Topic: Digital Microphones and Their Advantages: Is This the Future?

[John Willett]

John - those are some sexy rig pics! You might have convinced me to start saving my pennies for a digital rig    Just wondering, how would you rate the battery life of your digital Neumanns vs. analogue ones?

BTW- I've been patiently monitoring this thread, waiting for you to chime in. Great input and insights!

Cheers

I get about 15 hours from a fully charged battery on my Nagra VI - I guess that powering two DMI-2P it will probably go down to about 7 or 8 hours.  I have never had to test it to destruction yet.

As I said earlier - I'm happy to send PDFs of my papers and a couple of other things I have on AES42 mics to anyone who PMs me with an e-mail address.

Damn! 15 OR 7 hours sounds pretty good to me. [sound of bryonsos counting pennies] One, two, three...

I should say that I have the high capacity battery option with my Nagra VI, not the standard one.

[John Willett]

--Are there any other digital mikes out there that are being used, and with what rigs?

There are over 40 different digital mic. options.  Neumann, Sennheiser and Gefell all make AES42 digital microphones as well as Schoeps (Neumann were the first).

Have any literature on that? Their site has traditionally been tough for me (as someone who has trouble with just english, let alone other languages), so it may be there and I've just missed it in the German language section/materials.

The Gefell is the MV230 measurement mic. with interchangeable heads.

PDF info is HERE.

[John Willett]

The Neumann one is 28-bits (NB: not 24) that uses two converters in a patented way to get true 28-bits.  AES42 specifies 24-bits, but the Neumann is 28-bit in the mic.

What mic has >160dB unweighted dynamic range?  And as you noted, AES is only a 24-bit protocol, but I still don't think 138dB dynamic range is achieved by a mic.  Maybe 130dB if we allow for some additional headroom above the max SPL spec.

TI markets a DAC with 32 bit output, never mind that it only has 120dB dynamic range . . .

It is nice to have dynamic range in the ADC wider than the mic because then the gain into the ADC may be fixed.  That still fits pretty comfortably inside 24 bit.

I think Neumann work on the premise of 130dB dynamic range.

[John Willett]

Have a link to the patent?  I'm curious what they did.

The easiest way to increase dynamic range is preemphasis/deemphasis, which you could implement the de-emphasis in DSP in the microphone so it's transparent to the user.  You can stack multiple converters to realize a 3dB improvement each time you double the channel count, but that doesn't work unless the dominant noise source in each channel is independent.  If the dominant noise source in two channels is common-mode you can feed them a differential signal.  I don't think either of the latter would be the case in a quality converter though.

ah, I see they used compansion by sampling the control voltage.  Clever enough I suppose although emphasis is a lot simpler, saving the need for a second converter channel.

It's in a paper by Stephan Peus - you can download the pdf HERE.

[DSatz]

All currently available digital condenser microphones are hybrids: They have the same kind of analog front-end circuitry as analog condenser microphones, but this is then followed by an A/D converter rather than a traditional analog output stage. If people would get that basic fact straight in their minds, it would immediately dispel two common, incorrect assumptions:

• that digital recordings made with digital microphones have any cause to sound different from digital recordings made with analog microphones, and:
• that recordings made with digital microphones will have wider dynamic range than recordings made with analog microphones.

Of course the analog first stage and the input of the A/D can (and should) be designed together and optimized for one another. But even with theoretically "perfect" matching, their combined performance would only equal, and could not surpass, that of an analog microphone that is optimally matched to an external preamp and A/D converter.

There's no special exemption from the laws of physics, in other words, and this fact is borne out by the noise figures of the available "digital" microphones from Neumann, Schoeps and others. The noise floor is still set by the 1/f noise of the capsule at low frequencies, and by the first analog stage at mid and high frequencies--exactly as in the analog microphones of the same manufacturers. And then the built-in A/D adds a little of its own noise, as it must. Don't be fooled by the number of "marketing bits"--look at the distance from the noise floor to the maximum level, i.e. the actual total dynamic range. It's no greater than the best analog microphone / preamp / A-to-D converter combinations already offer.

The main advantages of "digital" microphones are of two kinds: operational (simplicity of connection, plus the sound quality isn't subject to RFI or analog cable losses), or else they are useful features based on digital signal processing, e.g. the Schoeps "SuperCMIT" shotgun microphone, which uses a rear-facing second capsule and DSP to achieve a narrower, more even pickup pattern than any other shotgun microphone and to maintain that narrow pattern down to far lower frequencies than any other shotgun.

--best regards

[page]

The Gefell is the MV230 measurement mic. with interchangeable heads.

PDF info is HERE.

sexy

thanks!

[notlance]

The main advantages of "digital" microphones are of two kinds: operational (simplicity of connection, plus the sound quality isn't subject to RFI or analog cable losses)....

