Topic: Is levels peaking at 0 to high?

[Gutbucket]

D Satz, can you please help us with this?

Correct me if I'm wrong (I'm getting in above my head here), but..

The platonic ideal of electrical energy theory is to maximize power transfer from one section to the next, with as little distortion as possible.

Except this application is not in need of efficient power transfer, but rather voltage transfer with as little distortion as possible.  As I understand it, matching impedances achieves clean, efficient power transfer.  But with most audio equipment where voltage transfer is of interest rather than power transfer, a output impedance lower than the input impedance is desirable, no?

The theoretical ideal of any amplifier is to have infinite gain, infinite input impedance, and zero output impedance.

^ This implies the voltage transfer model of low-output-impedance>high-input-impedance rather than power transfer model of matched impedance, no?

[snip..] My mentor took the specs pf the Sony D5, focusing on its' input impedance. Then looked at the output impedance of several microphones we were choosing from. (Sennheiser 421, Sennheiser 441, Shure SM58's, Beyer M88, Beyer M201, Beyer M160). He advised us to buy the Beyer M201's because their output impedance was almost an exact match of the Sony's inputs. He told us that combo of mics/deck should almost never distort until we reached the max input SPL of the Beyers. (may have been 126dB???). I notice listening back that we RARELY had input distortion on those M201 tapes.

Just looked up the spec sheets for each of those microphones and all of them state an impedance of 200Ω. Well, except for the SM58 which Shure states as: rated at 150Ω, 300Ω actual, so effectively identical to the others, when compared to..

The TC-D5 spec sheets I found don't state the microphone-input impedance, but list line-in impedance as being: either 47kΩ or 100kΩ (two different spec sheets), elsewhere I found mic-input impedance reported as being 7.6KΩ or 4.6KΩ for the pro model with input transformers.  In any case, a significantly higher input impedance than the microphone's output impedance.

[edit] ..which seems to fit the voltage transfer model of low-output-impedance>high-input-impedance.

[DSatz]

Gutbucket, could we maybe take the discussion about impedance matching and power transfer, etc., into a different thread? It's more a historical issue (especially the huge influence of telephone technology on early professional audio in the U.S.) than a practical one nowadays, and it's almost entirely separate from level-setting in digital recording. Plus it's well settled; there isn't any real controversy or difference of opinion involved, that I'm aware of at least. Well, maybe some manufacturers of $200/foot speaker cable have their own unique concepts of physics ...

You've definitely got the basic ideas straight, though. All analog audio connection-making since some time in the 1940s or 50s has been based on voltage rather than power transfer, unless the signal is encoded as RF or the source impedance is very high (condenser microphone capsules!) or non-resistive in character (condenser microphone capsules, tape heads and phono cartridges). Except for the few remaining objectors to woofer damping, power amplifiers aren't impedance-matched with loudspeakers in domestic hi-fi systems the way they used to be in the vacuum tube era. And even balanced, analog audio connections in studios use bridging loads rather than matching ones, now that nearly 100 years have passed since Western Electric's system for electrical recording of phonograph records was introduced.

By the time of the TCD-5[M] the only real concern with those dynamic mikes would be that a few of them, including the Beyer M 201 if I remember correctly, and certainly the Sennheisers, were available in alternate models with different output impedances. German tape recorders still had DIN sockets with ca. 3.4 kOhm input impedance, and some people used "medium-impedance" microphones (ca. 600 Ohms actual output impedance) to drive those, at signal levels intermediate between what we now think of as mike level and aux or line level. But as long as you had microphones with 150-200 Ohm output impedance, IIRC the inputs of the TCD-5 were about 1.5 kOhm, so no problem there; the decision should have been based instead on the directional pattern and sonic characteristics of the microphones.

That said, the Beyer M 201 (I keep saying "Beyer" to distinguish it from an early Schoeps mike with the same model number) was a nice, non-hyped-sounding hypercardioid, sturdy and reliable. I think they still make a version of it, though they may have reduced its bass response when they introduced the "TG" series; I could never quite get straight information about that. -- The real sonic gem IMO among the microphones you listed was the M 160, though with its very low sensitivity (like only 1 mV/Pa) it wasn't always a practical choice. Given its weak signals and the unbalanced inputs of that cassette recorder, it would surely have had lots of RFI problems, and even more so in more recent times.

But if you want to carry this on any further, let's please move it to a different thread.

--best regards

[Gutbucket]

Thanks for that clarification, and no need for a more in depth discussion on my part.  Was questioning a few things the others mentioned  that didn't make sense to me, yet didn't feel I could speak authoritatively about them. Your understanding of these topics is considerably deeper than my own. 

I hope no one takes my any of my questioning posts above as confrontational, my intent is a clearer understanding for us all, including future readers of the thread. 

Since I have your ear, can you also clarify the bit below which got buried by other discussion that followed?  Big thanks for your participation here and your always excellent explanations which help make TS such a great resource.

