Interface self noise or ???? Puzzle

[ManFromGlass]

I have been doing some testing of a number of interfaces that I have. I am bewildered by the noise a number of EQs from different companies have been showing me. I test the interfaces with nothing plugged into them, in or out. I have a 1 year old Mac Studio. The quieter one is an RME. The noisier an El Cheapo. I am running Logic Pro. Should I be concerned?

[Hugh Robjohns]

I have been doing some testing of a number of interfaces that I have.

Okay... what software are you using to do this testing, and how are the interfaces connected/configured?

I am bewildered by the noise a number of EQs from different companies have been showing me.

I'm confused! Are you testing hardware EQs via your interfaces? Or are you using the FFT analysers in EQ plugins to provide a very dodgy form of noise measurement?

I test the interfaces with nothing plugged into them, in or out.

Not a good idea! Unterminated inputs can pick up interference and generate a lot of noise. It's better to use a 'shorting plug' or a representative source impedance to provide a meaningful noise floor.

Should I be concerned?

I wouldn't be. The EQ fft is not a calibrated measurement tool. It probably has an auto gain-ranging function which will raise the noisefloor to ensure the frequency components are visible, and it will lack the required averaging across multiple scans needed to properly assess the noisefloor.

...and the reason those EQ plots slope upwards is because they're designed to give a flat response to pink noise (and music), rather than the white noise generated by electronics!

If you want to know what your interface A/D noise floor is, use a shorting plug, and record a 10 second track then use digital gain in the DAW to pull the noisefloor up to something that you can read on the meters (-20dBFS etc). Then subtract that amount of added gain from the meter reading to get a reasonable idea of the noise floor figure.

[alexis]

...Not a good idea! Unterminated inputs can pick up interference and generate a lot of noise. It's better to use a 'shorting plug' or a representative source impedance to provide a meaningful noise floor. ...

Hi Hugh,

Thank you for another great answering post (this one to MFG's questions), one quick question echo'd back to you if I may please:

I've noticed on-stage the phenomenon you refer to above, e.g., horrible noises immediately terminated by plugging a cable into one or another of some piece of hardware.

But I never made the mental leap (microns in length?) to consider a similar phenomenon might be occurring in my home studio ... that I could be adding noise to my recordings if I don't have a cable (or a "shorting plug", is that just a dummy plug?) in every port of my interface.

Just to clarify, I don't hear "horrible noises" now through my interface with some 1/4" ports empty (well, actually I do, but those noises are mainly directly related to my recording and mixing efforts ), but I'm wondering if you're saying I'm actually adding low level noise to my recordings with every input/output port not filled with a cable or "shorting plug" ...

Thanks for any further education!

[Hugh Robjohns]

If an input is unused it should be muted or faded down (and preamp gain minimised), in which case whatever noise is coming from the front end won't go anywhere.. so no real need to worry about shorting plugs.

You might want to use shorting plugs on unused channels in particularly hostile RF environments, but that would be an exceptional situation.

Shorting plugs (in which the hot and cold terminals are linked for a balanced input, or the signal terminal grounded for an unbalanced input) are only needed when testing a system's noise performance.

The horrible noises you hear on stage are usually either because something isn't being grounded correctly until its plugged into something else, or because an input has high gain /high impedance and is using the unterminated cable as an aerial to pickup nasties via the ether... which can't get in when properly terminated!

BTW if measuring the noise floor of a mic preamp, a shorting plug will give a lower (better) figure than can really be achieved... and some manufacturers do this deliberately to make their specs look better than they really are.

To test a mic preamp the test plug should have a low noise 150* Ohm resistor strapped between pins 2/3 to represent the nominal output impedance of a connected mic.

*Other values can be used if appropriate for the preamp or measurement standard in use (eg, 30, 50, 200 or 600 Ohms).

[alexis]

If an input is unused it should be muted or faded down (and preamp gain minimised), in which case whatever noise is coming from the front end won't go anywhere.. so no real need to worry about shorting plugs....

