in need of some expertise and guidance

[Paul Mailath]

I'm shooting tomorrow but as soon as I'm back I think I have to turn everything off exept the booth and I'll record into the laptop and see what happens - seems my editing setup is creating the audio room from hell

[Paul Mailath]

Originally Posted by Greg Miller

I'll bet you dollars to donuts the UPS units are the problem. Almost all of them put out square waves, which by definition are full of harmonics. They get into audio, and can even wipe out AM radio reception. Get away from those and I'll wager your noise will be much less.

that's a bugger, we tend to have brown-outs and the UPS was to supply power to the 3 RAID servers & both editing computers during those times. One of the computers records the audio.

would it be possible to put the UPS units in another room and run long extension cables to minimise the problem? (15ft further)?

[Andrew Smith]

Using a laptop (running for battery) with an SSD drive might sidestep all of these issues. You can record on that and then transfer to your RAID storage at the end of each recording session.

Andrew

[Greg Miller]

Originally Posted by Paul Mailath

would it be possible to put the UPS units in another room and run long extension cables to minimise the problem? (15ft further)?

First let me clarify what I've said before. Most UPS units do not put out clean sine wave, therefore they have lots of harmonics; however some UPS units do put out clean sine wave (thereby much less noise to worry about). We don't know which kind of units you have. (The clean ones are less common and more expensive, even so you may have these.) So this may be your only problem, or one of your problems, or not a problem at all. Just speculation until you try some tests.

A few years ago I had a real nightmare with UPS noise. I had included it here, but have decided to remove it because it isn't strictly applicable to your situation. Bottom line ... your extension cord solution may or may not work. The only way to tell is to try it. We haven't even definitely identified the source(s) yet.

I endorse Paul Johnson's approach: if possible turn off and/or unplug everything except the dBx and PC, see whether the noise is gone. If it is, then turn things on one by one until you find the culprit(s). If the noise isn't gone, then get the UPS units turned off and unplugged and see if that's the source.

I also like Andrew Smith's suggestion to use a separate laptop, battery powered and disconnected from the mains, to do your audio recording. Or, for that matter, you could use a shall portable audio recorder, again running on battery. In that case you could eliminate the M-Audio as well, just feed the dBx mic strip into the recorder. That may seem radical, but it may be the least radical of your alternatives.

Be of good cheer ... you may find just one insignificant piece of gear causing the problem. Don't forget to consider the physical proximity of the mic to all the other gear, especially display screens. That 120Hz frequency in your noise spectrum makes me suspect it is not all mains related (although that might be something picked up acoustically by the mic).

Let us know what you learn from your tests.

[Paul R Johnson]

In my experience of live sound, you see somebody coming towards you with a laptop and cable. You know this is where the show audio is coming from. They get closer and you see the Apple symbol. You breath a sigh of relief because it's 99% certain they'll plug it into the mains (another relief) and the connectors will go into your left and right and everything stays silent and hum free.

They get a bit closer and you CAN'T see the Apple symbol and you start to think about where the DI boxes were last time you saw them, but you might not need them.

Then they get closer still and you see the words of doom - Dell - on the damn thing and start looking for the one DI box of your half a dozen that might get rid of some of the noise, but of course you can unplug their power supply and hope the batteries last. They then tell you the batteries are dodgy, but don't worry as they have the power supply.

People think it unfair to single them out, but this has happened so many times, seeing the word worries me. It happens on my old Dell too. I wonder if their policy of including a check of the power supply validity on their products is partly to blame? IS it data on the DC line that allows the laptop to identify the PSU? That irritating message that says the power supply is the incorrect one, and battery charging has been disabled convinces me something apart from DC is on the way into the laptop. Maybe this makes it worse? No idea, but that awful noise going through a 12K PA system can be disturbing in the extreme. Others do it to a degree, but Dells are the absolute worst, and loads of people run off batteries for this exact reason - but nobody seems to complain to Dell. EVERY Dell I've had for years does this, and their desktop machines are quite noisy too - not as bad as laptops, but I avoid them now if I can, apart from in the office running nothing audio or video.

