Any suggestions for papers on low noise signal combining?
Steve
Assuming they are the bufferless version, aren't they all simply in
parallel?
...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
Help save the environment!
Please dispose of socialism properly!
** Likely all the mics are simply wired in series and supplied with power
from a DC current source and the signal taken from the top most mic. The s/n
advantage comes from *correlation*.
The 10 noise signals ( from the internal JFET pre-amps) are un-correlated
while the sound arriving at each co-sited mic will be heavily correlated.
So the output signal will be 10 times that of one mic, while the self noise
will be only 3.16 times - so a 10 dB improvement is possible.
I suspect the mics are arranged to create a directional pick up pattern too.
BTW
Dynamic mics have much less self noise than typical electrets - cos there is
no internal pre-amp. Means it is easy to get a 10dB or more improvement,
just by using one instead.
..... Phil
o...@uakron.edu wrote:
> My boss laid a unit on my desk today. Its got ten electret mics fed
> into one low noise opamp. Any ideas how they did this without a ton
> of noise. All I can see from the board layout is it is NOT the classic
> op amp summer.
Electret mikes usually have built-in FET stage. Are you sure they are
combined in parallel, not in series?
Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com
They are available WITHOUT the FET. I have an RFQ on my desk
requesting a complex chip that will use Electrets sans FET's.
>On Wed, 16 Dec 2009 17:58:19 -0600, Vladimir Vassilevsky
><nos...@nowhere.com> wrote:
>
>>
>>
>>o...@uakron.edu wrote:
>>> My boss laid a unit on my desk today. Its got ten electret mics fed
>>> into one low noise opamp. Any ideas how they did this without a ton
>>> of noise. All I can see from the board layout is it is NOT the classic
>>> op amp summer.
>>
>>Electret mikes usually have built-in FET stage. Are you sure they are
>>combined in parallel, not in series?
>>
>>
>>Vladimir Vassilevsky
>>DSP and Mixed Signal Design Consultant
>>http://www.abvolt.com
>>
>
>They are available WITHOUT the FET. I have an RFQ on my desk
>requesting a complex chip that will use Electrets sans FET's.
Right. Some of ours come with FETs, some not. Some "dynamics" are
really electrets in drag.
** So what ??
The kind with inbuilt JFET pre-amps are about a million times more common.
PLUS, to comply with the OPs post they cannot be naked condenser capsules.
.... Phil
Steve
Steve
** Totally stupid bullshit.
... Phil
** Be nice of YOU to tell us what the fucking thing is for - pal !!!
But it sounds a lot like a simple "beam forming" microphone array. These
are hot stuff now for use with hands free phones and also elaborate
conference mic arrays that can be steered remotely to pick single voices out
of an audience.
These sorts of directional mic arrays certainly do reduce " noise " - but
it is ** AMBIENT ** noise pickup, not self noise !!
In any case, YOU are 100% WRONG in your assumption that there ought to
be a noise problem when combining the mics signals. Wired either in series
or parallel, correlation still works to improve the s/n ratio.
.... Phil
If this is work, it should probably stay there, for the benefit of
your employers and clients.
RL
Mixer front ends still add thermal noise, "pal", and yes, its a
conference mike.
Steve
>
>"krw"
>>
>> Some "dynamics" are really electrets in drag.
>
>
>** Totally stupid bullshit.
>
In fact, you are. You and DimBulb; a perfect pair.
>>>
>>> Some "dynamics" are really electrets in drag.
>>
** Totally stupid bullshit.
Why don't ya go drop dead -
you VILE LYING pile of sub human GARBAGE.
.... Phil
** Wot an over- snipping cunt.
>> In any case, YOU are 100% WRONG in your assumption that there ought to
>> be a noise problem when combining the mics signals. Wired either in
>> series
> > or parallel, correlation still works to improve the s/n ratio.
>
> Mixer front ends still add thermal noise,
** Not to electret mics with NOISY internal jfet pre-amps - fuckwit !!
And not at ALL when the summing is done PASSIVELY prior to the gain
stage.
You KNOW NOTHING !!
> "pal", and yes, its a conference mike.
