I've been promising myself for a long time that I would do some bench testing of the quantitative characteristics of the GL2's mic jack. Heaven knows that Canon isn't going to tell us any more than is in the user manual.

Well I finally got to it. I used some excellent audio analysis software that I recently discovered, DSSF3 (www.ymec.com), along with a good digital multimeter, a 5K potentiometer, various adaptor cables and a bunch of hookup wire. DSSF3 includes tone generator, oscilloscope and THD (Total Harmonic Distortion) applets. [Edit: I barely scratched the surface of the software. For you real pros, the DSSF3 suite has truly impressive (i.e., over my head) capablities for analyzing room accoustics and more. The site offers a full 30 day trial, as well as extensive screen shots, help and information. I'm plugging the site and the software because its existence encouraged the work below]

For all my measurements I played a 1000 Hz sine wave tone through my computer's D/A ouput (line out), then through the potentiometer into the GL2's mic input jack, then out the GL2's headphone jack and into the computer's D/A input (line in) for processing by the oscilloscope and THD applets. For simplicity I worked with just the left channel.

Here are the highlights:

MAXIMUM TOLERABLE INPUT VOLTAGE WITH MIC ATT OFF: 14 mV (-37 dBV).
Above 14 mV the sinewave immediately started to distort and the THD started to rise rapidly from its nominal 0.2% up into the tens of percentage points. Could the distortion be heard? Yes. When headphones were substituted in, even these old ears could hear the changes starting a few mV above 14 mV. IMPORTANT: The distortion occurred regardless of the setting of the GL2's level controls, which could easily be adjusted to show a nice -12 dB level indication on the meters even when the input was cranked so high that the sine wave was squashed into a nearly square wave. This is why you have to be careful using hot mics (e.g., ME66, AT4073) with the GL2.

MAXIMUM TOLERABLE INPUT VOLTAGE WITH MIC ATT ON: 120 mV (-18.4 dBV).
Same comments as above, except the nominal THD below 120 mV was 0.6%

USING THE GL2 LEVEL METERS TO MEASURE INPUT VOLTAGE
If you're unsure as to what you're feeding the input, here's how to find out using the GL2 itself. The GL2 meters don't directly measure the input levels, they measure the level at which the amplified analog audio is being applied to the digital converter--an important thing to monitor. However, at a given setting of the contol knob the displayed levels do correlate to the input levels. Therefore, we can simply set the GL2's level control knobs at the indicated percentages of full scale (a visual approximation), note the level meter indication in dB and obtain the input voltage level from the table below. The GL2's external level meter is labeled numerically in dB, while the LCD meters have corresponding markers without numerical labels. Actually, only three of the columns below are really needed to estimate most input levels, the rest are kinda superfluous. I would use the MIC ATT ON, 25% and 50% columns, and the MIC ATT OFF, 50% column.

Input Voltage (mV) vs. GL2's Level Meter (dB) at Several Audio Gain Control Knob Settings

          Knob Settings, ATT ON        Knob Settings, ATT OFF
  dB     25%       50%       75%       25%       50%       75%
   0      245        40         13          27          5          1.1
 -2      160         33         10          18        4           0.9
 -6      100         20        6.5          11         2.4         0.5
-12      50          10          3            5         1.1
-20      20         3.7         1.1          2          0.3

Caveats:
1. I assume that my GL2 is typical
2. A 1000 Hz sine wave is a typical standard for measurements, but the effective voltage of real world audio signals behave a bit differently. Nevertheless, the above should be good approximations; certainly better than not knowing where you are.

I hope someone finds this info useful. I also made some obserations about using the DXA-8 but I will post those separately.

Edited 2/23 for general embellishment and to clarify use of the table.

Good information, Fred. I honestly would have expected lesser results, not better. Kinda surprised at how reasonably close to standard the inputs are overall. Not quite as sweet as you'd hope, but then again, consider what it is you're testing. A 10K camera it ain't.

Thanks a lot for doing this. I have a Gl-2 and a ME-66 and have been struggling a bit with distorted audio.

My knowledge of what I am doing is pretty limited here, and I could use some advice. I have the option of using the mic att switch or inserting a attenuator inline prior to the mic input. Should I use the pad set at -10 and leave the mic att switch off or use both? What would you suggest for the levels (I usually put one high and one low).

