Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Jim Williams]

If those natural sounds are so offensive, no one would have bothered singing until modern processing electronics were available.

[klaus]

(...) As I understand it, the KM184's response would be similar to that of a 1/2" omni  B & K measurement mic. What would be the difference in your opinion?

Can you cite a source for your 'understanding'? MY understanding is quite different. A quick look at the frequency response of the KM184 demonstrates its (deliberate?) high frequency boost.

[David Satz]

In addition, the B & K measurement microphone is a pressure transducer with an omnidirectional pickup pattern, while the KM 184 is about 50% a pressure-gradient transducer and only about 50% a pressure transducer, resuting in a cardioid pickup pattern. No one would ever mistake the one microphone for the other, I think, even if the free-field, 0-degree frequency response of both microphones was identical (which I'm rather certain it is not).

--best regards

[Timtape]

In addition, the B & K measurement microphone is a pressure transducer with an omnidirectional pickup pattern, while the KM 184 is about 50% a pressure-gradient transducer and only about 50% a pressure transducer, resuting in a cardioid pickup pattern...

 My apologies, I meant the omni capsule of course, the KM183. 

[Timtape]

If what you say were correct, why don't more people use B&K/DPA mics? Their static specs are certainly hard to beat: s/n and frequency linearity are exemplary.
Yet, they are not considered "musical" by most artists, music engineers and producers, and rarely find usage in recording studios.

What may be missing here?

Perhaps that measurement mics are calibrated (and that may mean hand selected samples) to a very strict standard, with something like a frequency response accurate within 0.3db. That will cost the company and therefore the customer more money. For audio work that sort of accuracy isnt normally required. Of course the same company could also market essentially the same mics but without such stringent specs. I cant be sure. Hard to find such information on the net.

Certainly cardioid mics are often  more desirable in an audio recording vs omnis and especially in PA situations. But as we know in this case, some  accuracy is traded off for the more useful cardioid type pattern with its selected rejection of unwanted sounds and less feedback susceptibility. This seems to reinforce the first point that in audio recording situations, some accuracy trade off is acceptable where it wouldnt be for strict measurement requirements.

[soapfoot]

Perhaps that measurement mics are calibrated (and that may mean hand selected samples) to a very strict standard, with something like a frequency response accurate within 0.3db.

The thing is, when discussing such a calibration against a reference, we must (at a minimum) answer the following questions:

"at what distance?"
"at what angle of incidence?

And when we do answer these questions, we must also understand the limited scope and relevance of the information presented.

If a microphone is calibrated to be "accurate within 0.3 dB" at, say, 1m directly on-axis, that's not going to be perfectly correlated with its frequency response at 1cm, or 10m, or at 90 degrees off-axis, or even at ten degrees off-axis.

And outside an anechoic chamber, off-axis coloration will always contribute to the sound of a microphone. Even if a microphone measures completely "flat" in all typical tests, it's still possible that it may perform poorly for music recording.

Subjectively, I've found off-axis characteristics to be one place where some of the better microphones really "earn their stripes," for me. Bleed/spill/leakage from off-axis sources into something like a U87 or U67 is less-often objectionable to me than spill into a lesser microphone, even if the on-axis characteristic of the lesser microphone is acceptable.

[Timtape]

 We were discussing omni mics.  An omni DPA 1/2" SDC and the omni 1/2" SDC measurement mic equivalent will share an essentially identical polar response. The only difference between the two mics will be that the measurement mic is calibrated to a  higher degree of  accuracy. Again, will that higher degree of accuracy make it less "musical"? If so, how? What causes the alleged qualitative difference to appear?

On your last point about the desirability of well behaved off axis response in say a cardioid mic I totally agree, but how does that relate  to the claim that a measurement mic is by definition  less "musical" - whatever that means? Omni mics of course have a better behaved polar response than cardioid types, but regardless, when comparing otherwise identical omnis with omnis  it's of course irrelevent.