** So what was the BIG FUCKING secret about that ?
Why not post a link to the stupid contraption ??
.... Phil
What sort of 'unit' ?
> Its got ten electret mics fed
> into one low noise opamp. Any ideas how they did this without a ton
> of noise.
Why do you think it would make the unit noisy unless designed by a
cretin ? Plenty of them about of course ! ;-(
> All I can see from the board layout is it is NOT the classic
> op amp summer.
You mean the inverting 'virtual earth' type ? Yes, avoiding that
structure *would* avoid excess noise.
> Any suggestions for papers on low noise signal combining?
Join the AES. Or come round and look at my data library which has some
excellent app notes / books on that and similar subects.
Graham
Very likely.
> Dynamic mics have much less self noise than typical electrets - cos there is
> no internal pre-amp. Means it is easy to get a 10dB or more improvement,
> just by using one instead.
But dynamic capsules cost more !
Graham
Quite so.
>
> They are available WITHOUT the FET. I have an RFQ on my desk
> requesting a complex chip that will use Electrets sans FET's.
>
> ...Jim Thompson
How many Gigaohms input impedance will the complex chip have ?
Graham
Uh ? What's a 'mixer front end' ? Obviously you're not an audio
professional and haven't a clue about current practice. You can make a
mix stage today using inexpensive off the shelf parts that has 100dB or
more dynamic range quite easily.
So what's the mix stage chip ?
Graham
Provide a link to one of these 'electret dynamics' or SHUT UP !
Graham
>
> But dynamic capsules cost more !
** But not 10 times more - you wally.
.... Phil
Not exactly frequency shaping, but probably time delay via networks.
You need delay to do beam forming. It is easier in radar since the
bandwidth is narrow. For audio, you are much better off using DSP
techniques where delay comes for free in the sampling.
I looked into doing a product with multiple microphones years ago. The
idea was not to do beam forming, which merely steers the signal, but
zone forming. It is possible with DSP to narrow the pickup to a zone
rather than have the beam go out to infinity.
The problem with beam forming is it is a patent mine field. As you
probably know, half the patents litigated are proven to be invalid,
but the cost to litigate makes developing the project unfeasible for
the small investor. Even a large Japanese firm I was dealing with
didn't want to do it.
It would be interesting to see how Polycom got their project to
market.
The issue is more related to ESD protection on the chip. You can make
a very high impedance amplifier in CMOS, but it would be a touchy
product in assembly. I don't recall running into a BICMOS with JFETs,
but that would be a better technology.
Just don't try to seperate them. You know what happens when you try
to divide zero by zero.
--
Offworld checks no longer accepted!
Nope, My experience is with 1980s audio when every open channel was
a problem.
I'm glad to admit that. I'm a laser/electro-optics guy.
Any specfic AES paper?
Steve
o...@uakron.edu wrote:
> My boss laid a unit on my desk today. Its got ten electret mics fed
> into one low noise opamp. Any ideas how they did this without a ton
> of noise. All I can see from the board layout is it is NOT the classic
> op amp summer.
>
> Any suggestions for papers on low noise signal combining?
>
> Steve
Crazy idea, Is there anyway to add them in series?
George H.
Phil, you're great! How much does a good Dynamic mic cost and who
makes it? Say in the $10 - $20 range.
George H.
Is there any mic switching or multiple acquisition paths? A good
conference mic would be directional in a bunch of directions and
select the dominant-signal mic, excluding the noise from other
directions.
John
They could just be in parallel. That would work if they were raw
electrets or if they were the more common electret+jfet.
But it sounds more complicated.
John
Wow, a trio of DimBulbs. We've hit the mother load!
You get a DimBulb Donkey? <shudder>
> How much does a good Dynamic mic cost
** The topic was mic *inserts*.
Do your own Googling.
... Phil
** Totally stupid bullshit.
Or a black hole.
That would be "mother lode" actually. Can't you even spell ?
http://www.askoxford.com/concise_oed/motherlode?view=uk
You're the DimBulb here.
Dynamic and Electret have specific and very different scientific
meanings and are constructed totally differently.