[Bruce Watson]

Originally Posted by Paul Mailath

My current setup is a Rode NT1-A to the DBX 286svia XLR then to the M-Audio via XLR then to the computer via USB and seems to be working okay to my tired old ears.

In my experience, electrical noise can enter electronics in a number of interesting and devious ways. The main ones are 1) through the mic cables, which is why we invented balanced XLR cables, 2) mains power, and 3) through the mics themselves.

Since you're already using XLR cables, I'm thinking you've probably taken care of #1. If you have a spare mic around that you can try in your existing chain of cables and devices, give it a try and see if your problem goes away or at least attenuates. That will quickly tell you if your mic is letting noise into the system. Since your mic is probably OK, that leaves power supply issues.

First thing to try is to turn everything off that you can -- all your extraneous computers, lights, HVAC, refrigerators, fans, etc. If the problem goes away or attenuates, start turning them back on, one at a time, testing as you go. This will help you find the culprit(s). But it won't help you deal with them really. Because the way you have to deal with them is to isolate them from your recording chain.

And the way to fully isolate them from your recording chain is to run your recording chain from a battery power supply that's giving you a pure sine wave output. Amazon lists a bunch (which probably doesn't help you a lot in Australia, but gives you an idea of what you're looking for). These "portable power stations" come in all sizes, just be sure you see "pure sine wave output" in the specs. Run your entire recording chain off your portable power station so that you're completely isolated from everything else, including your local ground.

One other thing to consider is whether the noise problem is electrical or acoustical. The mic might just be telling what it hears in the room with you. A buzzing fluorescent light can make both electromagnetic and acoustical noise, for example. But again, if you turn everything off, this problem should go away or at least attenuate.

If you still have noise problems, it's beyond what I can trouble shoot from where I'm sitting. Sorry.

[Paul R Johnson]

There are some important side issues here.
Firstly, few recording studios have to use UPS devices to remove hum and noise.

All the techniques here are trying to cure a fault, or a faulty device.

Critical factors are the fact that if we different ground potentials, then current flows. Sure, it causes noise, but it is trying to act as a protective measure to stop you dying!

Most power supplies nowadays are small switch mode devices and our UK ones have a plastic ground pin - needed for our sockets as they have a shutter fitted to prevent fingers being stuck in. Ground - the real safety ground outside tends to be connected in quite a lot of domestic buildings to the neutral line that feeds in with the live conductor. This means that there is a direct connection between the electrical power to the ground conductors in your home or studio. In the UK, adding extra grounds via rods outside (something we used to do to get a clean technical ground) is now not allowed. Multiple ground/earth points are now deemed electrically bad news.

We've covered the step by step connection system, and realistically, consider what we have? A pretty simple system. Mic, processor, interface, laptop. We have two sorts of noises - data buss noice and plain old fashioned hum. The data bus noise is coming from the computer down the USB cable then getting into the audio in the interface. The hum is probably coming ib via the DBX - but, is usually reliant on another ground connection to the system? Is the computer PSU a three conductor type with a ground? I assume it must be because hum like this needs either two ground sources, or none. If it is floating because the DBX isn't connected to real ground (two min mains cable??) nor is the computer, then you could try adding a ground connection to see if this improves the noise or makes it worse.

The data bus noise is going to the hardest to remove I suspect.

Curing that mains hum is the first thing to do. You're also compressing quite hard I note as when you speak the hum goes way down, but on the pauses, suddenly pops back in. Maybe you could back off a little on the compression, but that assumes you can find the origin of the hum.

Out of interest, what does it sound like plugged directly into the interface with the DBX unit removed from the chain completely?

[Greg Miller]

Originally Posted by Paul R Johnson

what does it sound like plugged directly into the interface with the DBX unit removed from the chain completely?