~Lee

Following along. Mr Satz, can you clarify a couple lines in your post above for me?  You said-

"..especially if your recordings are ever played back in analog." [and later..] "All told, to me it seems very likely that there would be some s/n benefit if you raised your levels at least somewhat, especially if you consider the noise of your recorder's (or any media player's) analog output circuitry."

Its the statements after "especially" that caught my eye.  Does that assume analog transfer of the raw recording from the recorder? Or that a digitally transferred recording will not have levels adjusted in the process of producing the released recording?  If so, I see how what you describe can be the case.  However, most tapers are digitally transferring the raw recording to computer and adjusting levels in editing software as part of the process of producing the finished output they and others will listen to.  That process typically includes increasing signal level to remove excess recording headroom that no longer serves a useful purpose in the released version, and instead becomes a potential detriment for the reason you describe.  When that is the case, I fail to see how analog playback or analog output circuitry noise would effect a recording originally made with excess headroom differently than one made with optimal recording levels.  I understand how the first may potentially have a lower s/n due to a higher noise floor, but that would be baked into the recording and independent of playback.

Just making sure I'm understanding you correctly and not missing something there.

Thanks for the discussion - in particular your mention of the different noise spectra of condenser microphones in comparison to preamps and the noise characteristics of preamps at various gain settings.

[DSatz]

I said earlier that I have a hard time believing that a recording with very low peak levels can be relied upon to be as quiet as one made with more conventional levels. It's not impossible--but it depends on a whole slew of particulars, none of which I can ascertain in anyone else's recording equipment. I totally get it about setting conservative levels, but at some point there's a price to pay in signal-to-noise performance, and I want to know where that point actually is for my microphones and my preamp and/or recorder rather than guessing.

From what I've seen in numerous messages on this board, I think some people's intuitions about gain and noise aren't how they actually work. Some people seem to think that the noise added by a gain stage will be, in essence, directly proportional to the amount of gain being added, so that lower gain settings will translate more or less directly into less noise. I've seen people say, for example, that for reasons of noise they'd rather connect a (relatively high-output) microphone to a recorder's line input than its mike input even if they have to turn up the record level control all the way and their recorded peaks don't get anywhere close to 0. They count on 24-bit or 32-bit recording to make up the difference. I'm pretty sure I remember seeing this w/r/t the Sony M10 recorder, for example. Unfortunately as far as I was able to tell when I tested one (not in a very elegant way), the 24-bit setting was hardly quieter than the 16-bit setting--there was less than a 2 dB difference as I recall. Now, I wasn't completely confident of my measurement result in that case. But if it was right or nearly so, then the option that the person chose based on their concept of how it should work, was probably a bunch noisier than the alternative that they rejected.

It's hard to do valid comparison testing in live recording situations; that's more of a test-bench project. Some kind of tone generator and voltmeter, separate from your recorder, are essential. And preferably there should be one other thing on hand that's less common, because when you connect a tone generator to the input of a mike preamp, the preamp doesn't "see" the same source (driving) impedance as that of your microphones, which in turn affects the noise floor. So it really helps if you have the kind of test setup, for the particular brand and type of microphones that you use, that a service technician would use if they work seriously on that kind of microphone. For condenser microphones with removable capsules, manufacturers make shielded "test heads" (a/k/a measurement adapters) that can be attached in place of a capsule, and some manufacturers will sell you one if you ask nicely. A test head places the same capacitance across the microphone amplifier's input, but you can feed signals through it, or you can short its input to measure noise (which still leaves the capacitance in place). The microphone amplifier, meanwhile, is powered in the normal way, so everything in the circuit is just as it would be in a real recording setup.

I have a test head like that for the type of microphones that I usually use (no secret--Schoeps Colette series), and I used it for my preamp noise tests. I set the preamps to 30 - 35 dB gain and the results that I got were not IN ANY WAY predictable from the spec sheet values for those preamps, which were all compiled at higher gain settings. I know of no way to predict or infer the noise performance of a preamp at a given gain level (and microphone source impedance) from knowing its noise performance at its highest gain setting with resistors across its inputs. I've never measured preamp noise at lower gain settings, but other people's measurements seem to indicate that for the best noise performance from typical preamps, you generally want a setting more in the direction of the highest gain that you can set without overloading anything--not the lowest.

So the remaining question is how that's affecting person X's recordings with recorder Y and mike preamp Z using microphones A and B, which I can't answer from here, not having their set of gear and the test equipment in front of me. But the assumption of getting similar (let alone better) noise performance from a preamp at low gain than with the same preamp set to a higher (but still safe) setting, isn't a safe assumption. It's definitely wrong in at least some cases, perhaps in many or even most, but I'd want to see a lot more measurement results before generalizing that far.

Thus my uncertainty, but thus also my sense that somewhat higher gain earlier in the "recording chain" could bring an improvement in signal-to-noise ratio without necessarily taking away safety.

