orwell
cyrano is offline   Reply With Quote
Old 03-15-2017, 12:03 AM   #17
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

Quote:
Originally Posted by cyrano View Post
It's not your AD's noise floor.

It's your environment's noise floor. Even without a guitar plugged in, a high impedance input will pick up a lot of RFI junk. Cell phone noise, FM/AM noise, your frdige when it switches on or off, the welding gear down the road...

To measure the noise floot of your preamp/AD, you need to plug a 150 or 220 Ohm resistance on the input, to simulate a mic. That's for mic preamps. I've never even bothered to measure an instrument input. As these are supposed to have a 1 MOhm impedance, they will be noisy when nothing is attached. And since guitar pickup impedance also varies wildly, you'd need to graph it.
Bingo. And this is also why it's silly to obsess over ultra high spec converters. The real noise floor issues are with the electronics you connect, and with microphone recordings, the background noise in the room.

High Z guitar inputs are more susceptible to noise than low Z mic or line connections.

Think about how noisy a single coil guitar is plugged into an amp on stage. Why would it be a lot quieter plugging the same guitar into a good simulation of the same amp?

Noise gates exist for a reason.

Last edited by drumphil; 03-15-2017 at 12:12 AM.
drumphil is offline   Reply With Quote
Old 05-08-2017, 05:09 PM   #18
Philbo King
Human being with feelings
 
Philbo King's Avatar
 
Join Date: May 2017
Posts: 3,201
Default

The best way to do this is with spectral gating. You get a measurement of the noise across frequency. Then you set a gate just above the noise floor for every frequency in the spectrum. This is how noise removal is done, and it works within limits. ReaFir is a little different; instead of gating it is subtracting the noise level at each frequency. Same result.

It is not presently possible to predict the waveform of noise. It is essentially random. You can use statistics to predict it's spectral shape, but that's about it. So there is no way to come up with an "antiwaveform" that can be added to it to cancel noise.

Last edited by Philbo King; 05-08-2017 at 05:39 PM.
Philbo King is offline   Reply With Quote
Old 05-08-2017, 06:39 PM   #19
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 11,044
Default

An expander could help.

Also a low-pass filter (part of the reason high gain amps don't sound quite so nasty in real life is because the high frequencies don't make it out of the speakers).

At the end of the day, there's just no way to decouple distortion from noise. Gates, expanders and good ol' fashioned editing are your weapons of choice unless you can live with the compromise of messing with your tone by using ReaFIR.
Judders is offline   Reply With Quote
Old 05-08-2017, 07:01 PM   #20
Fergler
Human being with feelings
 
Fergler's Avatar
 
Join Date: Jan 2014
Posts: 5,205
Default

Try this out, and adjust the Max Volume of Low Pass 1's parameter modulation to match your guitar's signal strength. Mine is set to a bridge humbucker of a 335 if that gives you any idea.

https://drive.google.com/file/d/0ByE...ew?usp=sharing

What it will do (besides the static hpf and lpf which you can turn off) is filter out the high frequencies when things are quiet, then rapidly stop filtering them when you strike a note. It's strength, release time, and threshold are all set to try and match the natural sustain of a string.

The effect is that as you hold a note out, the sustain slowly loses harmonics as they drop into the noise floor. Sounds pretty smooth! When used with a regular gate set very modestly, it is really, really effective. Don't have any demos atm to show you but trust me. Well worth tinkering and learning how it behaves with your particular guitar.

Filter 2, the high shelf, is there to help boost my particular guitar's high end but it relaxes to no effect as the guitar sustains, I'd turn that off too if you don't need it.

Enjoy!

p.s. ReaFir is good as additional fixing for parts of a recorded song (don't use it live), I particularly find it useful for exposed guitar parts where a gate isn't cutting it (pun intended). But I usually automate it off once other things are playing just in case it makes anything sound funky, not in a good way. Try 4096 on best quality mode to start with (sample some noise from a 'quiet' area first, ofc). Higher sampling introduces a lot of pre-ringing and I find 4096 to be the best compromise (it is also the default).

Last edited by Fergler; 05-08-2017 at 07:07 PM.
Fergler is offline   Reply With Quote
Old 05-08-2017, 07:03 PM   #21
Tod
Human being with feelings
 
Tod's Avatar
 
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
Default

Quote:
Originally Posted by Judders View Post
At the end of the day, there's just no way to decouple distortion from noise. Gates, expanders and good ol' fashioned editing are your weapons of choice unless you can live with the compromise of messing with your tone by using ReaFIR.
Hi Judders, I'm sure you have a reason for saying this, but really and truly I have never had a problem with ReaFIR.

I think the most important thing about ReaFIR is that it needs to be set up properly, every time you use it. For example, make sure you're just getting the noise floor when you set up the Subtract Mode.

Also make sure you uncheck the "Automatically build noise profile" when you're done calibrating the subtract frequencies.

The absolute most important thing, is that your testing the actual noise you want to eliminate. If you set it up using anything other than the noise, then that's probably going to mess things up. It's very easy to miss other sounds an frequencies in the section you're analysing.

ReaFIR is a great tool, but until you understand it's characteristics, as well as when and when not to use it, you might not get along nice with it.
Tod is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 02:56 AM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.