Microphones and mic preamps

[AdiT]

I have some questions that may be clarified here.
My audio interface mic preamp has the following specifications(simplified):
max input level +10 dBu, dynamic range 110 dBA, input noise at max gain -130 dBu and gain range 50 dB.
If I set the mic preamp gain at 50 dB the dynamic range will be 80 dB or 60 dB ?
I'm asking this because I wat to know if the microphone preamp is generating noise before the mic if I use a mic with a sensitivity of -50 dB and a noise level of 25 dB, regardless of the ambient noise.

If DPA 4090 is recommended as suitable for miking aplications on studio and by some is recommanded as a ''without self noise problems'', I wonder if Neumann MA1 could be a alternative and a proper microfon for recording a violin from 0,5m.
Does anyone that has a MA1 tried to record a violin or a acoustic guitar, if so how it was the noise ?

Neumann MA1 microphone has the following specifications: sensitivity -30 dB, equivalent noise level 39 dB (highest), maximum SPL for THD 0.5% 132 dB, does this means a dynamic range of 93dB?
DPA 4090 mirophone has the following specifications: sensitivity -34 dB, equivalent noise level 38 dB (highest), maximum SPL for THD <1% 123 dB, does this means a dynamic range of 85dB?
Dates was taken from the user manual of the manufacturer.

[Hugh Robjohns]

This is a complicated topic, not least because manufacturers often fail to provide the required specifications or supply them in incompatible formats.

I don't have time to go through your numbers now but will take a look tomorrow to try and answer your questions. In the meantime this article by Michael Williams might be helpfiul:

http://microphone-data.com/media/filest ... ecs-10.pdf

[James Perrett]

I think you need to tell us the actual maximum or minimum gain as your noise figure is given at the max gain but we don't know what that gain is.

Also I'm wondering why you are looking at relatively noisy measurement mics? Something like the Neumann KM183 is around 10dB quieter than the DPA and even the humble Line Audio Omni is 5dB quieter.

[Tim Gillett]

Diaphragm size limits noise performance, whether measurement mic or not. The Neumann and DPA mics you mention appear to be 1/4". The 1/2" type will be less noisy.

[ef37a]

I shall be most interested to see Hugh's figures.

In the meantime, can I offer my limited experience with some pretty humble gear?

An SM57 into an NI KA6 needs the latter at full gain for speech or acoustic guitar at 50mm and 200mm respectively. The resultant noise is almost all 'room' noise even in my leafy suburb* at 1am!

Going even cheaper my limited tests with a Behringer UMC204HD showed the preamps were no noisier than the KA6 but had a tad more gain and so still very acceptable.

The preamps on my spanking MOTU M4 are even better. More gain and noise with dynamics is really all just the room.

But the preamp/mic is not the end of the story. There is a tendency for people to replay at above 'natural' levels, especially on cans and so the noise level, whatever its source is boosted.

*Acoustic recording at night has not really been possible since Nov 1. We can only hope WW3 is done now!

Dave.

[Hugh Robjohns]

My audio interface mic preamp has the following specifications(simplified):
max input level +10 dBu

This will be at minimum gain, but we don't know what that minimum gain is...

dynamic range 110 dBA

It's very odd to denote dynamic range with an A-weighting... but presumably they are basing the DR figure on the A-weighted noise level which is normal.

So there's a claimed 110dBA between the noise floor and clipping point in the preamp... but we don't know at what gain (it's probably at minimum gain, but we don't know). In any case, this is really not a great figure by modern standards. A figure in excess of 120dB is typical these days, even from quite modest equipment.

input noise at max gain -130 dBu

Without knowing the maximum gain, or the measurement specs (impedance, temperature, bandwidth) this figure is completely meaningless. It's also about 1dB off theoretical perfection, which doesn't seem likely given the other specs.

...and gain range 50 dB.

So is that +10 to +60dB? Or 0 to +50dB? Or +20 to +70dB? It makes an important difference to both the practicality and the specs!

If I set the mic preamp gain at 50 dB the dynamic range will be 80 dB or 60 dB ?