[Gutbucket]

Thanks. A couple follow up questions:

Is is possible to make any generalizations about the different noise spectra of condenser microphones vs. preamps vs. the acoustic noise floor of a quiet recording environment, that might help folks unable to make noise measurements discern which of those is the primary contributor to the noise-floor in a recording they make?


In an effort to maximize s/n in the absence of actual noise measurements, I fall back on the practice of trying to achieve the lion's share of needed gain at the forward-most end of the signal chain, followed by setting recording levels to leave sufficient yet not excessive headroom to accommodate expected and somewhat predictable SPL increases and unpredictable peaks.

The first part of that is mostly about setting things up initially. And I've assumed more sensitive microphones helpful in that the microphone itself represents the earliest gain stage in the signal chain, thus requiring less gain from the preamp stage that followings (outboard or in the recorder). All good as long as the microphone's output is not so high that it overdrives the following stage.

The second part about setting gain levels is less predictable and involves more variables, including the type of music, specific program material, distance to the source, and in the case of many TS members, additional constraints such as not being able to monitor or adjust levels during the performance.  Some situations are going to require significantly more headroom than others, and the key becomes determining the Goldilocks amount - not too much, not too little.

Do you find this to be a reasonable approach?

[DSatz]

Condenser microphones have three or four main noise sources: the input noise of the first active device in the amplifier circuit, the capacitive source impedance of the capsule, and random air motion around the diaphragm. In "traditional" condenser microphones there's also some noise from the very high-value resistor through which the capsule is charged (polarized). Mike preamps have mainly the input noise of their first active device, but also the noise from the microphone including that of its source impedance.

These noise spectra tend to increase gradually toward higher frequencies, roughly like pink noise, except for the capsule. Its source impedance is highest at lowest frequencies, basically "1/f", and its level is quite significant--billions of Ohms at the bottom of the range. But its audible effect depends on subsequent amplification, since our hearing is so insensitive at low frequencies and low SPLs.

So if you look at a swept-frequency graph of noise output of any condenser microphone in a silent room or enclosure, going from the lowest to highest frequencies you'll see first the capsule noise with its downward slope, crossing over (normally somewhere on the left side of the graph) to the rising, more-like-pink noise of the other sources. The greater the capacitance of the capsule, the lower both the frequency of this transition and the amplitude of the noise will be, "all other things being equal". IIRC the transition is generally below 1 kHz for most studio microphones, sometimes well below. Maybe not the ones with ultra-small capsules, though.

As you can well imagine, with such an uneven distribution of noise across the frequency spectrum, any frequency weighting that you use will have a huge effect on the result; likewise the averaging (integration) over time intervals, which tends to erase the impulse noise that is by far the most disturbing to listeners. This complexity tends to cloud discussion, and to produce frustration among those who want simple answers to reasonable questions such as "how much noise does this microphone produce?". Unfortunately there's no easy answer.

--In your description of how you approach "gain staging" overall, the only assumption I question is that more sensitive (higher-output) microphones -> generally more desirable. That's only a "sometimes" thing. As we've been discussing, it's not necessarily productive to try to relieve a good mike preamp of its duty, and it can be counterproductive.

--best regards

[capnhook]

This complexity tends to cloud discussion, and to produce frustration among those who want simple answers. Unfortunately, there are always people ready to offer simple answers to complex questions. I wish that people wouldn't buy those answers so readily, here or in the world outside of sound recording (which I'm told exists).

https://taperssection.com/index.php?topic=202622.msg2400252#msg2400252

[Gutbucket]

Thanks, that helps.  I certainly wish to avoid the oversimplification of any conclusions, as well as the beating of a dead horse discussion into the ground of which I may already be guilty..

In the gain staging approach I mention above I'm assuming the use of preamps built into the recorder, and/or one of the small external preamps typically used for small unbalanced microphones.  The quality of such preamps has improved over what tended to be found in "prosumer" recording gear of years past, yet is nothing world-class. 

Since the external preamps you regularly use and have tested "in-the-chain" are presumably of better quality than the ones I'm using, yet were all found to have noise characteristics that must be measured to be accurately determined, in the interest of maximizing s/n without such measurements available to me, I default to choosing the higher sensitivity option between two otherwise identical micrphones as long as the higher output isn't going to overdrive whatever stage that follows.
^
My take away from the discussion here is that this presumption is insufficient in itself, yet without figuring out some way to do the actual testing, is at least reasonable when combined with setting the gain of the adjustable preamp stage which follows somewhere around the middle of its range and not overly low or high.  This is somewhat analogous to guessing at the appropriate RPM range at which to shift a manual transmission, without knowing the actual torque curve of the engine, based on the general shape of most torque curves.


The new 32-bit world may change this if what is going on under the hood is auto switching between multiple paralleled preamp stages with  each having been tuned for maximum s/n within its level window.