You may have seen an electret mic *marketed* as dynamic by some bunch of
half-wits but that doesn't make it valid.
Graham
In this instance to do what was required would require 10 capsules of
whatever type - you wally.
Graham
So, you're as thick as 'krw' are you ?
Graham
>>> But dynamic capsules cost more !
>>
>>
>> ** But not 10 times more - you wally.
>
> In this instance to do what was required..
** Was NOT in evidence when I posted my comment.
" BTW
Dynamic mics have much less self noise than typical electrets - cos there is
no internal pre-amp. Means it is easy to get a 10dB or more improvement,
just by using one instead. "
You fucking CONTEXT SHIFTING MANIC POMMY ASS.
... Phil
Dynamic mikes have a heavier diafragm, and because of that are less sensitive.
So you may gain at one point, but have less signal to process, still more noise.
This is *my* view, IIRC Phil has a different idea about that.
>Graham
>
> Dynamic mikes have a heavier diafragm, and because of that are less
> sensitive.
** Yaaaaaawwwwwwnnnnnnnnnnnnnnnnnnnnnnnnnnnn.....
> So you may gain at one point, but have less signal to process, still more
> noise.
** Supposition - not fact.
Spend a few hours ( or several years in my case) studying makers spec
sheets or better still the real things.
Shut the FUCK UP till you have !
Wog IDIOT !!!!!
..... Phil
>
>"Jan Panteltje is a Wog Wanker "
>
>> Dynamic mikes have a heavier diafragm, and because of that are less
>> sensitive.
>
>
> ** Yaaaaaawwwwwwnnnnnnnnnnnnnnnnnnnnnnnnnnnn.....
>
>> So you may gain at one point, but have less signal to process, still more
>> noise.
>
>** Supposition - not fact.
I have tried the dynamics.
And the electrets.
>
> Spend a few hours ( or several years in my case) studying makers spec
>sheets or better still the real things.
>
> Shut the FUCK UP till you have !
Well, how about studying your prescription?
LOL
> > Any suggestions for papers on low noise signal combining?
> Crazy idea, Is there anyway to add them in series?
Sure, if you transformer-couple the outputs; use 9 transformers
to put ten in series. The individual units need filtered DC on the
power terminal to bias 'em, though, and there's DC on the outputs.
I'm not happy paralleling the (FET) outputs; seems like the DC
levels would interact poorly. One could, however, use a low-Z
input, high-Z output amplifier on each unit (grounded base transistor)
and just tie all the outputs (collectors) together, with suitable
bias.
More troubling, the multiple small microphones are a phased array
of elements; you'd want to form the array with respect to the desired
microphone directional gain (which will be frequency-dependent).
That's it. Change the subject when tou can't do anything else.
Typical loser attack.
Hmm, can't you just AC couple them into the opamp?
George H.
Eeyore was essentially history, then along comes the dorko twins to
feed the troll.
That's OK. Since my filter scheme relies on a "key" file rather than
having to change the program, I see all kinds of possibilities for
even better troll/troll-feeder filtering ;-)
...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
Help save the environment!
Please dispose of socialism properly!
"> I'm not happy paralleling the (FET) outputs; seems like the DC
> levels would interact poorly."
Hmm, can't you just AC couple them into the opamp?
** Yep - as I have already said, there will be an improved s/n ratio as a
result.
When several similar mic capsules are paralled - the random noise voltage
drops, compared to one. However, the signal output will remain essentially
the same (as from one) long as the capsules are sited together in a
directional array.
This simple fact has got some here bamboozled.
..... Phil
>krw wrote:
>> On Thu, 17 Dec 2009 02:25:00 +0000, Eeyore
>> <rabbitsfriend...@removethishotmail.com> wrote:
>>
>>> krw wrote:
>>>> "Phil Allison" <phi...@tpg.com.au> wrote:
>>>>> "krw"
>>>>>> Some "dynamics" are really electrets in drag.
>>>>> ** Totally stupid bullshit.
>>>>>
>>>> In fact, you are. You and DimBulb; a perfect pair.
>>> Provide a link to one of these 'electret dynamics' or SHUT UP !