There isn't enough information to say, I'm afraid. You'd need to know the self-noise of the mic, the EIN of the preamp at 50dB gain, and the maximum input level at 50dB gain... none of which we know.

I'm asking this because I want to know if the microphone preamp is generating noise before the mic if I use a mic with a sensitivity of -50 dB and a noise level of 25 dB, regardless of the ambient noise.

Sensitivity of -50dB relative to what? Noise level relative to what and measured in which way? These details are critical if you want to calculate anything meaningful.

But taking a guess that the self-noise is a CCIR figure, it's a pretty noisy mic -- but not untypical of a miniature capsule. The EIN figure for the preamp above would imply a stunningly quiet preamp, but it is unreferenced and the other figures suggest it's actually a relatively noisy preamp.

If DPA 4090 is recommended as suitable for miking applications on studio and by some is recommended as a ''without self noise problems''

The DPA 4090 is intended for close miking and high sound pressure levels, so although its self-noise is quite high it isn't an issue in that kind of application.

I wonder if Neumann MA1 could be a alternative and a proper microfon for recording a violin from 0,5m.

Self noise and max SPL figures are similar to the DPA. The specs suggest it is flat within +/-2dB from 20Hz to about 15kHz, although it's not clear if that's with the applied correction data or au naturel.

If you already have an MA1 measurement mic it would be worth trying, and you'd probably find it's okay... but as a general policy, measurement mics rarely make ideal recording mics.

Neumann MA1 microphone has the following specifications: sensitivity -30 dB, equivalent noise level 39 dB (highest), maximum SPL for THD 0.5% 132 dB, does this means a dynamic range of 93dB?

Manufacturers tend to use the A-weighted noise figure when calculating DR, so that would be 27dBA, giving a dynamic range figure of 105dB. But that's based on a peak level where there is minimal distortion (0.5%)... and the mic will take more than that before it clips... but we don't know how much more. So the practical DR figure for the mic is probably greater than 105dB, but we don't know how much greater. Importantly, that figure is only realisable if the acoustic input to the mic approaches 132dB SPL.. which isn't likely from a violin 0.5m away!

DPA 4090 mirophone has the following specifications: sensitivity -34 dB, equivalent noise level 38 dB (highest), maximum SPL for THD <1% 123 dB, does this means a dynamic range of 85dB?

The max level (at a whopping 10% THD) is quoted as 134dB, and the A-wtd self noise is 23dBA (at best), giving a dynamic range of 111dB. DPA don't provide a max SPL figure for 0.5% distortion, so we can't compare directly with Neumann's figures, but it can take 123dB at less than 1% THD, giving a potential dynamic range figure of 100dB. it sounds worse... but actually, you would be likely to get quieter results from the DPA because you're never going to approach the max SPL, and the self noise is 4dB lower than the MA1.

Hope that shines some light through the fog of incomplete specs!

[Arpangel]

Going even cheaper my limited tests with a Behringer UMC204HD showed the preamps were no noisier than the KA6 but had a tad more gain and so still very acceptable.

Dave.

I can confirm this Dave, especially with my ribbons, my UMC leaves nothing to be desired, crazy, at the price.

[ef37a]

Going even cheaper my limited tests with a Behringer UMC204HD showed the preamps were no noisier than the KA6 but had a tad more gain and so still very acceptable.

Dave.

I can confirm this Dave, especially with my ribbons, my UMC leaves nothing to be desired, crazy, at the price.

Yes, Behrry's pre amps have always been quiet way back to the ill fated BCA2000. You get a lot of folks saying "Ooo! Don't buy a cheap mixer! Noisy 'king things. I have had a Xenyx 802, a Wharedale 16-4 a Xenyx 1202 and an A&H ZED10 and all have been within a dB or so for noise, i.e. bloody good. The A&H might have more headroom?
I would rather AIs have low noise and reasonable gain, 60dB say and limited headroom you can always buy some XLR 'slugs' for the drum kit! A lot cheaper than Cloudlifters.

