anechoic chamber - whispering - ESL

[plodsmeade]

Hi all

Excuse the length of this.

I have a slightly unusual project for which I will be recording a person whispering. The recording will be done in an anechoic chamber, and is intended for playback on an electrostatic loudspeaker. In terms of noise the weak link in the chain is probably the speaker and the playback environment, but the latter may well be a very quiet place (including the possibility of the chamber itself) and I would like to give the speaker the best chance I can. So I have been researching the best way to maximise the strengths of these recording and playback scenarios with the choice of mic and recording device, and indeed the best way to do the recording.

So for the mic, I need:

1. low noise
2. high sensitivity
3. low distortion
4. very fast transient response
5. extended HF response

1 and 2 suggest a large diaphragm condenser.
3 - 5 would suggest a small diaphragm condenser.
3 and 4 are probably most important in terms of maximising the capabilities of an ESL.
I’m not totally sure about 2: my thinking is that less gain will be required from the preamp and so there will be less noise, but I have doubts.

I have been looking at this list of microphones ordered by self noise. I can can get hold of a Rode NT1-A or NT2-A which seem very impressive as far as 1 and 2 go above. But I can also rent something like the Sennheiser MKH20 or Schoeps MK2+CMC6, or perhaps something like the DPA 4006A which is not on that list (15dB(A) self noise and 40mV/Pa sensitivity, very good). All of these would be superior in terms of 3 - 5 above.

One answer might be to use many mics on the source and decide later, which brings me to the recording device. I have a Tascam US600 which I aways thought sounded pretty noisy, and looking at the specs gives a fairly vague SNR of '98dB or more(LINE IN to LINE OUT with Gain knobs at minimum, 44.1kHz, JEITA)’. I have been looking at renting something better and have found RME Fireface800s are available (Dynamic range AD: 109 dB RMS unweighted, 112 dBA). But it occurred to me that perhaps a high-end film industry portable recorder would be better, such as the Sound Devices 633 (A/D Dynamic Range 114 dB, A-weighted, typical). The specs are similar and as I might have to have the recorder inside the chamber, I can leave the laptop outside and choose a noiseless device.

For the recording, I need to be wary of plosives and probably quite big sonic differences in the whispering off-axis from the mouth, as the mic will need to be fairly close. I suppose that because of the chamber and the whispered source the consequences of polar pattern choice will be limited to proximity effect, which should be minimal anyway with whispered sound, but perhaps I am wrong here - my hunch would be to go omni to keep everything natural and make positioning easier.

So to cut a long story short, and because I’m even boring myself now, if anyone has any advice or suggestions with regard to:

mic choice
recorder choice
mic placement methods
multi-miking strategies for a single close source
whether in this instance 192kHz might be worth it
how best to analyse a recording for 1, 4 and 5 above
recommended hire companies in and around London

I would be very glad to hear it.

cheers

[Hugh Robjohns]

So for the mic, I need:

1. low noise
2. high sensitivity
3. low distortion
4. very fast transient response
5. extended HF response

1 and 2 suggest a large diaphragm condenser.
3 - 5 would suggest a small diaphragm condenser.

Yes, physics can be frustrating, can't it! and point 5 would demand a miniature diaphragm mic, too!

However, I would suggest (from practical experience) that the best option would be a Sennheiser MKH20. It has pretty low self noise, high sensitivity, extremely low distortion, good transient response and a decent bandwidth.

... and looking at the specs gives a fairly vague SNR of '98dB or more

The problem you have here is a confusion of specs. Any decent converter will manage a dynamic range of at least 115dB. That's not the quality-limiting factor. The mic preamp, on the other hand will be a potential weak point as far as noise is concerned. As you say, a high output mic (like the MKH20) will help. RME interfaces are generally pretty good. The Sound Devices stuff is good, but designed for robustness and handling difficult signals, rather than the ultimate in low noise. I presume the anechoic chamber has mic tie lines to the outside world, so there shouldn't be an issue recording to a laptop outside.

