Jay,
I hate to say trial and error but this is basically what happens using Alister's procedure as a starting point. Most of the inputs on Sony machines use a much higher impedance so it is not just a question of matching up the impedance number.
The idea is to see if you are getting too much or too little output from the device you are sending to the camera. Don't just rely on Tone for this as it is easy to adjust tone to some level and not hear distortion from regular audio like voice and music. Try different dynamic sounds to see if the circuitry adds pops or clips or other noise without the level overmodulating.(Ie the phrase Testing 1 2 3) If the level is within range and the sound doesn't sound clean then you might be in an instance when the trim is coming into play. Listening to the headphone outputs of the camera or playing back tests is the best way to see if the mic and camera are getting a long.

The best way to avoid having distortion from a source while using auto is to set the source up in manual before giving the auto function the job. Try and remember what trim settings work well with each piece of gear and check them before recording.

If I may I'd like to add that I really recomend updating to ver 1.11 firmware to get better audio. Before I did this I had big problems avoiding distorsion and clicks when the wind blew or other sounds peeked. It was really hard to find the right sensitivity level.
After upgrading the the camera is much more tolerant to wider dynamics.

Yes, Ola, that would help with the EX1. However, with the EX3 the firmware upgrade is not necessary.

In a nutshell, if I understand all this correctly, and there is the excellent chance that I do not, you set the camera’s Audio Level controls (which some are calling “audio gain”) for channels 1 & 2 at “5”. Then you go into the Audio Set menu, under Audio Input, and set the Trim CH-1 and Trim CH-2 (which are calling “attenuator settings”) based on your mic’s sensitivity—this is merely a starting point. These settings are then adjusted based upon what you’re recording. Isn't that what the Audio Level controls are for?

Actually, this approach seems a bit bass-ackward to me. In a studio situation, this works just fine, but in a run-and-gun documentary environment this would be close to impossible to use.

Also, it would be an ENORMOUS help if everyone used the same nomenclature, based on the piece of equipment being discussed.

Another thing, unfortunate as it is, we are not all at the same level of understanding in all things. When those who are expert in any given subject write as though their readers have the same level of understanding they do, their explanations are of little value.

Am I too far off base here?

It's awkward, and that's why I recommend a very conservative approach - setting the trimming to an aggressive -30 dB (or so), keeping the output level knobs at the neutral 5. When it's too low (watching the meters), and you're absolutely sure no loud event can happen suddenly, you _might_ rise it to some 7-8 with the knobs, but you don't have to - because, as I said before, the S/N is high enough in this camera to safely rise the sound level to your liking in post, without introducing any hiss.

This way you are protected against sudden events (strong wind guts, or somebody firing his shotgun at you :)), which - with less relaxed trimming like in the [-40, -58] dB range - might trigger the limiter, which, acting as it does, can lead to compression distortion (pumping, and clicking noise).

Piotr, your comments are very interesting.

I'm encountering a problem in weddings during the cutting of the cake - there are usually a number of poppers going off. These produce a sharp but high volume sound (like a muffled gun shot). The audio always pumps following such poppers leading to a horrendous sound.

My settings are: auto audio levels (I cannot work manual here as I have no time to control the levels) and a trim level of -41 dB (both on the internal mic and on the external mic). The external mic is a Sony short shotgun type.

So, if I understood you right, if I choose a trim level such as -58 or -74dB would I eliminate such problem or is it due to the fact that I'm keeping the audio levels on automatic?

No, No , and once more - No !

Avoid auto setting like a plague - your pumping is caused by using it.

And remember about the "reversed scale" of trimming: your suggested -58 or -74 dB is actually far more relaxed trimming than what I'd recommend (of course, exact values depend on your mic sensitivity, and the actual "sound stage"). It works the other work round: lower (absolute) number gives more aggressive trimming, so -30dB is "more trimmed" than - 58dB.