I maintain that a far better way to specify AI 'gain' would to give a figure in mV for a specific level in a DAW e.g. -20dBFS or 0dBFS. 'we' all know that a dynamic mic puts out about 1.5mV for 1Pa some are a bit hotter, others, 7b e.g. colder.

Dave.

[Arpangel]

Going even cheaper my limited tests with a Behringer UMC204HD showed the preamps were no noisier than the KA6 but had a tad more gain and so still very acceptable.

Dave.

I can confirm this Dave, especially with my ribbons, my UMC leaves nothing to be desired, crazy, at the price.

Yes, Behrry's pre amps have always been quiet way back to the ill fated BCA2000. You get a lot of folks saying "Ooo! Don't buy a cheap mixer! Noisy 'king things. I have had a Xenyx 802, a Wharedale 16-4 a Xenyx 1202 and an A&H ZED10 and all have been within a dB or so for noise, i.e. bloody good. The A&H might have more headroom?
I would rather AIs have low noise and reasonable gain, 60dB say and limited headroom you can always buy some XLR 'slugs' for the drum kit! A lot cheaper than Cloudlifters.

I maintain that a far better way to specify AI 'gain' would to give a figure in mV for a specific level in a DAW e.g. -20dBFS or 0dBFS. 'we' all know that a dynamic mic puts out about 1.5mV for 1Pa some are a bit hotter, others, 7b e.g. colder.

Dave.

The 57 has always been a shocker as far as gain, I tried it with the Berry on a recent session, ended up using a Beyer instead, didn’t need as much gain.
But overall, the Berry amps are cool.

[AdiT]

Also I'm wondering why you are looking at relatively noisy measurement mics? Something like the Neumann KM183 is around 10dB quieter than the DPA and even the humble Line Audio Omni is 5dB quieter.

At those very small diaphragm mics I was impressed by their almost perfect omnidirectional pickup pattern up to 16 kHz.

I was looking at KM183, but I want to be sure that it worth the extra money for my needs. An alternative was Audio Technica AT 4022 but I don't know why is so rarely recommended and used. Oktava MK 012 is also a option.

But I'll do the purchase only after I'll be aware of their limitations.

[AdiT]

...and gain range 50 dB.

So is that +10 to +60dB? Or 0 to +50dB? Or +20 to +70dB? It makes an important difference to both the practicality and the specs!

With this I wanted to say that the mic preamp is capable to amplify the signal with 50 dB.

When I wrote in my first post that specifications are simplified I wanted to say that numbers are rounded, to make a easier calculation.

I want to know an example with a real situation when a mic is introducing 10 dB of noise to the recorded signal because of it's noise level and one situation when the mic preamp is adding 10 dB of noise to the signal(makes the dinamic range with 10 dB smaller) because the level from the microphone it was amplified. Please choose you the speciffications of the gear just to be apropriate to a real one. Let's not consider the room noise for a easier calculation.

[AdiT]

If you already have an MA1 measurement mic it would be worth trying, and you'd probably find it's okay... but as a general policy, measurement mics rarely make ideal recording mics.

I don't own one, but I wonder why measurements mics rarely make ideal recording as long as they have a low enough noise, are handling the SPL without distortions and have the most omnidirectional pickup patern.

[ef37a]

" Let's not consider the room noise for a easier calculation."

I am not sure what 'calculation' you are meaning but I do not think you can dismiss room noise because for us ordinary mortals it is often the 'noise floor'.

Of course, you might have an incredibly quiet room but I think the lowest ambient noise I have ever measured (on a cheap SPL meter) is around 25dBC although I suspect that is the noise floor of the meter!

And, just to add? Some active monitors are not THAT quiet either!

Dave.

[Hugh Robjohns]

...and gain range 50 dB.

So is that +10 to +60dB? Or 0 to +50dB? Or +20 to +70dB? It makes an important difference to both the practicality and the specs!

With this I wanted to say that the mic preamp is capable to amplify the signal with 50 dB.

Ah. Okay. In that case the 130dBu EIN figure is even more unlikely!