For the recording, I need to be wary of plosives and probably quite big sonic differences in the whispering off-axis from the mouth, as the mic will need to be fairly close. I suppose that because of the chamber and the whispered source the consequences of polar pattern choice will be limited to proximity effect, which should be minimal anyway with whispered sound, but perhaps I am wrong here

You're wrong here! Proximity effect Is determined by the source proximity to a pressure gradient mic. The clue is in the name. The volume of sound -- whether spoken or whispered -- plays no part.

However, you are right than an omni mic would be the best choice -- hence my suggestion of the MKH20. Pressure operated (omni) mics don't suffer from proximity effect, and are far less susceptible to plosives too. But you also have the advantage, in an anechoic chamber, that the critical distance is huge -- meaning you can place the mic at a comfortable distance from the performer without compromising the sound perspective.

Dare I ask what the idea of this recording/playback is all about?

H

[John Willett]

For your application, the only mic. I would look at is an AES42 digital mic.

The noise floor on these is really so very low that the first time I used them I thought they were not working as the noise floor was so low - until the music started.

For your use I would look at the Neumann D-01 or, if that is too expensive, the TLM 103-D or the KM-D series.

If you are recording in the field, then the AETA 4MinX recorder has direct AES42 inputs as does also the Sound Devices 788T. Or you can record directly into a laptop by using a Neumann adaptor, though the Neumann DMI-2 or DMI-2p give more control.

The Sennheiser MKH 8000 series with the MZD 8000 is also an AES42 mic., but the Neumann has the ADC right after the capsule and, uniquely, the Neumann uses a special 28-bit converter for the conversion.

I wrote up my first sessions with the Neumann KM-D series in these forums <a href="/forum/showflat.php?Cat=&Board=MRT&Number=425894&Searchpage=1&Main=425894&Words=neumann+km-d&topic=&Search=true#Post425894" target="_blank">HERE</a>.

The reason the noise floor is so low is that the ADC is right at the capsule so there is virtually zero analogue noise. You don't have to back-off the recording level to allow for headroom (which adds noise) and, also, there is none of that low level RF noise that gets in. In normal recording on a good system, most of this is too low level to worry about - but, in your case, an AES42 microphone will eliminate this and is, IMHO, the ony real solution for what you want to do.

I have written a couple of AES papers on AES42:-

Digital Microphones - what's it all about?

Digital Microphones - AES 42 and all that

NB: Please note that these links will be switched off at the end of April 2016 as "Copy" are discontinuing the service (I will then have to ind a new site for them).

I am happy to answer questions if you want to send me a PM.

PS - I have just noticed that Hugh Robjohns is recommending an omni. ALthough I agree with him and the MKH 20 and the Gefell M221 are my two favourite omnis - in your case I would put the Neumann KM-D series KM 131-D as first choice with the Sennheiser MKH 8020 with MZD 8000 as second choice.

[Urthlupe]

I have written a couple of AES papers on AES42:-

Digital Microphones - what's it all about?

Digital Microphones - AES 42 and all that

NB: Please note that these links will be switched off at the end of April 2016 as "Copy" are discontinuing the service (I will then have to ind a new site for them).

I am happy to answer questions if you want to send me a PM.

Thankyou John - those really were a useful read for me.

Loopy

[Hugh Robjohns]

For your application, the only mic. I would look at is an AES42 digital mic.
The noise floor on these is really so very low....

Yes, good call. If the OP is going to hire stuff anyway, why not?

H

[John Willett]

For your application, the only mic. I would look at is an AES42 digital mic.
The noise floor on these is really so very low....

Yes, good call. If the OP is going to hire stuff anyway, why not?

H

Well, I'm close enough to the West Midlands to help if he needs it - I have several digital mics.

[The Red Bladder]

Dare I ask what the idea of this recording/playback is all about?

I bet it's something completely bonkers!

It sounds like one of those gigs one used to get in the 70s, where some mad woman would come into the studio, perfectly convinced that if we had a microphone sensitive enough, we would be able to record the voices in her head.

In the 70s, I recorded some nut-job guru drinking herbal tea, who wore a purple dress and smelled of a heady mixture of dope, incense and urine. But my career reached a new zenith, with the seminal recording of a man eating celery in 4.0 surround sound.