I suggested two reasons why measurement mics are less common in non measurement applications. You didnt acknowledge or address either point but made a point about polar patterns which is true but seems irrelevent to the point being discussed.

Tim

[klaus]

I can think of an issue that could make measurement mics inherently unsuitable for music recording: transducer and electronics that are customized for an ultra-linear rendering of single, sequentially fed sine waves.

You get that kind of response by using extensive corrective electronic and acoustic measures, which typically move the overall response of such a mic away from simple but musical (and decidedly not linear!) circuitry and capsule geometry.

The all-time greats still widely used today to record music do not measure well at all; look at a graph of an original ELA M251, for example: nothing but valleys and peaks, and audible differences from one mic to the next. But what seems to slip through quite nicely despite such gross non-linearity (or because of it) is the music.

Or to put it simply: it seems that the more complicated a microphone's circuitry, the less musical. Somehow non-linear musicality trumps the ultra-linear response of measurement mics.

[Timtape]

I can think of an issue that could make measurement mics inherently unsuitable for music recording: transducer and electronics that are customized for an ultra-linear rendering of single, sequentially fed sine waves.

Again I refer you to the earlier point.  Frequency response  over the audible range within a fraction of a db tolerance has been achieveable in  measurement mics for decades - and without external corrective EQ.  That sort of linearity is not normally needed for music and voice recording - but will hardly make recordings made with such a mic "less musical". Some people do indeed use measurement mics for recording. I personally think they are wasting money on a level of technical accuracy that nobody will appreciate just listening to music.

So  using a measurement mic and amp combination that has even more accuracy (ultra linear) at certain spot frequencies for a music recording would be even more pointless, and even more  needless expense.

Some mics are more suited to different tasks. The ELA M251 might have some roughness in response (I havent seen a graph or used one) but I suspect it's a very useable mic not because of some  roughness in response but in spite of it. And the response wont be very rough or nobody would use it for high quality work.

Then there's the issue of response graphs that have been "corrected" or "smoothed" We could discuss that separately...

And why  only talk about condenser mics? If roughness of response made a mic more "musical" then high quality condenser mics would not have needed inventing. Cheap, garden variety dynamics have bags of non linearity. But we pay more money for an MD441 dynamic which was carefully designed to rival the linearity of a condenser - but doesnt quite get there.

I dont agree that the more complicated the mic's design the less musical. The old AKG D202 had the unusual complication of two transducers. The ELA M250 had two backplates I believe. The U87 has two diaphragms. The new Shure KSM 8 stage dynamic mic is a complicated and more expensive design which achieves pretty flat response, the best absence of proximity effect and smoothest polar pattern in such a mic ever. Before it the whole range of EV "D" mics which aimed for similar performance in a rugged package but which also entailed a more complex design.

The wise old design rule is "as simple as possible - but no simpler"...





   

[Michael O.]

I have to agree with KH here: very generally speaking simple proves best. Klaus is talking about simplicity in circuitry, and the EV variable-d design that was mentioned could hardly be more elegant in that regard (no circuitry). They're simply a transducer, a connector, and sometimes a transformer. The vents that create the pattern and avoid the typical proximity effect are innovative and clever, but not complex in their implementation. The corrective electronics he refers to is the sort within a (condenser) mic's internal circuitry (e.g. the hi frequency de-emphasis network in a U87).

If you follow this whole thought experiment to its logical conclusions it all becomes moot. Regardless of the accuracy of any given microphone we would have no perfectly accurate way to then output the recorded sound. And if such an accurate playback medium existed we would need an equally well-suited listening environment along the lines of an anechoic chamber to fully appreciate it (with our imperfect and indivually-tuned ears, at that). So, even if our music went from air, to perfect mic, to perfect circuitry and recording medium, and out of a perfect playback system, we would need particular listening conditions to even begin to accurately discern such a signal path.

But ultimately this all comes down to something completely subjective/aesthetics/personal preference. And to generalize, people certainly seem to gravitate toward non-linearity/coloration/roughness in sound.