That is why I could never get too excited about "digital" mics; they only solve problems I don't have since I seldom encounter RFI or analog cable losses.  And old fashioned analog mics are simple enough to connect, I figure.  A "digital" mic at the current state of the art just seems to be moving the ADC from one end of the cable to the other.

[DSatz]

notlance, that is well put and a thoroughly understandable viewpoint. It's not too far from my own opinion, although I have to admit that all other things being equal, from an engineering standpoint the A/D converter does belong as close to the microphone capsule as possible.

Someone earlier in this thread hoped that Schoeps would come out with a two-channel digital amplifier that would allow a pair of their Colette capsules to be connected. That would be my preferred approach as well, since it would eliminate the synchronization problem and obviate the need for sampling rate conversion at the recorder's inputs (not that I'm afraid of a good, modern SRC, but I just don't happen to own one!). I've certainly suggested it to them.

However, even if that were available, I wouldn't spring for it unless it had greater capabilities than existing, analog microphones have. And I just don't see that yet, at least for the kinds of recording I am usually involved in. I think that we are only in the early days of DSP for audio, though, and am hoping to be pleasantly surprised some day.

--best regards

[John Willett]

The main advantages of "digital" microphones are of two kinds: operational (simplicity of connection, plus the sound quality isn't subject to RFI or analog cable losses)....

That is why I could never get too excited about "digital" mics; they only solve problems I don't have since I seldom encounter RFI or analog cable losses.  And old fashioned analog mics are simple enough to connect, I figure.  A "digital" mic at the current state of the art just seems to be moving the ADC from one end of the cable to the other.

This is what I thought until I started to use digital microphones.

Although what DSatz says is correct there are things that he did not say.

With an analogue mic. the signal travels along a cable that can pick up RF interference and noise - it then goes into a mic. pre. that adds noise of its own and amplifies the signal.  You have to allow headroom for peaks, so you back off a bit to allow for this.  It comes out of here on another cable and goes to the ADC.  Again, you have to back-off the ADC to allow headroom for peaks.

The advantages of digital mics are not in the specs compared to an analogue mic., but the fact that you get a much better s/n ratio because you don't have to all the backing off you have to do with analogue and having it in digits from the mic. keeps the signal immune from RF interference and the other noise added by an analogue system.

RF interference may not be obvious, but can be there as an increase in the noise floor.

This is explained better in the paper by Stephan Peus and in my AES papers (send me a PM with an e-mail address if you want to read them - I'll send them by return).

When I first started recording with digital mics I kept thinking we had not recorded anything - until the music started - because the noise floor was so low.

What I have been sending to those who asked are: my two AES papers, an article I wrote for LineUp magazine on my first recording session with AES42 mics, another article from LineUp (not by me) on the Abbey Road digital sessions and the White Paper that has been mentioned.  Stephan Peus's paper is downloadable from the Neumann website, but I can include this as well - so, 6 x PDFs in all, if you want them.

[DSatz]

John, this comes down to elementary issues of logic and of making fair comparisons. If a certain headroom must be allowed for peaks, then use the same yardstick (or metre stick if you will) for the digital microphone as for the chain consisting of analog microphone + preamp + A/D converter. It is misleading to compare only the worst-case analog microphone / preamp / converter arrangements with only the best-case digital microphone arrangements, e.g. by setting the digital microphone's gain for the particular SPL of the event which one is recording, while pretending that similar adjustments somehow cannot be made in a recording with analog microphones.

Say a digital microphone is designed so that full-scale digital output occurs at the onset of audible distortion in its analog first stage (e.g. somewhat above 130 dB SPL perhaps). If one never adjusts that microphone's gain via its interface/controller, its signals will probably fail to use the uppermost 15 or even 20 dB in many of the recordings that one makes with it. Thus one would prefer to increase the gain of the digital microphone somewhat, via its interface/controller. But doing so has exactly the same effects on dynamic range as raising the gain of the analog mike preamp or the input to an outboard A/D controller. Digital gain is not noise-free, nor does it magically evade headroom limits.

--best regards

[F.O.Bean]

I don't know the tech aspects of digi mics, but I don't see them becoming the "standard" for MANY YEARS TO COME Besides, we can customize a sound we want w/ analog preamps and ADC's/Recorders

[John Willett]

John, this comes down to elementary issues of logic and of making fair comparisons. If a certain headroom must be allowed for peaks, then use the same yardstick (or metre stick if you will) for the digital microphone as for the chain consisting of analog microphone + preamp + A/D converter. It is misleading to compare only the worst-case analog microphone / preamp / converter arrangements with only the best-case digital microphone arrangements, e.g. by setting the digital microphone's gain for the particular SPL of the event which one is recording, while pretending that similar adjustments somehow cannot be made in a recording with analog microphones.