I want to know an example with a real situation when a mic is introducing 10 dB of noise to the recorded signal because of it's noise level and one situation when the mic preamp is adding 10 dB of noise to the signal(makes the dinamic range with 10 dB smaller) because the level from the microphone it was amplified.

Okay. Firstly, under normal conditions the mic isn't really "introducing 10dB of noise to the recorded signal".

A mic has an inherent noise floor due to the nature of its transducer and (in the case of a capacitor mic) its impedance converter. In a capacitor mic this self-noise can be anywhere between about 5dBA and 25dBA depending on the mics design, and particularly on the size of its capsule diaphragm -- the smaller the diaphragm the noisier the mic will be.

However, in most cases, that self-noise will be completely swamped by the ambient acoustic noise, so you can't ignore it.

Even in my well treated studio in a quiet house, in a quiet village, at the dead of night the ambient noise floor rarely drops below 25dBA, and during the day it's easily 10dB or 15dB higher than that. So unless you're trying to record fleas burping La Marseillaise the microphone self-noise is rarely a practical problem... and it definitely won't be an issue if you're close miking a violin from 0.5m away -- not even if the mic has a self-noise of 25dBA.

The mic preamp is the usual source of electronic noise and all mic preamps add a few decibels of noise, and typically more noise with more gain...

Back in the 70s and 80s 'project studio' preamps were all terribly noisy, but it's fairly rare to find duff preamps these days, and even quite modest designs can match or exceed the specs of professional standard gear from the 70s and 80s.

So if you had a preamp adding 10dB of noise you need to get a new preamp! 3-4dB of noise is more likely (the venerable Neve 1073 added about 4dB), and state of the art designs are quieter still.

The dynamic range of the recording also depends on how well suited the microphone is to the particular source and situation. Mics with very high SPL capabilities tend to have higher noise floors, and there's no benefit in recording a violin which might reach 95dB SPL with a mic that can handle 142dBA, for example. In that situation lots of the microphone's dynamic range capability would be unused and wasted.

[Hugh Robjohns]

At those very small diaphragm mics I was impressed by their almost perfect omnidirectional pickup pattern up to 16 kHz.

Sure... one theoretical benefit of tiny capsules... but do you actually require a near-perfect omni polar pattern at 16kHz? I'd very be surprised if you do.

I was looking at KM183, but I want to be sure that it worth the extra money for my needs. An alternative was Audio Technica AT 4022 but I don't know why is so rarely recommended and used. Oktava MK 012 is also a option.

There are lots of options. The 4022 is a great mic, but it hasn't been around as long as the KM183 and it doesn't have Neumann or Schoeps on the badge... The Octava is also great and you can swap out capsules, as you can with the Rode NT55 which is also very good value for money.

But I'll do the purchase only after I'll be aware of their limitations.

If you're looking for omnis (and I'd use an omni if recording a solo violin) please bear in mind that many are equalised for the diffuse field, some for the near- or free-field, some specifically for close-miking, and a very few are user-switchable (either electronically or mechanically).

This acoustic equalisation affects the high frequency response quite dramatically, and if you use the wrong type in the wrong situation the sound could be much brighter, or much duller, than you intended.

For close-miking a violin at 0.5m you need the free- or near-field type. A diffuse-equalised mic will introduce a significant HF peak which will be most unpleasant on a violin!

(You can correct this HF boost by mounting the mic so the sound arrives side on, rather than directly on axis, but that can introduce other compromises and really isn't idea!)

[Hugh Robjohns]

I wonder why measurements mics rarely make ideal recording as long as they have a low enough noise, are handling the SPL without distortions and have the most omnidirectional pickup pattern.

The best measurements mics are great for recording -- the famous DPA 4006 started out in life as a measurement mic. And there's a reason it is as expensive as it is!

However, when it comes to budget 'measurement mics' intended for simple speaker and room calibration, not so much! In a crude measurement application linearity and distortion aren't that critical because a mic's individual response anomalies can be compensated for in the alignment process via a bespoke correction file -- which is available for the MA1. That's not practical in a straight music recording situation.

