This is more of a thought experiment than a question. Let's set aside the notion of color and I'll concede that if you want a certain tone, a certain preamp may give it to you.

My question assumes the following - you want to capture the electrical output of the microphone accurately, adding as little as possible in terms of noise or color.

Do we still need to add gain before Analog to Digital Conversion (ADC)? Or can we just adjust our ADC so that instead of setting 24 bits (clipping) to a voltage of say +24dBu, we set 24 bits to just over the maximum output voltage of the microphone?

Of course, there's a wide range of microphone output levels, so maybe we leave a gain stage in but make it all gain reduction - is that quieter than adding gain?

I'm assuming that the reason for this is that we don't have the ability to do hi resolution (bit depth) ADC at very small voltage levels cleanly, and that we end up with less noise overall by adding gain with a good preamp before ADC. Because of this, I assume a digital mic like the Neumann D-01 has a fixed gain stage between the capsule and ADC (or as part of the ADC), though their website isn't clear about this.

I'm curious to have my assumption confirmed by the folks more knowledgeable on the hardware side.

Answering my own questions a bit - Neumann does some clever stuff with dual ADCs and compression as described in this article https://en-de.neumann.com/product_files/48/download, but it's still not clear (to me, anyway) whether the analog circuits have any gain or if all of the signal processing is at the voltage levels of the capsule output.

What spawned this thread was my musing whether it would increase fidelity to have the gain adjustments on a normal interface or mixer be all gain reduction to take a signal DOWN into the optimum range for the ADC, rather than raising it into the optimum range by adding gain.

...Dead link...

When I click on it a PDF file downloads...

My favorite? The Neumann SDC in the Sennheiser clip.

But I'm funny like that.

D.

Years ago when I was actively interested in the AES42 standard, and digital mics, what I read from Neumann is that there are no analog circuits in their digital mics. That they convert to digital from the leads coming off the diaphragm, or the capsule. So gain in this sense is undefined. But so is most noise. The Neumann digital mics are still said to be eerily quiet.

To answer the question you asked in the tread title, "Do we still need preamps" the answer is no, we do not. It's not our level of technology that's holding us back, it's a cultural problem. People don't like change: "we've always done it this way" and all that.

Interesting. While I see all of the advantages in moving AD as close to the source as possible, it seems like there would be cost savings to eliminate the analog preamp on digital interfaces and consoles as well.

funny how dated the pdf looks in terms of the gear used at that time...

i do use a few neumann digital mics: they are very quiet but still need a 'digital preamp' (or at least a conversion kit), mainly to power the condenser capsules - neumann added a few more gimmicks to their preamps/interfaces that one may find useful (or completely useless).

dunno why but imo the digital version of the km184 sounds a bit different from the analog version, regardless of what preamps/converters i'm using (mostly studer).

personally, i'm not much of a fan of using different preamps to add some specific sonic characteristics; to me, the mic, the pattern, the (direct) sound of the instrument, the (reflected/ambient) sound of the room and their relative level give me enough choices.

for some odd mics, i find it interesting to play with different impedance settings but compared to the time i spend with other gear, the preamps don't get much work.

and then there are some mics that have such a high output that a preamp indeed becomes unneccessary, at least in some situations...

To answer the question in a short way, it doen't matter what the voltage level is past the point of over coming the difference in common reference impedance (when all grounds are near the same impedance). This is somewhere around -20dbfs typically. Because relative bandwidth is effected by the way they use a passive circuit to derive the signal's zero crossing reference. This is the main reason why the s/n is different with different amplitudes.

I had to sit back for a little to figure out how to post this without stepping on a lot of toes in this sector, but oh well. You see every converters I bought, I opened to see if they actually made the ADC circuit correctly, they haven't yet. They are using the "suggested circuit" which was designed to just test functionality, but technically its not correct in ac signal terms. The way the test circuit sets up the last part of the audio path, the quiescent voltage is derived internally has a very high impedance profile and that DC voltage is referenced to the zero crossing of the signal. Ideally, you want zero crossing to have a finite impedance between the other signal commons (signal grounds in laymen tech terms) in the circuit (bias +, bias-, bias gnd (if not floating/dual voltage) ).

I've attached the typical design which is in the IC datasheet. This is how almost all of the converters are built like, and its not correct. The other attachment is a schematic to a converter that is built correctly. The tech guys will see that the Q voltage (Vref to some) is infinitely referenced to one of the signal's "ground" to make the impedance of the zero crossing to be finite. So in this one, the s/n and bandwidth is constant throughout the total range (-144 dbfs).

Is there not also an issue of noise at play here: either self-noise from the capsule or residual RF noise in the surrounding air from cell phones etc ? Thus you want gain added to the mic (ideally as early in the chain as can be) so that the noise is 'drowned out' by the signal, getting the S/N ratio as high as possible ? Noise elimination (or minimization) would have to be given some consideration in all this.