[Tim Gillett]

Just a comment on the nature of the source here. Whispering, having no vibrating vocal chord component, doesnt have a great dynamic range. While the sibilants may be at normal speaking strength, the vowel sounds, normally produced by the vocal chords, will not nearly match the volume of normal speech or singing, let alone a shout.

I'd have thought a huge dynamic range recording chain is not needed here. Low noise of course, but that's about all.

[Tim Gillett]

Cant top that one but I did know a customer who was asked to consult on the most suitable microphone to record the sound of a pig's carcass rotting...

[MarkOne]

Cant top that one but I did know a customer who was asked to consult on the most suitable microphone to record the sound of a pig's carcass rotting...

Inspired by Yoko Ono:

“We were doing a song, ‘Dead Rat,’ on Yoko Ono’s Approximately Infinite Universe. I had no idea what it was going to be. The music was like, ‘da da da da da da da da da da,’ and then there’d be five seconds of silence. Then the music would start again, and there’d be another five seconds of silence. Finally, I found out what this silence was. It was the dead rat solo. Yoko brings a shoe box into the studio, and inside the shoe box is this freshly killed rat, size large. She says, ‘Okay, you know where the band stops playing? That’s where the rat takes it.’...I put the rat on a stool, and I had my assistant put a [Neumann] 87 over the rat. I ran the tracks, the music stops, the rat takes it. Nothing happens. I turn around and say to my assistant, ‘You [ ****** ], get out there and put the mic in the right place. About four inches up and a little to the left.’ So he goes out and moves the mic over the rat a little bit. We start the music and it gets to the rat solo. I turn around and say, ‘What do you think, Yoko?’ She says, ‘It’s much better like that.’ And she’s dead serious.”—Jack Douglas, January 1991

[Hugh Robjohns]

I'd have thought a huge dynamic range recording chain is not needed here.

I don't think 'huge dynamic range' was listed as one of the OP's original requirements. It came up in passing as an equipment specification that might give some clue as to the achievable signal-noise ratio, which it does, of course (although not as a limiting factor in this situation), and in the context of the benefits of digital mics.

H

[Tim Gillett]

I'd have thought a huge dynamic range recording chain is not needed here.

Correct... but I don't think 'huge dynamic range' was listed as one of the OP's original requirements, anyway.

No, quite correctly it wasnt listed. Rather I was responding to what John wrote which I must be misreading, but maybe that's for him to clarify if necessary.

[Exalted Wombat]

Dare I ask what the idea of this recording/playback is all about?

In the 70s, I recorded some nut-job guru drinking herbal tea, who wore a purple dress and smelled of a heady mixture of dope, incense and urine. But my career reached a new zenith, with the seminal recording of a man eating celery in 4.0 surround sound.

Speaking of seminal recordings, I did a few things for a guy called George Harrison Marks.

It all ended. There was going to be a riverboat gig. On a boat called the Marchioness. Then George died...

[ManFromGlass]

I didn't participate in the recording session but a friend of mine had her internal body sounds recorded from the inside by an experimental music composer. I didn't get to hear the final composition. This was in the 80s.
(What images are going through your minds right now?)

[plodsmeade]

Hi all, thanks for the responses, really great.

Proximity effect Is determined by the source proximity to a pressure gradient mic. The clue is in the name. The volume of sound -- whether spoken or whispered -- plays no part.

I was thinking in terms of frequencies within the source, i.e. since there are no voiced sonorants in whispering, the low-end boost from proximity effect might not be relevant. But perhaps there will be breathiness which is broadband enough for it to be an issue.

For your application, the only mic. I would look at is an AES42 digital mic.

I had heard about these but had no idea the situation was so advanced. Thanks for the link to those papers. It seems like the perfect solution.

Dare I ask what the idea of this recording/playback is all about?

I bet it's something completely bonkers!

Not quite bonkers exactly but you are all on the right track - it's for an art work, involving a 6ft ESL monolith, through which is played the whispered voice I mention. There are other reasons for wanting an ESL but I thought it best to take advantage of their properties to make the voice as 'present' as possible without using headphones, and it seems to me that I've got a good shot this way. I don’t know how familiar you are with ESLs, but this will be my first build so fingers crossed.

