Making decent recording of lecturer in a hall
 

The lecturer will be lav miked. But recording only the
lav will be dry. I want to record the spaciousness of
the hall with stereo miking out in the audience.
The stereo is also for ambiance/audience reaction.
The lecturer's voice is amplified through large
loudspeakers at the ceiling level, 30 feet up.
The problem is that when I try to mix the
ambiance recording with the lav recording in
post, it sounds echo-y. If I 100% mute out the
ambiance stereo mikes while the lecturer is
speaking it collapses the stereo sound field.
If I mute out the ambiance while the lecturer
is speaking, and then lift if for applause, you are
switching between stereo and mono, which will
sound unusual to the listener.

I've heard someone say to make an ambiance
recording of the room noise for a minute before
any audience gets there, and then loop it over
in post throughout the lecture, but then you don't
get the audience reaction.

Anyone have any thoughts on this? Am I making
some mistake in how this should be miked?
Or do I just have to turn the ambiance mikes
down more in post and live with the echo?
Would miking the audience mono instead of
stereo be of any help?

How about trying these variations on the technique sometimes used to clean out vocals when preparing Karaoki tracks ... don't know what NLE assets you have but some audio editing programs allow you to process a stereo track into a fresh track with each channel only containing the part of the signal that is different from the other channel. Since the room and audience sounds will be different on each mic but the sound from the ceiling speakers will be similar at both, the resulting track will have the speaker's voice suppressed while preserving the audience and room sounds.

You might also try inverting the phase of the track recorded with the lav and mixing it with both channels of the ambiance stereo track, nudging it back and forth and adjusting the levels until it cancels the speaker's voice recorded in the those channels. When you get a good cancellation, bounce the results to a fresh ambiance track. Now mix the original, un-inverted lav track with the new ambiance tracks as usual.

Room tone and ambiance are not the same thing. The advice you got was for room tone, which is certainly a good thing to have to on hand, but it's not the same thing as the audience ambiance you're wanting to capture. OTOH, I don't know why stereo micing on the audience would be especially important - general audience sounds aren't localized and even a mono track can be panned equally to left and right during post. This lack of localization would be especially apparent while the speaker is talking - your camera is going to be on him, shooting more or less from the audience's perspective, and most if not all the audience will be behind you and not visible to the camera. The audience would be relatively quiet and so your sound track will be dominated by the speaker with the audience sounds at most a low level and diffuse background murmur. It's only when you're doing the reverses and showing audience reaction shots that stereo imaging might be be a factor. About the only time I think I'd worry too much about it would be if I was planning on mixing to 5.1 surround where you have localization in all directions and there is something really different taking place in different parts of the audience - say, a political speech with hecklers concentrated in just one part of the room. Otherwise recording the audience in mono and panning it to both channels in the mix should be fine. And remember too, depending on your distribution a lot of your intended audience is likely to be hearing it in mono anyway. Even if they have two speakers a lot of TVs blend both stereo channels into mono. Even if it doesn't mix down to mono, many of the theoretically "stereo" TVs have the speakers too close together to give any signifigant stereo imaging. About the only people that will really hear it in stereo are people with relatively high-end TVs, home theatre systems, or who view it on a computer with stereo speakers - for everyone else it won't matter much what you do.

try pulling the audience track up a bit (time shift-earlier) so that you reflections in the hall line up better with the speaking track. this will make the echo less noticable. if the pa is very loud, you will still have to duck the audience a bit when there is no reaction happening. if you listen closely you'll find this happens in a lot of broadcast shows (academy awards comes to mind-sorry ed...).

other options would include lowering the pa level, or spread it out over more speakers at lower level (not usually feasable). or use many audience mics mixed into a stereo track-hopefully increasing the audience to pa ratio. i've done as many as 12 mics on an audience...