One final (or maybe initial) consideration is that in this context "accurate" has no determinate definition.

[Timtape]

I have to agree with KH here: very generally speaking simple proves best. Klaus is talking about simplicity in circuitry, and the EV variable-d design that was mentioned could hardly be more elegant in that regard (no circuitry). They're simply a transducer, a connector, and sometimes a transformer. The vents that create the pattern and avoid the typical proximity effect are innovative and clever, but not complex in their implementation. The corrective electronics he refers to is the sort within a (condenser) mic's internal circuitry (e.g. the hi frequency de-emphasis network in a U87).

Hi Michael, I'm  not sure what you are saying here. Are you saying that Neumann made a mistake in using the internal de emphasis in the almost legendary U87 mic? How could it ever have come to be such a highly regarded mic? BTW I didnt know the U87 had internal corrective EQ. Interesting. Which other mics have similar circuitry?

If you follow this whole thought experiment to its logical conclusions it all becomes moot. Regardless of the accuracy of any given microphone...

 but "the myth of the accurate microphone " is the topic of the discussion...

we would have no perfectly accurate way to then output the recorded sound. And if such an accurate playback medium existed we would need an equally well-suited listening environment along the lines of an anechoic chamber to fully appreciate it (with our imperfect and indivually-tuned ears, at that). So, even if our music went from air, to perfect mic, to perfect circuitry and recording medium, and out of a perfect playback system, we would need particular listening conditions to even begin to accurately discern such a signal path.

Quite and it puts into perspective the obsession some people have with getting just the right microphone, as if microphone accuracy is a fundamental problem, when in practice, the playback and potential for hugely variable listening conditions is far more problematic if fidelity to the source is our standard.

But ultimately this all comes down to something completely subjective/aesthetics/personal preference.

That's certainly a factor but it depends on the weighting we give it. If people's preference was so widely divergent, group listening to live symphony orchestras would be impossible to please most attendees, as it would be for any group listening activity. Unless a person has a moderate to serious hearing impairment there are good general guidelines as to what most people regard as good, listenable sound. Professional cinemas for example are only able to deliver good acceptable sound to most film goers because they are guided by such parameters and indeed there is a long history of this, with ongoing progress and development.

 

And to generalize, people certainly seem to gravitate toward non-linearity/coloration/roughness in sound.

With certain sounds, and in certain musical genres yes but not so much in others. Classical music recording has a long tradition of fidelity to the sound of the performed music although there will always be some disagreement as to what constitutes the ideal listening position, balance of instruments, amount of venue reverberation etc, including listening level, which is also related to Fletcher Munson.

One final (or maybe initial) consideration is that in this context "accurate" has no determinate definition.

Well if we limit ourselves to the initial topic of the thread: the accuracy of microphones, we might find there are  indeed objective standards and tests that are empirical and repeatable. Once we confuse this aspect with personal subjective preference - which is to some degree true - and the potentially huge variability of real world listening conditions, both of which you mentioned and are certainly true, we have changed topics. Again I thought the question was the accuracy of microphones.

Tim

[panman]

Well if we limit ourselves to the initial topic of the thread: the accuracy of microphones, we might find there are  indeed objective standards and tests that are empirical and repeatable. Once we confuse this aspect with personal subjective reference - which is to some degree true - and the potentially huge variability of real world listening conditions, both of which you mentioned and are certainly true, we have changed topics. Again I thought the question was the accuracy of microphones.
Tim

I find this thread has been good so far and good points have been discussed. To me this topic requires a lot of deviations to make any sense at all or how would you make limits to this discussion? Too tight borderlines and what is there to discuss?

[Jim Williams]

One should come to the logical conclusion that accurate microphones do not exist at this point of human development.

Measurement microphones are not in any way accurate as to encoding the reality of what a mammal can detect. Like all microphones they are an air pressure to electron converter with many factors of actual hearing left out, like location sensing and the ability to seperate multiple sounds at the same time.