He's already provided a sample of that, and it's been discussed in earlier posts. The hum goes away, and is replaced by data buss noise. The M-Audio's preamp is picking up data noise internally; the dBx is picking up mains hum. You might want to go back and check the thread from the beginning, it will answer a few questions, but will raise some new ones.

[Paul R Johnson]

Thanks Greg. I did and just still want sure why the two issues seem to be linked or separate, and if the DBX is removing the data hash or just gating it off? Mic/interface only with and without the power supply will show if the power supply is the cause, or at least the route in. We don't know if the OP is doing the testing in the way we think he is. We've agreed on possible identification and now we are suggesting methods, but we can't advise if we're not sure of the exact circumstances. That's all. I'm not sure of the test conditions, that's all. Something simple could be in place and we don't know. If we guess it can't help.

[Paul Mailath]

Originally Posted by Paul R Johnson

Then they get closer still and you see the words of doom - Dell

laughs
looks over at the laptop
shit!!

[Paul Mailath]

okay,

I've run some more tests and recorded audio

1 - everything in the room on but recording via Audacity into the laptop

https://drive.google.com/open?id=1jk...-783VkkY5nSWmX

2 - everything in the room off except the Synology servers and the ups

https://drive.google.com/open?id=1G-...QnDZRXS9ohW43v

3 - everything off *powerpoints are still on but everything is switched off

https://drive.google.com/open?id=15R...QFUNutxYQ4sHSE

It's the quietest the room has ever been but I could still hear a faint hum and it was coming from the speakers. I switched those off and there is still a faint noise but I can't determine it's origin.

I can still the same hum when I bypass the DBX processing - buggered if I know

[Paul R Johnson]

Paul - I'd be very interested in if your Dell can have it's headphone output connected to a grounded audio system with silence. I like the computers, just the connection headaches - so if yours is silent, the model would be very handy for knowing a good one!

On the audio file front.

It's not just being off - it's connected - because the ground paths are intact even when powered down.

I know it's a pain - but what would be good would be a comparison between the totally isolated simple path - mic - interface-laptop totally on batts.
then
same thing with power connected to the laptop. You also mention speakers? Are these active types, connected to mains power, or have we got an audio amp in the system too?

I just have the feeling we're not seeing the complete audio path. The DBX unit, even if powered down has a ground path, so would an amp, but the speakers often don't.

Those harmonics on the audio - is it me, or are they wrong? Looking at them they seem to be 60Hz, 120Hz etc - but is not Australia a 50Hz country like the UK? More mysteries as we go on?

[Greg Miller]

Originally Posted by Paul R Johnson

Those harmonics on the audio - is it me, or are they wrong? Looking at them they seem to be 60Hz, 120Hz etc - but is not Australia a 50Hz country like the UK? More mysteries as we go on?

Paul, I have already mentioned that twice, in two earlier posts! My suspicion is a computer display running at either 60Hz or 120Hz scan rate. I have been very specific about this, including mentioning physical proximity between mic and any displays.

[Greg Miller]

Originally Posted by Paul R Johnson

Thanks Greg. I did and just still want sure why the two issues seem to be linked or separate, and if the DBX is removing the data hash or just gating it off?

It's obvious from the samples provided that there is NO data hash in the sample where the mic goes through the dBx (but with all processing disabled). The dBx is not removing the hash ... it is not there to begin with. There is data hash only when the dBx was out of the picture, and the mic going straight into the M-Audio. I asked Paul to provide a sample of dBx noise, in "bypass" mode, specifically for this reason. It is very obvious from that sample. This is clear from previous audio samples and earlier posts.

[Greg Miller]

Originally Posted by Paul Mailath

okay,
I've run some more tests and recorded audio

Paul, the processors change the level of your voice. Therefore they make any signal-to-noise comparisons meaningless.

Please record *everything* including your voice with the processors bypassed.