Is this all for future planning or have you actually recorded it?
If this is still in the planning stage, there are two important components to the presenter wearing a lav to keep in mind. First, the lav track may not be as dry as you think it will be. This is a double-edged sword. It will get you more audience response but it leads to the next problem which is: Second, a lav and PA are always a very delicate balance unless significant effort was put into taming potential feedback. You may not only be dealing with the presenter echo in your audience mics, but the presenter's PA sound may be a significant problem in his own lav track.
As others have mentioned, going mono for the audience and keeping the PA output as low as possible for a good presentation are important ways to minimize your problem.
Lets assume you can't control the PA speaker placement. A micro head-worn mic will give you much better sound than a lav in any PA situation. But this will also make the presenter's track even drier in relation to the audience reaction.
For the audience, I would use a directional mic or mics that have clean off-axis response. I have used a Rode NT3 for this effectively even though generally the smaller the diaphragm the better off you are in this situation where PA sound is mixing in with the live audience. I have also used AT873r and EarthWorks SR71 mics in these situations where time coherence is more important than self-noise. Placement and direction of these mics are also very important.
Lastly I'll repeat what I usually say in this situation where recording and PA take place together. It's very important to keep the two pathways as separate as possible. I take two signals from the mic preamp, one to be controlled for recording and one for PA. The only control that affects both is the preamp input trim, which is set properly ahead of time. Then you can do whatever you need to each individual signal without causing a problem in the other.

Dave,

Use a Countryman E6 on the lecturer and have a runner or two with mics to get the questions.

Fageddabout the stereo.

Ty Ford

There are no questions, just applause, laughter, etc.

No audience responses? Tnen double fageddabout it!

Why even bother with stereo? Stick up a couple of mic for the house and be done with it.

Ty

So, your recommendation would be to place a hyper
at the stage facing back at the audience, and just
run it low in post (~ -24dB)? What about putting an
omni at the back of the audience for ambiance, and
mixing that low with the lecturer's lav?

e6 and mono-stereo
 

Ty's suggestion for the e6 is a great idea.

We use one at our church. Because of the way it is positioned it does a pretty good job of picking up some good ambience (especially laughter).

here's their web site if anyones unfamiliar with this mic. I don't think you can buy it directly from them, we got it from some place in chicago.

http://www.countryman.com/html_data_sheets/e6data.html

RE: Mono to Stereo
(something simple I've tried) --I'm sure I'll prove that I am not a live sound expert with this post.

In this setting (not for recording a group of musicians), I've doubled the mono track (panned them right and left) and slightly offset one track, or applied a tiny bit of reverb to one of them.

I think that may be part of the brilliant solution offered in one of the earlier posts, but it's brillaince was too advanced for me, so I'm not sure. (Don't worry Steve, I've printed it up and I plan to read it later after I've had more coffee and put my sick babies down for a nap).

Dave,

You seem to be very concerned about ambience. Intelligibility rules for lectures. Ambience can only decrease intelligibility.

If you can record some ambi and mix it in in post (and have the budget for post) fine. Most of my clients who want this sort of work don't have budget for post. You get it live to CD or to HD.

Regards,

Ty

Thanks, Greg, hadn't thought of trying a time
shift, but it sounds like it might offer some
help here.
Now, about "ducking the audience" such as is done
at awards shows, are you meaning to just
pull down the level of the audience mikes in
post (during the times when there is really
not much going on other than murr-murr)?
And then lift the audience mikes when there is
reaction, such as applause or laughter? That's what the
awards shows do?

Another practice for post shows is to drop the high end of the audience so that there aren't any conflicts in the sibilant and intelligibility ranges of the main speaker/focus subject. Some NLE's/DAWs can automate this; others can't. But it does make a discernable difference. On the "psychological side," viewers will actually feel a sense of "settlement" when this occurs following applause, laughter, oohs, ahhs...

This makes sense, too. Do they drop the highs
just when the presenter is speaking? Or is it
throughout the whole show, that the audience
mikes have attenuated highs?

Quote:
Originally Posted by Douglas Spotted Eagle
.... Some NLE's/DAWs can automate this; others can't. ...

Which NLE's/DAWs can automate this? (I use Vegas)

Vegas can do this, since version 4.
If you right click an audio header and choose "FX Automation envelopes" you'll be able to insert an envelope that will control any FX parameter that is automatable. This works for most 3rd party plugins as well, provided they have automation capability. (WAVES, iZotope, Kjearhus, etc)