[AdiT]

I wonder why measurements mics rarely make ideal recording as long as they have a low enough noise, are handling the SPL without distortions and have the most omnidirectional pickup pattern.

The best measurements mics are great for recording -- the famous DPA 4006 started out in life as a measurement mic. And there's a reason it is as expensive as it is!

However, when it comes to budget 'measurement mics' intended for simple speaker and room calibration, not so much! In a crude measurement application linearity and distortion aren't that critical because a mic's individual response anomalies can be compensated for in the alignment process via a bespoke correction file -- which is available for the MA1. That's not practical in a straight music recording situation.

After I explore the specs of DPA 4006 and compared it to other mics the only omni mics that I know equalised for close recording remains the MA1 and 4090. The unknown that remains is what frequency response MA1 has without correction.

The price difference could be gived by the materials used because 4090 is handeling the cold weater but MA1 dosen't have to.

[AdiT]

At those very small diaphragm mics I was impressed by their almost perfect omnidirectional pickup pattern up to 16 kHz.

Sure... one theoretical benefit of tiny capsules... but do you actually require a near-perfect omni polar pattern at 16kHz? I'd very be surprised if you do.

Because I need a mic for close distances, I chose the polar pattern benefices instead of a higher SPL handling.

There are lots of options. The 4022 is a great mic, but it hasn't been around as long as the KM183 and it doesn't have Neumann or Schoeps on the badge... The Octava is also great and you can swap out capsules, as you can with the Rode NT55 which is also very good value for money.

I'll choose one of these mics If I wot find the freq response without correction of the MA1 in the near future.

[AdiT]

...and gain range 50 dB.

So is that +10 to +60dB? Or 0 to +50dB? Or +20 to +70dB? It makes an important difference to both the practicality and the specs!

With this I wanted to say that the mic preamp is capable to amplify the signal with 50 dB.

Ah. Okay. In that case the 130dBu EIN figure is even more unlikely!

I want to know an example with a real situation when a mic is introducing 10 dB of noise to the recorded signal because of it's noise level and one situation when the mic preamp is adding 10 dB of noise to the signal(makes the dinamic range with 10 dB smaller) because the level from the microphone it was amplified.

The mic preamp is the usual source of electronic noise and all mic preamps add a few decibels of noise, and typically more noise with more gain...

So if you had a preamp adding 10dB of noise you need to get a new preamp! 3-4dB of noise is more likely (the venerable Neve 1073 added about 4dB), and state of the art designs are quieter still.

I found on Youtube a channel named Podcastage, where is tested a SSL 2 wich according to manual for microphone input has an EIN (A-Weighted, 150 Ω termination) -130.5 dBu and an Dynamic Range (A-Weighted) 110.5 dB. Using a 150 Ohm resistor insted of a mic after ading amplification ended up with only 70 dB of dinamic range. There is no room noise or microphone noise just a dummy load.

What caused this situation?

It's not something particular to this interface also with other is the same.

The title of the Youtube video is ''Solid State Logic SSL2+ USB Audio Interface Review / Explained (Differences between SSL 2 & SSL 2+)'' and the situations describes above is at 6:52.

[Wonks]

I found on Youtube a channel named Podcastage, where is tested a SSL 2 wich according to manual for microphone input has an EIN (A-Weighted, 150 Ω termination) -130.5 dBu and an Dynamic Range (A-Weighted) 110.5 dB. Using a 150 Ohm resistor insted of a mic after ading amplification ended up with only 70 dB of dinamic range. There is no room noise or microphone noise just a dummy load.

What caused this situation?

That figure was with a full 62dB of gain applied, so any background noise has been amplified by 62dB, plus the addition of any noise form the amplifying circuit.

Add 62 dB of gain to 70db and you get 132dB, which is basically the same as the EIN figure (measured with no gain applied).

You can't add gain without the background noise level rising as well with it. The more gain you add, the louder the interface's background noise will get.

Also think about what your actual signal level from the mic will be like if you need a full 62dB of gain. 20db to 30dB of gain is a typical maximum gain setting for most practical miking applications, often less, so your dynamic range is typically in the 90dB to 100dB region, and often more.