Say a digital microphone is designed so that full-scale digital output occurs at the onset of audible distortion in its analog first stage (e.g. somewhat above 130 dB SPL perhaps). If one never adjusts that microphone's gain via its interface/controller, its signals will probably fail to use the uppermost 15 or even 20 dB in many of the recordings that one makes with it. Thus one would prefer to increase the gain of the digital microphone somewhat, via its interface/controller. But doing so has exactly the same effects on dynamic range as raising the gain of the analog mike preamp or the input to an outboard A/D controller. Digital gain is not noise-free, nor does it magically evade headroom limits.

--best regards

An AES42 digital mic. has very little analogue in it at all - basically the capsule and an fet and not much else.  This is then fed directly into the DAC.

There is no need to adjust the gain at all as it it digital gain and will not affect the quality; though I do adjust the gain as it makes monitoring easier.

But, every bit of analogue circuitry adds its own little bit of noise and the more there is in the analogue path the more noise there is - OK, on a good system this can be very small, but it's still there.

I am not saying that digital mics are perfect and analogue mics are rubbish - there is room for both and will be for many years.

But digital microphones have great advantages and the more I use them the more I like them.

I have been using AES42 digital microphones in anger for 5 years now, so I am speaking as someone who uses them, has read up on them, had long discussions with the designers of them (Neumann, Schoeps, Sennheiser and Gefell) and written papers on them (and before my papers were published they were sent to Schoeps, Neumann, Sennheiser and Gefell for review so they could make comments/corrections before I presented them).

People tend to forget all the little bits in the analogue chain that slowly decrease the specs of an analogue mic. that does not happen with an AES42 mic.

And, yes, I know it was you who translated the White Paper that has been mentioned previously - it's a good paper, though I don't agree with all that is said in it.

[John Willett]

I don't know the tech aspects of digi mics, but I don't see them becoming the "standard" for MANY YEARS TO COME Besides, we can customize a sound we want w/ analog preamps and ADC's/Recorders

Probably quite right.

But there are many advantages in digital microphones over analogue ones that, for some applications, make them a better choice.

And you can customise a sound in a DAW if you want to.

Personally I see AES42 digital mics will probably mainly be used in the classical arena where you are attempting to capture a performance as perfectly as possible.

Though there is a PA company in France that is all digital from the mics to the power amplifiers.  They find that the sound-check time is drastically reduced and the hiss and noise out of the loudspeakers is so quiet that they forget the system is on; which happened when they went for a pre-show meal at the café round the corner and left the system on.  Someone started up a motorbike by the stage and there was a very loud rendering of this sound through the system.

[SmokinJoe]

Correct me if I'm wrong... a 24bit A/D can basically break an analog signal into 2^24 increments across it's full range.  With the A/D in the mic, if full scale is 130db and you are recording quiet acoustic material than only goes to 80db, you are failing to optimize a big portion of your bandwidth.  Digital gain is like digital zoom on a camera... it makes it seem closer (louder) but it doesn't enhance resolution.  If you recording is all at -50 on the meter and you have to add that all in post, that's digital gain too.  With an analog preamp chain you can better optimize the full range coming into your A/D.

Now if you can set something back at the control box/recorder to tell the FET front-end to increase analog gain before the mic's A/D, that's completely different, but I didn't get the impression it works that way.

If you tell me that resolution issue is peanuts compared to the overall increase in signal accuracy of a much shorter analog chain, I can't argue with you on that point.

[John Willett]

Correct me if I'm wrong... a 24bit A/D can basically break an analog signal into 2^24 increments across it's full range.  With the A/D in the mic, if full scale is 130db and you are recording quiet acoustic material than only goes to 80db, you are failing to optimize a big portion of your bandwidth.  Digital gain is like digital zoom on a camera... it makes it seem closer (louder) but it doesn't enhance resolution.  If you recording is all at -50 on the meter and you have to add that all in post, that's digital gain too.  With an analog preamp chain you can better optimize the full range coming into your A/D.

Now if you can set something back at the control box/recorder to tell the FET front-end to increase analog gain before the mic's A/D, that's completely different, but I didn't get the impression it works that way.

If you tell me that resolution issue is peanuts compared to the overall increase in signal accuracy of a much shorter analog chain, I can't argue with you on that point.

There is virtually no analogue at all in an AES42 digital mic.  -  you have the capsule and the FET ad very little else.  This is fed directly to a dedicated ADC that is matched to the capsule.  The capsule is, in effect, driving the ADC directly.

There is no line driver stage in the mic..  Everything is taken away and the capsule goes directly to an ADC that is designed and matched to the capsule - in the case of Neumann it's a 28-bit ADC.

I know it can be difficult to get your head around, but you need to stop thinking like it's a normal analogue microphone.

As I said, I am very happy to send PDFs of info I have to anyone here who sends me a PM with an e-mail address - and most of these PDFs are very readable and not filled with difficult to understand technical stuff.