I will be using the mic tonight on a boom to collect conversational discussions in a noisy room. It does a great job isolating voices but overloads when the occasional loud laugh occurs.

Adjusting the GL2's levels won't help you unless the problem is that you have been recording at too high a recording level all along and have been hitting the 0 dB ceiling. But I'm guessing you already know enough to be avoiding that, and that your problem is distortion due to overloading the input. The GL2 level controls have no effect on that. Setting the channels for different levels would only mean recording the distorted sound at different levels.

The ME66 puts out 50 mV at 94 dB SPL (sound pressure level). Most of that will appear at the GL2 input because of the much greater impedance of the mic jack (5900 ohms) compared to that of the microphone (200 ohms). I've seen 90 dB SPL compared to a lawnmower a few meters away, and to a rock band at low to average loudness. My sense is that a loud laugh a meter away peaks at well over 100 dB.

At 105 dB SPL the ME66 would put out 150 mV, which would overload the GL2's input even with MIC ATT ON (it can only take 120 mV). Adding the -10 dB pad in line would bring it back down to about 50 mV, which would be okay. You should be good to about 110 dB with MIC ATT ON and a -10 dB pad. Unfortunately, if you set your GL2 so that loud laughs peak at, say, -3 dB, then normal conversation will be at about -45 dB. Maybe a bit better than that with boom pole operation, or if I've overestimated the loudness of the laugh, but you get the idea.

The only really good way out of this is to use limiting such as provided by the BeachTek DXA-8.

--
Fred

Hey Fred, nice work!

Did you get a chance to compare EE to what actually gets recorded on the tape? Were you assuming them to be identical or would it have taken too much additional time to do that test also?

If you have two cameras you could set up a second camera that videotapes your testing, perhaps have it videotape the viewfinder of the camera you are doing the tests on and verbalize your testing as you do the tests. The camera that is actually being tested could be pointed at your various testing devices.

Then you might be able to compare your ee results with what actually gets recorded on tape and then compare the two by reviewing the two tapes that you created.

Because you videotaped the viewfinder you will have instant time-code offsets between the two cameras.

<<<-- Originally posted by Bill Ball : Thanks a lot for doing this. I have a Gl-2 and a ME-66 and have been struggling a bit with distorted audio.

My knowledge of what I am doing is pretty limited here, and I could use some advice. I have the option of using the mic att switch or inserting a attenuator inline prior to the mic input. Should I use the pad set at -10 and leave the mic att switch off or use both? What would you suggest for the levels (I usually put one high and one low).

I will be using the mic tonight on a boom to collect conversational discussions in a noisy room. It does a great job isolating voices but overloads when the occasional loud laugh occurs. -->>>

Before you convert the XLR signal to a mini-input you might want to try a -15DB pad and then turn off any auto gain functions that can be turned off via the menu.

Thanks, guys.

We used the ME66/GL-2 last night with a -10db pad inline before the mic input, the camera mic att switch turned off and with two levels of settings on the mic channels (about 50% and about 25%). This worked really well. Still had the on-axis sensitivity of the me66 without overloading the camera.

Bill

My experience over the years has been if you have to set your camera microphone input level lower than 9:30 or higher than 2:30-3:00 (12:00 being the middle point) you probably have an audio level mismatch somewhere in the system.

Did the 25% setting fall within the 9:30-3:00 guideline?

<<<-- Did the 25% setting fall within the 9:30-3:00 guideline? -->>>

It would have been at 9:30. That was only to save us from the unexpected loud outburst. The channel set at 50% (or 12:00 to use your labeling) will be the one we will use.

Glad it worked out, Bill. I was forgetting that your application was to record conversation, which covers a relatively narrow dynamic range (compared to, say a stage musical), so my conclusion that you were going to need signal limiting was overkill.

The loudest normal conversation (excluding the laughter outbursts) from a meter or two away should weigh in at about 70 dB SPL, which would excite the ME66 to the level of about 10 mV. Inserting a -10 dB pad would cut that to about 3 mV at the GL2's input. According to the table in the original post above, that would result in recording level peaks of about -6 dB with the GL2's gain set at 50% with MIC ATT off. Perfect.