And have a look at where most of that noise peak is on the dynamic graph. 60Hz, the mains frequency. The rest is the more general background hiss and even at 62dB gain, that was sitting around -100dB. You may get a lot lower level of 50Hz/60Hz noise by careful placement of the interface and moving other power supplies away from it. Whilst the test is a good real-world one, it may not have been optimised for the lowest possible mains noise pickup.

Plus, for a lot of miking applications, you can simply apply a high pass filter which will significantly reduce the level of that 50Hz/60Hz noise down to around the general background noise level.

[Hugh Robjohns]

I found on Youtube a channel named Podcastage, where is tested a SSL 2 wich according to manual for microphone input has an EIN (A-Weighted, 150 Ω termination) -130.5 dBu and an Dynamic Range (A-Weighted) 11w0.5 dB.

Okay. I've not watched that specific video yet but (1) as a general rule you have to be careful as not everyone on YT knows what they're talking about. Some really do, but they are in a minority! And (2) the EIN figure is meaningless unless the gain applied during the measurement is stated too (along with the source impedance, measurement bandwidth and, strictly, the temperature). And finally, (3) the dynamic range measurement is normally taken with a line input at unity gain.

Using a 150 Ohm resistor insted of a mic after ading amplification ended up with only 70 dB of dinamic range. There is no room noise or microphone noise just a dummy load.

What caused this situation?

This is entirely normal and inevitable. The self-noise of a 150 ohm resistor is -130.9dBu (at a room temperature of 20C and with a measurement bandwidth of 20Hz to 20kHz).

If you amplify that noise floor by 60dB, say, it will be at -70.9dBu. If the mic preamp overloads at, for example, 0dBu you have a potential dynamic range of 70.9dB. If you hadn't applied any gain at all the dynamic range would have been 130.9dB.

So because the preamp's overload point is (usually) fixed, adding more gain inevitably reduces the dynamic range because it inherently raises the noise floor.

[Wonks]

Okay. I've not watched that specific video yet but (1) as a general rule you have to be careful as not everyone on YT knows what they're talking about.

It's Podcastage, which is the channel Mike Stranks often guides people to because the chap does good mic tests. Link to video: https://youtu.be/ygeCS69MM7w

The noise figures being quoted by AdiT are purely from a software LUFS meter display and real-time software frequency analyser being displayed on the screen. The presenter doesn't quote any figures himself.

But it doesn't seem to include any A-weighting to the noise, whilst the EIN figure is A-weighted, so it's certainly not comparing like for like.

[ef37a]

+ 1 Wonks to your comment about hum pickup. In any recording setup where there is a mixture of signal and mains cables there is always the potential for hum. Wall rats and line lumps are the very devil for spewing out 50Hz (or if a really poor one, at full load, 150Hz).

Even just mains cables can produce measurable hum, it is good practice IMHO to run a 'zero signal' recording once a month or so and stuff the resultant .wav through RightMark Analyser or similar. You will usually see hum peeking several dB above the noise floor and careful re-siting or/and lead dressing can reduce that quite a bit.

Can I also commend Douglas Self's Small Signal Audio Design to the OP? There you will find a wealth of real world information about noise, , its sources and measurement protocols.

Dave.

[AdiT]

The self-noise of a 150 ohm resistor is -130.9dBu (at a room temperature of 20C and with a measurement bandwidth of 20Hz to 20kHz).

Would a 50 or 250 Ohm resistance made a big difference ?

[ef37a]

The self-noise of a 150 ohm resistor is -130.9dBu (at a room temperature of 20C and with a measurement bandwidth of 20Hz to 20kHz).

Would a 50 or 250 Ohm resistance made a big difference ?

Not a lot because thermal noise rises as to the square of the resistance i.e. for a 10 times resistance increase noise goes up by about 10dB.

What WOULD make a difference is that a 250 Ohm pre amp input R would attenuate the mic signal by nearly 6dB! Plus playing ducks and drakes with the response.

Dave.