Tascam Dr-100 MK II - Operatic Event in 2 days

[kwkeirstead]

Hi, I have two concerts to video Saturday nite and Sunday matinee (opera singers/15-member orchestra) and could benefit greatly from advice/experience with the following proposed Tascam DR-100 MK II setup.

I did a trial recording on opening night with a VideoMic Pro on mycamera and you can get an idea what the ambient sound is like at a distance of 40 feet. The highlights extract is very short in duration.

https://vimeo.com/129101764

The problem is the brass and more particularly the piccolo go over the singers, which surprised me, given you can usually hear opera singers in the next county.

So, my plan it to put a Tascam DR-100 MKII at the front of the stage (on unidirectional) with the mic pointing toward the singers. With this setup, the orchestra will be in three rows 5-15 feet behind the mic.

I will be able to simulate this to an extent by getting my wife to stand in our outside porch 5 feet away from the mic, with the TV turned up in the TV room (first omni, then directional) but no clue right now how much of a difference this will make or whether the simulation will be anywhere close to conditions in the hall.

The hall, actually, is quite nice. It has a capacity of 200 (including a balcony) and no echos. They have no sound board.

Any suggestions other than why did I not bring the Tascam to the venue for my test. Answer : Supidity

 

[kwkeirstead]

The lighting does not change in this production but the hall is part of a community center and some of the lights are burned out so depending on where people stand, there can be dark spots.

I should point out the clip above is a first time recording off a new Libec tripod/head.

One finger is about all it takes to go from 0-60. My previous tripod practically needed a crowbar to get it past the starting friction so I will need to be very very careful.

Please feel free to bash the video clip. I want this recording to be as good as it can be within the "one-trick pony" constraints I have to work under.

The cam, by the way, is a Panasonic AG-AC160A. I am at 1080/60 24p.

While we are at it, I notice a few dark scenes in my clip, so I probably should put the gain at +6 and pay close attention the the waveform.

I was very interested to learn that some of the newer cams automatically apply gain when you hit open.

Does that work well or are we talking about yet another auto feature that is best turned off?.

 

[Noiz2]

THere are two possible problems that jump out at me. First the distance is going to mean you will need to post sync the two sound tracks, but then I guess you would anyway... But as a general note remember that sound travels a LOT slower than light.

The second you won't know till you try to mix them, and that is you will probably get some strange phase issues. They may not be a big issue hard to tell. In theory if you can get the voices to be in sync then the stage recorder is not going to pick up a lot of the orchestra so? Because of how you are planning to set this up you will have things coming in at three different offsets. THe onstage sound through the onstage recorder then the orchestra bleed to the onstage recorder and then the camera mic. As long as the orchestra bleed is not very loud you can probably make it work.

This is not a great set up and it's not likely to sound great, but I think you know that. It might be good enough though, kind of depends on what your expectations are.

Other issues you may have are that IF the DR100 is anywhere close to the floor and it's a sprung wooden stage (probably) you may pick up a LOT of foot traffic. Also probably not a problem from where it will be but stages with a fly loft (if curtains and such can be raised up above the proscenium arch it's a fly house), it can get very cavernous sounding if you mic picks up any sound from the loft.

Also your problems with balance may have to do with the mic. I don't know but because of what and who it was designed for I wouldn't be surprised to find that the VMP is a bright mic with out a lot of low end. That is going to make things like piccolo's stand out and will diminish big bass voices. It won't really "fix" it but a little EQ in post might tame that a bit and bring things back in balance.

Last note. THe syncing you are going to have to do, you should do in a sound editing program, or with PluralEyes. Most video editing apps will only let you move things in whole frame increments and to get these in phase you are going to need a much finer resolution than that. Sound apps will let you move stuff at the sample level. The difference of about 1/30 of a second vor a NLE VS 1/48,000 of a second for a DAW.

 

[kwkeirstead]

Thanks Noiz2. I have experienced the creaking floor problem so best have a stool not on the stage for the mic. Good job you reminded me of this problem.

As for synch issues, I use the on cam mic only as a reference and then feed this plus the Tascam signal into PluralEyes. I have found that it converts half a day of work down to a few minutes. It won't work on very short clips but I never have any of these.

 

[Steve House]

Thanks Noiz2. I have experienced the creaking floor problem so best have a stool not on the stage for the mic. Good job you reminded me of this problem.

As for synch issues, I use the on cam mic only as a reference and then feed this plus the Tascam signal into PluralEyes. I have found that it converts half a day of work down to a few minutes. It won't work on very short clips but I never have any of these.

The point is sound travels far slower than light, about 900 FPS. Imagine your camera with its mic (that you're using for sync reference) is 30 feet from the sound source. The light from the source arrives at the camera essentially instantaneously while the sound arrives about 1/30 of a second, one full video frame, later. That means your sync reference track is one frame late. PluralEyes will line up your production sound with the delayed sync waveform, not the video frames, and so your final sound will be out of sync with picture by the same amount as the in-camera scratch track. If I recall correctly, typical broadcast standards allow for a +0 to -1/3 frame maximum deviation from dead on.

 

[Rick R]

"The light from the source arrives at the camera essentially instantaneously"
-- With the wearer's wireless mic signal not too far behind (unless they're using a new 2.4Ghz system.. in particular, the new Sennheiser DI, with an alleged latency of a whopping 19ms !

"the sound arrives about 1/30 of a second, one full video frame, later
Yes, around 1 millisecond per foot.. easier to calculate and compensate (for audio folks anyway).

 

[patrickbaldwin]

Last note. THe syncing you are going to have to do, you should do in a sound editing program, or with PluralEyes. Most video editing apps will only let you move things in whole frame increments and to get these in phase you are going to need a much finer resolution than that. Sound apps will let you move stuff at the sample level. The difference of about 1/30 of a second vor a NLE VS 1/48,000 of a second for a DAW.

FCPX will allow subframe editing of the sound at 1/80th I think.

 

[Noiz2]

FCPX will allow subframe editing of the sound at 1/80th I think.

FCP would let you "edit" at a subframe rate but that was a blind edit. You couldn't for instance shift by a subframe while looking at picture. PluralEyes uses that feature to do it's syncing.

FCPX may have improved this over FCP? Since I do sound post and FCPX doesn't allow for export to other systems I haven't used it or had any projects come from FCPX yet.

For the OP he could probably use that feature to make a calculated shift in his on camera sound to compensate for the lag and then use PluralEyes to sync the stage recorded sound.

Sub frame would either be 1/80 or 1/100 of a frame, I don't think there is a hard standard - I've seen both. It's good enough for lip sync but it's probably not good enough to get rid of phasing, though maybe. Probably depends on how close the subframe gets you to dead sync. And as I said if you don't have a lot of cross bleed it's probably not going to be a killer.