I mainly record live music events, where the only outburst in the sound level is typically the audience applause - and yet I never trim my mic above (or below, is you consider it's in the minus domain) some -38 dB.

For events like yours (weddings, fireworks, etc) I'd even go higher (OK -lower, if you follow me), to some - (20-30) dB.

So, too avoid pumping and "clicking":

- no Auto!!!!
- aggressive trimming (around -30 dB); you can safely rise the sound level in post

I hope I was clear enough - forgive me if not, I'm really tired right now :)

Piotr

Brian, the higher the minus number is (-58 vs. -30) the more volume (louder) you get (don't ask me why). So I would think you want to do in the opposite direction.

Try it out. You'll see what I mean.

Wind Filter Info Correction!
 

Ted, I owe you an apology. It appears that I was mistaken in the above statement.

Today I got a Rycote Softie (aka deadcat) for my NTG-3. Miami, being on the coast, is always windy. Even with the Softie, I noticed considerable wind noice. Just for the heck of it, I switched on the Wind Filters. By golly, they worked!

If the mic were in a blimp with a deadcat and the WF applied, it would be dang near perfect.

My apologies to all for the misinformation!

No problem, Jay.

Regarding the Trims, think in terms of what level you are allowing them to trim to. If you are limiting their trim down at -54 or -78, you are hardly letting them trim at all. If you allow them to trim up to -30, then it is trimming more agressively and that is why it sounds quiter.

Now, you wouldn't have to manually adjust these doing run n gun except in the most extreme situations, perhaps. What you are adjusting for is the mic that you use. So, the reason that you start with gain at 5 is to give you the most flexibility after you've calibrated your camera to your mic.

Once the Trims are set according to your mics, then you have your gain control available to tweak for the situation. Just keep peaking low - no more than 2/3 of the scale. Then, if the odd loud noise happens, it hopefully won't clip. When the audio is too soft, then you bring it up in post.

Because you have calibrated the trim for your mics, and recording levels of the gain control aren't cranked, and because the audio system of the camera is good, you'll have a low signal to noise ratio and therefore the ability to bring it back up in post without hearing a lot of noise.

Thanks for the further explanation, Piotr. I've read your last post several times, and it's beginning to sink in now.

Ted, thanks for your explanation, too. That was a great help!

I guess that's what's confusing me--calibrating a microphone. I can calibrate a monitor because a "standard" is provided. I can calibrate the white balance because a "standard" is provided. I can calibrate exposure because a "standard" has been provided.

Is there a "standard" tone somewhere, for example, provided for calibrating mics?

Tone won't work, as it is internal. You are adjusting for the the mic and the environment simultaneously. Perhaps "calibrate" is to optimistic a term - maybe "adjust" is all you can hope for.

With the gain at 5, Allister recommended speaking into the mic. The idea here is to approximate the situation you expect to be in and adjusting the trim accordingly. You can still adjust gain as necessary for the situation. Just keep peaking from getting to close to the top of the scale.

Jay, please take this with a grain of salt. I'm learning right along with you and this is all based on a little bit of background in audio, Allister's recommendations for this camera and my testing. Tomorrow, we might all say, "No, six is where the gain should be when you adjust - whoever heard of five?"

Okay, I'm with you, Ted. The trims aren't a magic bullet. They're just trying to get you within the ball park, whatever it may be. Sound still is what it always has been. I can deal with that.

So, excuse me if I'm going to ask a stupid question - but does this mean that even in run n gun situations one should NOT use automatic levels?

So if I have understood right Piotr and the others are saying that ideally one should:

1)set audio on manual control

2)start by setting the audio levels on 5

3)trim at a level of about -30dB


These settings should give a relatively safe starting point and should the levels start getting too close for comfort one can adjust the audio levels?

Or have I understood the whole procedure wrongly?

Audio is not my strong part.........clearer explanation would be most welcomed!

Yes Brian - you got it right :)