Just remember that your protection channel at 25% will only protect you from hitting the 0dB ceiling on recording level, it won't protect you from overloading the GL2 input. The -10 pad will protect the input to about 94 dB SPL with MIC ATT off. If your loud laughs go much above that (I guessed it might hit 105 dB SPL in an earlier post, but I'm not at all sure), you'll probably hear some distortion even in the 25% channel.

I'm way too analytical. If you had posted your problem as a new post, the experienced hands probably would have responded immediately and simply, "use a -10 pad with the ME66." But I just love it when everything make sense.

Fred, that's some great information to have when looking for new video mics. I'm curious, though, what the db reduction rating is for the internal MIC ATT. Somewhere I thought I read that it was 10db. Is an external one still better?

I often record from very loud sources (aka mixer), so I've got two 20db attenuators (one for each channel) which do fairly well when combined with the on-camera attenuator. I may still have to lower the send level on the mixer somewhat, but this gets it close (without distortion -- been there, done that!).

Thanks, Jeremy.

The MIC ATT ON setting gives -20 dBV (dB based on voltage) attenuation to the input. That's the GL2 spec on page 156 of the manual, and that's approximately what you see in my tablulated results (-20 dB is 1/10). That equates to -10 dBW ( dB based on watts), which is why you may have heard it that way.

Based on my work, it behaves no differently from an external pad.

Thinking about a way to properly set a mixer send level is one of the things that prompted my little exercise above.

Ah, that would explain my confusion. I didn't realize you could measure it in different ways.

So my 20dB attenuators would be -20dBV? I got them from B&H:
http://www.bhphotovideo.com/bnh/controller/home?O=productlist&A=details&Q=&sku=135093&is=REG

Thanks for the clarification!

Jeremy

Yes, dBV is the convention when decbels are applied to electronic signals, so unless otherwise specified, dB in that context is dBV.

for &nbsp;-6 dB the output voltage is 1/2 the input voltage
for -10 dB about 1/3
for -20 dB 1/10
for -40 dB, 1/100
etc.
+dB inverts the above fractions.

For historical reasons, dBu = dBV + 2.2

To be a bit more specific, dBV means the level is measure relative to 1 volt. Thus -6 dB means it is 0.5 volts, and -60 dBV is 0.001 volts

dBm is ir referenced to a milliwatt (which translates to 0.775 volts across 600 ohms).

dBu means it is measured relative to 0.775 volts (without regard to impedance/resistance)

dB is just a ratio (no specific units) although it maybe implied by the device in question.

At one time this there was a dBv (small v) whihc was equal to dBu, but that was confused with dBV, so they probably dropped the designation staying with dBu)

Could anyone tell me how to display the table in Fred's post correctly,please?
I get strings of "&nbsp;" rather than a table.

Could anyone tell me how to display the table in Fred's post correctly,please?
I get strings of "&nbsp;" rather than a table.

He attempted to insert HTML code which won't work. See the table at bottom left of this page.

-gb-

It worked at the time, but has gone awry since.

Alan, I can email you a readable version if you want.

Thanks Fred - I have emailed you!

Well, after studying Fred's findings I have taken the plunge. My DXA-8 arrived from LTF this morning (for the same price in £ as B&H quote in $), and I have been playing with it under controlled conditions with various of my recorded sound tracks. The camera/Microphone/Century teleconvertor/Beachbox looks most impressive! Fortunately it sounds just as impressive.

My problem is that I regularly record waterfowl where there is a general background chatter, but every now and then a Mallard or a Coot lets rip close to the camera. If I attenuate the levels to cope with this without overloading my XM2 input, the background chatter is often at -30dB or more, and is lost in the noise and hiss.

With the DXA-8 in place, edited tracks with Mallard calls normalised at -3dB can be recorded on the XM2 via the ME67 with the limiter on so that they are still limited to -3dB, but the background chatter level in the XM2 can now be as much as 12dB louder than on the recording, without overloading the XM2 inputs on the peaks. With the limiters off, the Beachbox gain has to be backed off or the camera audio is clipped with clearly audible distortion.

Another big plus is that I can now use the K6 with the XM2 MIC Attenuator off.

Now to try it in the field!

Alan, you're just the type of chap I was thinking of when I shared my data. I'm glad you found it useful. Good luck!