>>
>> Wow, a trio of DimBulbs. We've hit the mother load!
>
>That would be "mother lode" actually. Can't you even spell ?
>http://www.askoxford.com/concise_oed/motherlode?view=uk
>
>You're the DimBulb here.
That's exactly what DimBulb says, dumb donkey.
>Dynamic and Electret have specific and very different scientific
>meanings and are constructed totally differently.
No shit? What a dumb donkey, DumbDonkey.
>You may have seen an electret mic *marketed* as dynamic by some bunch of
>half-wits but that doesn't make it valid.
You read just as well as DimBulb, too. Yep, DumbDonkey an
AssholeAllison really are DimBulb's DimBulb.
>
>krw wrote:
>>
>> On Wed, 16 Dec 2009 22:37:10 -0500, "Michael A. Terrell"
>> <mike.t...@earthlink.net> wrote:
>>
>> >
>> >krw wrote:
>> >>
>> >> On Thu, 17 Dec 2009 11:29:21 +1100, "Phil Allison" <phi...@tpg.com.au>
>> >> wrote:
>> >>
>> >> >
>> >> >"krw"
>> >> >>
>> >> >> Some "dynamics" are really electrets in drag.
>> >> >
>> >> >
>> >> >** Totally stupid bullshit.
>> >> >
>> >> In fact, you are. You and DimBulb; a perfect pair.
>> >
>> >
>> > Just don't try to seperate them. You know what happens when you try
>> >to divide zero by zero.
>>
>> You get a DimBulb Donkey? <shudder>
>
>
> Or a black hole.
Allison?
That would be a 'black ho', wouldn't it?
I saw that, sorry I didn't mention it. (Guess I was a bit cranky
cause you wouldn't give me the name of a good/cheap dynamic mic...
google found nothing for less than about $100. But as is usual I was
probably searching for the wrong thing... half the battle is knowing
what something is called.)
"> This simple fact has got some here bamboozled."
Correlation effects in noise are not always obvious. (At least it
took me a while to figure it out when I first encountered it.)
George H.
For your failure to explain or show a picture or diagram of how
the mics are placed, I can not determine what the purpose was.
It would be a big problem to do this just to reduce noise.
There has to be some kind of directionality involved.
greg
I saw that, sorry I didn't mention it. (Guess I was a bit cranky
cause you wouldn't give me the name of a good/cheap dynamic mic...
google found nothing for less than about $100. But as is usual I was
probably searching for the wrong thing... half the battle is knowing
what something is called.)
** See:
http://www.askshop.co.uk/shopping/microphone-capsules-mic-insert-dm2.html
YOU FUCKING FIFTIETH WIT !!!!!
... Phil
> My boss laid a unit on my desk today. Its got ten electret mics fed
> into one low noise opamp. Any ideas how they did this without a ton
> of noise. All I can see from the board layout is it is NOT the classic
> op amp summer.
>
> Any suggestions for papers on low noise signal combining?
Parallel arrays of identical small capsules can be used as a substitute
for a large-diaphragm capsule to reduce the electrical and Brownian
movement noise.
If identical in-phase signals from two capsules are combined, the
resulting signal increase will be 6dB. If the random noise from two
identical capsules is combined, the noise increase will only be 3dB
because the two noise sources, being random, will partially cancel each
other. The result is an improvement of 3dB in the S/N ratio.
This can be done with many more capsules, to give a significant
improvement in S/N ratio (I have used a stack of 32 capsules for one
specialised low-noise array). The result will not be quite as good as a
large diaphragm, because the capsules (being round in shape) will have
to be spread out over a larger area for a given size of diaphragm
surface, so the phase differences between capsules for off-axis signals
will be greater.
However, the capsules can be arranged in a vertical stack, which reduces
phase problems in the horizontal plane at the expense of worse phase
effects in the vertical plane. That may not matter as long as the
performers are all in the same horizontal plane (which they often are).
--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk
They are in a arc, 3 each side with uneven spacing and 4 touching
each other in the middle about the vertex.
Steve
When you talk uneven spacing it sounds like your not trying to get
high phase addition or cancellation. Its still going to be
frequency dependant.
greg
What is claimed? Phased array? Directional? Selective with
frequency?