Yes, measurement microphones use the same technology as recording microphones, diaphram, impedance converter and output stage. There are not any "electronics customized for an ultra linear rendering of single, sequentially fed sine waves". No such specialized amplifiers exist, either they have low THD specs or they do not. Amplifiers do not discriminate between differing wave shapes as long as the bandwidth can accomodate all of them.

After 100+ years of microphone development we are really no closer to that goal than we were 100 years ago. Maybe that's why microphone designs from the 1930's have not yet been bested.

[Timtape]

One should come to the logical conclusion that accurate microphones do not exist at this point of human development.

Hi Jim,

Of course all mics have their limitations. The question then is "accurate to what degree, and for what task?" The mic is part of a  chain. Typically is it the weak link in that  chain,  in what respect(s) and by how much? We'd need to look at specific real life cases.

Measurement microphones are not in any way accurate as to encoding the reality of what a mammal can detect. Like all microphones they are an air pressure to electron converter with many factors of actual hearing left out, like location sensing and the ability to seperate multiple sounds at the same time.

The human ear  needs a brain to interpret the raw data. When we make a music recording we leave it to the brain of the unknown listener at home to do the interpreting. To compare the microphone with a mammallian ear PLUS the mammalian brain seems unfair and invalid.The ear on its own has the same basic limitation of a  microphone. It cant "interpret" sound.  That's not its job. (although the human ear does incorporate a basic kind of "limiter" but only to protect it from damage from loud sounds but even there I'm not sure whether or not that mechanism is triggered within the ear itself or by a signal from the brain). Again, when we make a music recording we leave it to the brain of the unknown listener at home to do the interpreting.

After 100+ years of microphone development we are really no closer to that goal than we were 100 years ago. Maybe that's why microphone designs from the 1930's have not yet been bested.

Again not sure which goal you are talking about. If you mean that we still havent artificially created a mammallian ear and brain that does the same job as well I'd agree. But we were discussing just the microphone as a transducer (and I guess the associated preamp) and its accuracy.
Sure the capacitor mic of the 30's is still the basis of modern day high quality mics. That it hasnt been bested by another design is inconclusive. It doesnt prove that it will or wont be bested in future, or in what respects and by how much.  We  dont know. We will only know if or when it happens.
In the meantime  we could discuss the microphones that we do have today...

Tim

[klaus]

(...) when we make a music recording we leave it to the brain of the unknown listener at home to do the interpreting.

In the field of analyzing any of the five senses, there is no reality outside subjective interpretation. So, objectively, nothing happens acoustically beyond and outside of sound reaching us and having an effect on us.

We continue to TRY correlating sound we hear to sound waves we measure, but that is only valid to the extent that it would allow us to identify specific measurable features of sound waves (like gross frequency anomalies) that we can correlate to dissatisfying sound and try to correct and improve it: listening is the corrective, nothing else.

The idea that certain idealized properties of measurements, like a ruler-flat frequency response, would get us closer to satisfying sound remains unprovable. So far, we can only verify by listening whether manipulating parameter x or y just got us closer to an aurally satisfying experience.

(...) we were discussing just the microphone as a transducer (and I guess the associated preamp) and its accuracy.

 'Accuracy' of a microphone can only be judged by hearing the musical event live, and comparing how close (accurate) the recording of that event comes to that experience. A determination of accuracy can therefore never be derived from any other (non-sensual) experience or method, see above.

Regarding the relatively primitive nature of microphones, even the best ones, compared to our hearing apparatus, I never get tired of citing this example:

Sitting in an auditorium, and listening to a quiet passage of the orchestra, I can tune out the noise of the person sitting next to me rustling her candy wrapper. If I were listening to a recording of this event with a microphone placed where my ear is, I would be royally annoyed by the disturbance.
The brain can analyze sound waves delivered from the ear's sophisticated pathways, and filter out unwanted noise. The relatively primitive rendering of reality by a microphone does not make this possible.