The best preamps, at least for the classical type of recording work, are the ones which bring no additional noise, character or colour to the signal ie the proverbial "straight wire with gain"

If they are using the reference test design (which most do) the common mode noise is from the zero grossing "ground" to all other signal "grounds". Since the referecing is passive via a simple capacitor, the Xc of the capacitor comes into play. On the datasheets you will find a listing of s/n to signal level. This was plotted with the test circuit, however, these set of statistics are not caused by the chip itself, rather it is the measurement of the application of the IC in the test circuit.

I think I answered this.

Problem is in this sector, they don't have schematics to these interfaces/products posted anywhere. Either they want to hide the circuit from prying eyes, or they want to prevent others from copying (Which doesn't work. If they want to copy they would just simply buy it and reverse engineer it like they do other devices).

Most likely they are using "the suggested or test circuit" to build the ADC circuit.

Apparently Neumann doesn't agree

One thing that gets me, is why do they leave the phantom blocking caps in the circuit when the phantom power is not used. Dynamic mics don't need the ac coupling and this is an instance of loss and non-linear phase shift (different phase angle shifts at different frequencies) .


Must be the cheapest way to build it.

I don't understand. I'm not asking why Neumann doesn't only make digital mics.

I'm asking why RME/Focusrite/Yamaha/AH/Midas and everybody else whose products take in analog signals and immediately convert them to digital for processing eliminate the preamp stage and just put in variable gain reduction or another technique to ensure the signal is in the ADC's acceptable input range, and then do everything else in the digital domain. I can think of some reasons:

There's no significant cost savings to do so, and some real costs in user adoption to something "new" and changing habits.
It's actually more expensive to eliminate the preamps.

The Neumann Digital KM 133 d mics are just so expensive that it's difficult to tout them as a budget saving choice

I guess I didn't know what you were talking about
sorry

No worries! I know the answer to why designs haven't changed is "money". I'm just trying to understand what makes the change costly, or why it doesn't make sense.

I understand the KM d bodies are only about $250 more than the analog ones so I guess they can save money. It's just when you get to their 133's that the 'hilarity' ensues!

Yup. Which is why I lost interest in digital mics. None of the manufacturers would implement an AES42 mode 2 interface in place of mic-preamps.

My thread on AES42 mode two field mixers. My thread on the state of the art of AES42 in 2014.

Which comes first -- the chicken or the egg? The manufacturers (at least the ones I talked to, SD, Schoeps, Neumann, and a couple of others, I forget) all said in a strong unison voice that there was no demand. So they'd produced a few mics just to show they could? I'm doubtful. But if no one was willing to provide a proper AES42 interface for the mics, that's a lack of interest from the manufacturers, not the buyers. I know I lost interest because of this.

I mean really. What they were offering (AES 42 mode 1) was just a digital version of an analog workflow. All of the PITA of analog, with none of the advantages of digital. AES42 mode 2 lets you do some pretty neat things on the fly -- if you've got an interface that lets you do them. And nobody did.

Now it's true you could always buy an additional interface box from Neumann (DMI-2 I think it was). Who wouldn't want more equipment to cart around, and more cables to run? And of course, an increase in equipment costs? So, you could pay more for the mics, then have to buy interface boxes for even more money, they you have to buy a recorder with AES 42 inputs (and you end up buying the recorder's unnecessary mic-pres and redundant A2Ds because there's no way to get the recorder without them). Etc.

No demand indeed. A self fulfilling prophesy.

Crap. I let go my anger about this way back in 2014. I don't need to dredge it back up again.

[QUOTE=Bruce Watson;13567865]Yup.


Now it's true you could always buy an additional interface box from Neumann (DMI-2 I think it was). Who wouldn't want more equipment to cart around, and more cables to run?


No demand indeed. A self fulfilling prophesy.

QUOTE]

I do agree with you. The DMI-2 does takes the place of a mic pre and an ad converter and outputs a single AES3 stereo pair.

Neumann has seemed to lose a grip on a current OSX version of the program that will 'program' it.

Tangent: i had a u67 on a small guitar amp recently, and eventually dispensed with the preamp altogether once i found it required about 40dB of total padding to sound ok and match properly. Sounded better without it, and levels were plenty hot. FWIW....not the same as the average usage here...

Wouldn't STAGETEC Nexus compact be exactly what you're talking about?

Also:

SCHOEPS CMD 42 Beta - The New Digital Colette Amplifier

Also:

Microtech Gefell MV 240 USB Microphone

dude is flooding the zone with shit again as soon as 'mic' and 'digital' pop up in the same sentence...