Outer mics have low pass characteristics, , middle ones have peaks at
high and low but are down in the mid range and inner 4 have high pass.
Steve
o...@uakron.edu wrote:
> On Dec 16, 9:23 pm, Eeyore wrote:
> > o...@uakron.edu wrote:
> > > AMBIENT ** noise pickup, not self noise !!
> > >> In any case, YOU are 100% WRONG in your assumption that there ought to
> > >> be a noise problem when combining the mics signals. Wired either in series
> > >> or parallel, correlation still works to improve the s/n ratio.
> >
> > > Mixer front ends still add thermal noise
> >
> > Uh ? What's a 'mixer front end' ? Obviously you're not an audio
> > professional and haven't a clue about current practice. You can make a
> > mix stage today using inexpensive off the shelf parts that has 100dB or
> > more dynamic range quite easily.
> >
> > So what's the mix stage chip ?
>
> Nope, My experience is with 1980s audio when every open channel was
> a problem.
It asn't a problem even then with well-designed gear.
> I'm glad to admit that. I'm a laser/electro-optics guy.
>
> Any specfic AES paper?
No need. It's the semi manufacturers that have produced the stunning parts
available today, not the AES. 100dB dynamic range is quite routine. 120dB is not
unusual on a single signal path. Try looking at app notes.
Graham
--
due to the hugely increased level of spam please make the obvious adjustment to my
email address
krw wrote:
> You read just as well as DimBulb, too. Yep, DumbDonkey an
> AssholeAllison really are DimBulb's DimBulb.
You're a DAMN MORON !
"Michael A. Terrell" wrote:
krw and you started the ad-hominem attacks because you have no relevant
knowledge to be able to do anything else. Just like the AGW crowd in fact !
Graham
Jim Thompson wrote:
> "Michael A. Terrell" <mike.t...@earthlink.net> wrote:
> >Eeyore wrote:
> >> Michael A. Terrell wrote:
> >> > krw wrote:
> >> >> "Phil Allison" <phi...@tpg.com.au> wrote:
> >> >>> "krw"
> >> >>
> >> >>>> Some "dynamics" are really electrets in drag.
> >> >>>
> >> >>> ** Totally stupid bullshit.
> >> >>>
> >> >> In fact, you are. You and DimBulb; a perfect pair.
> >> >
> >> >
> >> > Just don't try to seperate them. You know what happens when you try
> >> > to divide zero by zero.
> >>
> >> So, you're as thick as 'krw' are you ?
> >>
> >> Graham
> >
> >
> > That's it. Change the subject when tou can't do anything else.
> >Typical loser attack.
>
> Eeyore was essentially history, then along comes the dorko twins to
> feed the troll.
>
> That's OK. Since my filter scheme relies on a "key" file rather than
> having to change the program, I see all kinds of possibilities for
> even better troll/troll-feeder filtering ;-)
>
> ...Jim Thompson
So you support those who talk crap to make a non-existent point like the AGW
crowd do you ? Make your damn mind up will you ?
Now reconsider krw's insane coment " Some "dynamics" are really electrets in
drag ". Note that he is unable to post a single link to an example, just as the
warmingists can't 'prove' AGW.
>
>
>krw wrote:
>
>> You read just as well as DimBulb, too. Yep, DumbDonkey an
>> AssholeAllison really are DimBulb's DimBulb.
>
>You're a DAMN MORON !
...and you, DumbDonkey, really are DimBulb's DimBulb.
You really are that stupid. I thought you just couldn't read. Oh,
well, DumbDonkey, have fun with your new lovers, DumBulb and Phyllis.
>
>
>"Michael A. Terrell" wrote:
>
>> Eeyore wrote:
>> > Michael A. Terrell wrote:
>> > > krw wrote:
>> > >> "Phil Allison" <phi...@tpg.com.au> wrote:
>> > >>> "krw"
>> > >>>
>> > >>>> Some "dynamics" are really electrets in drag.
>> > >>>
>> > >>> ** Totally stupid bullshit.
>> > >>>
>> > >> In fact, you are. You and DimBulb; a perfect pair.