That's great, thank you, Hugh!

Since you didn't mention output 1/4" ports (the "back end"), or line in ports, I'm guessing they don't contribute to noise either in a non-hostile RF environment (because no preamp involved in either?), so it's OK if they're unplugged as well ...

Please disabuse me of this idea of incorrect - and thank you again!

[James Perrett]

Since you didn't mention output 1/4" ports (the "back end"), or line in ports, I'm guessing they don't contribute to noise either in a non-hostile RF environment (because no preamp involved in either?), so it's OK if they're unplugged as well ...

Please disabuse me of this idea of incorrect - and thank you again!

They'll contribute noise if they are routed through to your output channel but if you get into the habit of only sending channels that you are using to the output you'll not get any additional noise contribution (unless your gear has very poor crosstalk between channels). When line inputs are concerned, the added noise is usually tiny but, technically speaking, not zero.

This is the kind of thing that engineers might have to think about when using a mixer with 50 inputs or more (very common in the days of tape) but shouldn't be an issue in a typical home setup where you are just dealing with a few inputs.

[alexis]

Since you didn't mention output 1/4" ports (the "back end"), or line in ports, I'm guessing they don't contribute to noise either in a non-hostile RF environment (because no preamp involved in either?), so it's OK if they're unplugged as well ...

Please disabuse me of this idea of incorrect - and thank you again!

They'll contribute noise if they are routed through to your output channel but if you get into the habit of only sending channels that you are using to the output you'll not get any additional noise contribution (unless your gear has very poor crosstalk between channels). When line inputs are concerned, the added noise is usually tiny but, technically speaking, not zero.

This is the kind of thing that engineers might have to think about when using a mixer with 50 inputs or more (very common in the days of tape) but shouldn't be an issue in a typical home setup where you are just dealing with a few inputs.

Perfect, thank you James!

[ManFromGlass]

Thanks all. Part of my problem was not knowing the proper questions to ask or how some plugins work measurement-wise. And I was using the dodgy method.
But again I learn from these pages!
Also as I have high frequency loss I was concerned that there were frequencies in my mixes that others could hear but I couldn’t. What I take away from this is that these EQ plugs are not an accurate test for noise in a system, although they did help me find a guitar pedal wall wart that was introducing a big jump in noise when I had the pedals plugged in. Changing the wall wart fixed that issue.

For interest - the RME has 8 balanced 1/4” in and outs with no mic pres built in.

[Hugh Robjohns]

A thought popped up, loosely related to unused inputs.

It's worth considering fitting the plastic 'bungs' which are available for male and female XLRs and quarter-inch sockets just to stop the dust getting in — especially for any unused sockets on upper surfaces.

JackCap do a good range on ebay (including for DIN and minijack): https://www.ebay.co.uk/itm/193326921343 ... media=COPY

...and Neutrik also make some for female XLRs (code NDF).

[ManFromGlass]

I never knew these existed. Similar idea like the ones for optical sockets.
I will admit to being a bit of a neat freak when it comes to maintaining gear.
I like to take the vacuum to all flat surfaces like keyboards, top of monitors and mixers etc so maybe a bit of a vacuum to sockets on the back of gear wouldn’t hurt

[Martin Walker]

A thought popped up, loosely related to unused inputs.

It's worth considering fitting the plastic 'bungs' which are available for male and female XLRs and quarter-inch sockets just to stop the dust getting in — especially for any unused sockets on upper surfaces.

JackCap do a good range on ebay (including for DIN and minijack): https://www.ebay.co.uk/itm/193326921343 ... media=COPY

Yep, I've filled all my unused sockets with JackCaps, to keep dust out

[James Perrett]

It's worth considering fitting the plastic 'bungs' which are available for male and female XLRs and quarter-inch sockets just to stop the dust getting in — especially for any unused sockets on upper surfaces.

Good idea - I can now envisage some frantic 3D printing happening next door over the next few days...