[ConcertinaChap]

ESL? English as a Second Language

[plodsmeade]

ESL? English as a Second Language

Sorry, Electrostatic Loudspeaker

[ConcertinaChap]

Thanks.

CC

[Tim Gillett]

For your application, the only mic. I would look at is an AES42 digital mic.

I had heard about these but had no idea the situation was so advanced. Thanks for the link to those papers. It seems like the perfect solution.

I beg to differ with John's take on this one. Low noise might be important for recording whispers but low noise is not peculiar to digital mics. Many conventional mics also have low noise. If you were running hundreds of metres of mic cable before the preamp, a situation where degradation of the signal eventually occurs it MIGHT be a factor but in your situation that seems unlikely to be a requirement.

[John Willett]

For your application, the only mic. I would look at is an AES42 digital mic.

I had heard about these but had no idea the situation was so advanced. Thanks for the link to those papers. It seems like the perfect solution.

I beg to differ with John's take on this one. Low noise might be important for recording whispers but low noise is not peculiar to digital mics. Many conventional mics also have low noise. If you were running hundreds of metres of mic cable before the preamp, a situation where degradation of the signal eventually occurs it MIGHT be a factor but in your situation that seems unlikely to be a requirement.

Have you *used* an AES42 digital mic?

Yes, I know good analogue mics can be quiet, but taking analogue out of the equation altogether does make an obvious difference.

I was amazed the first time I used AES42.

Especially in this case where the recording will be in an anechoic chamber - having just the capsule and FET before the ADC would be an advantage.

[John Willett]

Dare I ask what the idea of this recording/playback is all about?

I bet it's something completely bonkers!

Not quite bonkers exactly but you are all on the right track - it's for an art work, involving a 6ft ESL monolith, through which is played the whispered voice I mention. There are other reasons for wanting an ESL but I thought it best to take advantage of their properties to make the voice as 'present' as possible without using headphones, and it seems to me that I've got a good shot this way. I don’t know how familiar you are with ESLs, but this will be my first build so fingers crossed.

Now this *does* sound interesting.

I hope there will be space round the ESL as these generally are fig.8 and radiate as much from the rear as they do from the front.

[Tim Gillett]

For your application, the only mic. I would look at is an AES42 digital mic.

I had heard about these but had no idea the situation was so advanced. Thanks for the link to those papers. It seems like the perfect solution.

I beg to differ with John's take on this one. Low noise might be important for recording whispers but low noise is not peculiar to digital mics. Many conventional mics also have low noise. If you were running hundreds of metres of mic cable before the preamp, a situation where degradation of the signal eventually occurs it MIGHT be a factor but in your situation that seems unlikely to be a requirement.

Have you *used* an AES42 digital mic?

It would be nice but no, I havent had that opportunity yet.

Yes, I know good analogue mics can be quiet, but taking analogue out of the equation altogether does make an obvious difference.

I was amazed the first time I used AES42.

Especially in this case where the recording will be in an anechoic chamber - having just the capsule and FET before the ADC would be an advantage.

Do we know how much quieter is say the TLM 103 D compared to the TLM 103 into a good low noise pre and ADC?
Not merely a subjective impression but formally conducted, standardised, falsifiable test measurements? How do the noise levels of the two examples stack up in a real world comparison?

[ConcertinaChap]

Do we know how much quieter is say the TLM 103 D compared to the TLM 103 into a good low noise pre and ADC?

Not merely a subjective impression but formally conducted, standardised, falsifiable test measurements? How do the noise levels of the two examples stack up in a real world comparison?

I'm sure if you look for this data, Tim, you'll find it.

CC

[John Willett]

The better noise level of an AES42 mic. is not so much in the mic. itself, but from the fact that you are not adding any analalogue noise and the fact that you don't have to "back-off" to allow for headroom.

As explained by Neumann's Stefan Peus in his AES paper and also in my AES papers (diagrams used with the permission of Stefan Peus).

[Tim Gillett]

Thanks John. Can you supply a link to the Stefan Peus AES paper?

Cheers, Tim