>> > >
>> > > Just don't try to seperate them. You know what happens when you try
>> > > to divide zero by zero.
>> >
>> > So, you're as thick as 'krw' are you ?
>> >
>> > Graham
>>
>> That's it. Change the subject when tou can't do anything else.
>> Typical loser attack.
>
>krw and you started the ad-hominem attacks because you have no relevant
>knowledge to be able to do anything else. Just like the AGW crowd in fact !
>
You really are trying to one-up DimBulb. You're a close second, I
gotta give you that.
Yawn... Same old crap, over and over.
> krw wrote:
>> On Thu, 17 Dec 2009 02:25:00 +0000, Eeyore
>> <rabbitsfriend...@removethishotmail.com> wrote:
>>
>>> krw wrote:
>>>> "Phil Allison" <phi...@tpg.com.au> wrote:
>>>>> "krw"
>>>>>> Some "dynamics" are really electrets in drag.
>>>>> ** Totally stupid bullshit.
>>>>>
>>>> In fact, you are. You and DimBulb; a perfect pair.
>>> Provide a link to one of these 'electret dynamics' or SHUT UP !
>>
>> Wow, a trio of DimBulbs. We've hit the mother load!
>
> That would be "mother lode" actually. Can't you even spell ?
That is funny... He calls everyone he is too lazy to argue with
"stupid". But it fits since "stupid" is the oldest/lamest insult
known to man.
And you are an obsessive lamer...
--
> Path: news.astraweb.com!border5.newsrouter.astraweb.com!
nntp.club.cc.cmu.edu!fu-berlin.de!uni-berlin.de!individual.net!not-
for-mail
> From: krw <krw att.bizzzzzzzzzzz>
> Newsgroups: sci.electronics.design
> Subject: Re: Microphones in parallel to reduce noise
> Date: Sun, 20 Dec 2009 23:01:29 -0600
> Lines: 13
> Message-ID: <v70ui5dab3qc3lnk03dprg2q1c21tkbr4a 4ax.com>
> References: <_tKdnRiTv5dG7rTWnZ2dnUVZ_sydnZ2d giganews.com>
<g4tii5ture8r5vcnjkoop7pdknrevs2l32 4ax.com>
<batii5popo65j738uqu8qgsms49ilpumng 4ax.com> <7otcb4F3qom5aU1
mid.individual.net> <lp0ji5hp620vbvo3c74opg2tdaje34pls8 4ax.com>
<016dcddb$0$14148$c3e8da3 news.astraweb.com>
<smbji55i4bgpuf6qkeb1rr2otaj16sfuph 4ax.com> <00600487$0$9595$c3e8da3
news.astraweb.com> <6fhli5d1d5to3kaql14pck2ne62d7lqn5g 4ax.com>
<4B2EE6DA.10F3B577 notcoldmail.com>
> Mime-Version: 1.0
> Content-Type: text/plain; charset=us-ascii
> Content-Transfer-Encoding: 7bit
> X-Trace: individual.net
c9l4h2pj2hpzIqRLDnb+ogu709XzMudvtONM2eB86D0XuTN3+n
> Cancel-Lock: sha1:ZXBNCxiU6InGa/Nxx9E+6C5ChxM=
> X-Newsreader: Forte Agent 5.00/32.1171
>
>krw <krw att.bizzzzzzzzzzz> wrote:
>
>> On Mon, 21 Dec 2009 03:09:14 +0000, Eeyore
>> <rabbitsfriendsandrelations notcoldmail.com> wrote:
>>
>>>
>>>
>>>krw wrote:
>>>
>>>> You read just as well as DimBulb, too. Yep, DumbDonkey an
>>>> AssholeAllison really are DimBulb's DimBulb.
>>>
>>>You're a DAMN MORON !
>>
>> ...and you, DumbDonkey, really are DimBulb's DimBulb.
>
>And you are an obsessive lamer...
Oh, lookie. A puppy is following me around. BAD BOY! Don't shit in
your food, dumb dog!
No, stupid, I call people stupid who are demonstrably stupid. It
really isn't hard to figure out, for most.