signal to noise ratio.

[Wiz]

I had a few hours to kill till my afternoon gig..and thought I would do some testing experimenting.

I was going to try and see If I could hear the differences between my preamps, regards their sounds of driving the inputs harder or softer and same on the outputs to see if I could hear the effects of transformer saturation.

But it didn't turn out as I expected.

My idea was to set up a SM57 with pop screen, so I could maintain a set distance.

Talk into the mic with the preamp set at 20,30,40,50,60,70 dB of gain, then repeat with the Input PAD set on my preamp (those that have it, JLM stuff, as it pads the input to the input transformer)

I would record each of those gain settings to a different track, then adjust the corresponding volumes of each track, so that the output of me talking, was the same volume.

The idea being I would hear A) the noise floor and how it performed vs gain of the preamp and B) any output transformer saturation. Then when I engaged the pad, I would hear how it effected the input transformer.

Seemed like a good idea at the time.

The results baffle me. Surely I have cocked up somewhere but I have been over it, 10 times.

I even went and got a 150 Ohm metal film resistor and put it across pins 2 and 3 of the mic cable connected to the preamp and stepped the preamp up through its gain range, no problems there... (will become clear why I did this soon, i promise.)

Here is how I set it up in logic



you can see as each track goes up in preamp volume, i drop the track fader by the corresponding 10 dB.

I originally did this with the gain plug, with each instance being -10db. But when things got weird, I tried the track faders instead, same outcome either way , as you would expect.

I have 48db of gain across the stereo bus.

Each track causes the stereo bus to peak at around -4dBFS, so no clipping, and the volumes are consistent.

I thought what I should hear was, each time I talk roughly the same volume, and tonal changes from the preamp gain, and transformers, etc

what i got, was a dramatic change in noise floor, which completely baffles me, I also hear tonal changes.

The preamp is a BAE 1073DMP, i tested my JLM 99NV and it behaves pretty much the same, regards the noise floor decrease as the gain increases.

Which makes me think that maybe, its not gain increase, but, ATTENUATION that is happening with the gain control on each preamp, possible?

Here is the recorded example.

I talk twice, repeating the same phrase, then lastly you will hear the noise floor recording.

This is all with the SM57 connected.

When I terminate the mic cable with a 150 Ohm metal film resistor, I don't have these noise floor issues at all. I might see the noise floor at -87dBFs for the majority of either preamp up till around 60 - 65 dB of gain , with the noise floor maxing out at around -60dBFS.

I will include two audio files, one AIF, one mp3


would love an explanation. Don't mind being called names, If I deserve it.

http://www.rentadrummer.com/wizodrop/pr ... 20test.aif

http://www.rentadrummer.com/wizodrop/pr ... 20test.mp3


cheers and thanks for any insight.

PS Rob and Rick, I am the annoying guy who phoned you both in the last couple of weeks, asking about pads and the like....this is why....


Peter

[Linear]

Interesting experiment.

I'm no tech expert but what I think you're uncovering is the noise floor of the microphone/preamp system rather than the preamp itself.

Microphones and preamps interact in very complex ways - and the design of a mic pre is really a set of compromises, with competing objectives of noise, frequency response and distortion depending on what is connected.

The mic pre 'sees' a load in the microphone and this has an effect on the performance characteristics of the mic pre. The difference between connecting a mic and connecting a resistor is the Noise Figure, measured in DB. Try a few different mics and see how this will change.

Also varying the resistor and running these tests for lowest noise will give you the optimum source impedance for the mic pre - it'll probably give you some insight into the designers objectives and what they were expecting would be connected most of the time to the preamp (They've done studies, you know. 60% of the time, it works every time...)

Chris

[rick]

Sorry I don't have the brain space to read though
This and download and understand what is what
But I am sure your learning something
Or about to ...

I read alot about transformer distortion and warmth
And stuff with preamps
Mostly I don't agree or perhaps i dont understand
Or understand too much I dunno

I think I will leave your questions
To somebody else to discuss

It's been a long month this week

[Wiz]

Thanks for the replies.

Perhaps i was verbose... But I was just trying to provide all the info I thought would be relevant...


I plan on looking into this more tomorrow.


I am just flummoxed as to which there is an apparent increase in signal to noise ratio as the gain of the premp increases.

[Linear]

Thanks for the replies.
I am just flummoxed as to which there is an apparent increase in signal to noise ratio as the gain of the premp increases.

I think you need to remember why there is a gain switch - to cope with varying input levels from the source.

What you're doing is varying the gain but keeping the source identical - then adjusting for levels of the recording. This isn't how you'd use a mic pre in the real world. I think that's why your getting freaked out about noise.

Correct gain staging is a balancing act, adjusting the pre to maximise level/minimise noise while leaving adequate headroom.

Chris

[Wiz]

To me it was the only logical way to hear the differences in tonality ...or at least attempt to hear them, by running preamps at varying gain levels. So that volume differences don't get in the way.

Perhaps I haven't explained myself very well... It's not about gain staging, per se.

Its about, preamp.transformer sound changes at different gain levels....and then me being puzzled by the resultant signal to noise ratio

[rob]

sorry, I'm to lazy on a sunday night to listen you your audio files and maybe my brain is fading, but can you sumarise your conclusions or results from this experiment?

If you have found that the noise floor doesn't increase in direct relationship to the increase in gain dialled up, I can explain that ... tomorrow!

[Wiz]

sorry, I'm to lazy on a sunday night to listen you your audio files and maybe my brain is fading, but can you sumarise your conclusions or results from this experiment?

If you have found that the noise floor doesn't increase in direct relationship to the increase in gain dialled up, I can explain that ... tomorrow!

Hey rob


Thanks for the reply.

As I increase the gain the noise floor goes down...relative to the signal. So when level matched, the SNR apparently increases. Eg at 20dB gain, there is more noise than at 60dB gain, level matched.

thanks for your help.

[Wiz]

Interesting experiment.

I'm no tech expert but what I think you're uncovering is the noise floor of the microphone/preamp system rather than the preamp itself.


Chris

I couldn't sleep this morning and got up very early...I was thinking about what you wrote Chris.

I think you are right.

I think thats it.

thanks again

Pete

[rob]

hmmm, where to start.... This is complex and it is hard to remotely account for the manner in which you did the testing.
But here are some factors:
measurement system ... your converters and DAW. The converters have their own noise floor and this is imposed on all results. How are you measuring the noise floor ... software metering? How accurate is this metering at the noise floor level?
Room noise... When I'm wanting to get a quick feel of the noise floor of a microphone I have a devil of time acoustically isolating it sufficiently for the acoustic noise level to be ruled out of the microphone electrical noise measurement. Even in an double airtight acoustically isolated box I have for this purpose, I won't get rid of the low freq noise from the environment, so if I want an accurate result I have to replace the capsule with a "dummy" head.
Preamp noise floor. Different preamp designs have a different relationship between gain and noise floor. Things like a 1073 have a constant noise floor for a number of adjacent gain settings. When you move the gain switch you are effectively adding a pad while the actual amp gain stays constant. However because you are changing the overall system gain of the pad and amplifier the amplified noise of the microphone and environment will change. But not always in direct relationship. What you really have to consider is the whole setup from mic to metering and there are a whole bunch of things contributing to the final noise floor. Some are constant, some change in direct relationship to the gain settings and some that change not as a simple one to one ratio.

See it is complex!

[Wiz]

Thanks for the reply rob... I know you are a busy guy, I appreciate it.

Also, thanks to everyone who replied.

I have a better handle on what's going on now...


Thanks all