Location sound; questions regarding recording methods and appropriate dialogue levels

[Jaws]

Hi Fellow sound guys! I am just getting into Location Sound Recording, and have done a few short low budget films. I have some questions that I am having trouble figuring out on my own, and I apologize in advance for my long winded question's/ramble.

My two main questions focus on recording methods and appropriate dialogue levels.


1) When recording a master shot in a dialogue based situation where the DP has made the frame too wide to be able to get close boom coverage providing low S/N ratio, and their aren't enough wireless radio mic's (or channels on the mixer(SD302)) to cover all the talent. What is the appropriate response and practice to still record quality usable dialogue, or is this just not an option based on the provided set up?


As a general question, what do people do when recording master shots with dialogue?
Typically when recording more than three actors I have put two wireless mic's on the talent and tried to boom from up high, but this obviously only covers some of the dialogue with the wireless and then the rest of the tracks are echoey and with a high noise to signal ratio. Do other mixers turn up their gain and just deal with the recorded noise in post? I have tried planting mic's but this is difficult to count on when concidering available production time, sporadic actor movements, and the physical reality of a location. When communicating with the picture department, we have agreed to just use the sound from the Master shots as a Guide track for intercutting the mediums and close ups or cutting out the recorded close-up dialogue and editing it to fit the master later. Although while considering my limited experience with editing sound, when using just lapel mic's in a Master and then trying to mix it in post the shotgun and lavs are quite difficult to match, even when adding reverb to the lapels. Additionally, I know it is important to maintain constant recording levels throughout a scene, when considering that a mic can't be placed close to the actor's mouth in a master shot, do mixer's turn up the Pot to make up for the difference to a medium shot regarding actors levels?


2) I also tend to have dramatically fluctuating signal levels when recording dialogue. In general I find that they are too low. In my last shoot I made a serious effort to request that the talent speak louder when their levels were coming up to low on the mixer. Is this a simple reality of doing sound mixing, or is there possibly something I am missing?


Here is my typical set-up;




MKH-416 shotgun mic AND TR-50 lavs with if needed G3 Transmitter/Receivers (once Lectrosonics)


-mic in- with48 phantom power, to-


Sound Devices 302 mixer (three channels)


-Line out to-


Tascam DR-100 Mkii Recorder (two channels) WITH a confidence return to the MIXER (which I stay on when I am recording)


Recording 24/48 quality onto two channels with a scratch track to the camera typically using a 1/8 cable.


At the beginning of each set-up I check my levels sending a 1k tone through the loop to calibrate the mixer to the recorder


For whispers or soft Dialogue I try to have levels between -30dBFS and -20dBFS
Average Dialogue between -20dBFS and -12dBFS
Loud Dialogue between -12dBFS and -6dBFS (placing the limiter around -3 on the SD 302 mixer





I would really appreciate any comments or advice that people have.
Thank you for your time and for reading my post!!

 

[Rick R]

"placing the limiter around -3 on the SD 302 mixer"
-- Does this mean the software limiter setting? If so, that would considered excessive by most folks, unless your levels are extremely low. But then you would have other gain stage issues. I set the 302 output limiter to around +18dB.
Everything else seems fairly logical though.
BTW, using a dark colored font makes your post difficult to read

 

[Jaws]

"BTW, using a dark colored font makes your post difficult to read"
Thanks for the heads up! This was my first post and I assumed that it would look normal when it was put online.

Regarding the Limiter, I made an error in my writing (I was thinking in terms of the recorder). And yes I did mean the software limiter, I intended to write -
that I set my Software limiter on the 302 to a setting of +18dB on the mixer.

Thanks for catching that!

 

[Richard Crowley]

I apologize in advance for my long winded question's/ramble.

Consider also apologizing for posting in black print over dark gray which is nearly impossible to read. (You can go back and fix it.)

1) When recording a master shot in a dialogue based situation where the DP has made the frame too wide to be able to get close boom coverage providing low S/N ratio, and their aren't enough wireless radio mic's (or channels on the mixer(SD302)) to cover all the talent. What is the appropriate response and practice to still record quality usable dialogue, or is this just not an option based on the provided set up?

1) Catch the dialog in the close-up shot(s)
2) Rent more wireless mics
3) Reduce the number of speaking parts
4) If you are using a "locked down" camera, you could consider masking the visible boom with "clear" replacement footage during editing.

As a general question, what do people do when recording master shots with dialogue?

It is not uncommon to not even attempt "clean" dialog in a wide master shot for the very reason(s) you outline. A "scratch" recording is always valuable, of course, for sync and continuity, etc.

Typically when recording more than three actors I have put two wireless mic's on the talent and tried to boom from up high, but this obviously only covers some of the dialogue with the wireless and then the rest of the tracks are echoey and with a high noise to signal ratio. Do other mixers turn up their gain and just deal with the recorded noise in post?

There is no good way to "deal with the recorded noise in post". You have been painted into a corner. You are simply attempting to do something with inadequate resources.

I have tried planting mic's but this is difficult to count on when concidering available production time, sporadic actor movements, and the physical reality of a location. When communicating with the picture department, we have agreed to just use the sound from the Master shots as a Guide track for intercutting the mediums and close ups or cutting out the recorded close-up dialogue and editing it to fit the master later. Although while considering my limited experience with editing sound, when using just lapel mic's in a Master and then trying to mix it in post the shotgun and lavs are quite difficult to match, even when adding reverb to the lapels. Additionally, I know it is important to maintain constant recording levels throughout a scene, when considering that a mic can't be placed close to the actor's mouth in a master shot, do mixer's turn up the Pot to make up for the difference to a medium shot regarding actors levels?

Trying to intermix lavs on some subjects, and others picked up with shotgun(s) is not easy even under ideal conditions. The situation you describe is very far from ideal. IMHO, it is not reasonable to expect to intermix lavs and shotgun, especially with amateur talent and/or amateur crew.

I also tend to have dramatically fluctuating signal levels when recording dialogue. In general I find that they are too low. In my last shoot I made a serious effort to request that the talent speak louder when their levels were coming up to low on the mixer. Is this a simple reality of doing sound mixing, or is there possibly something I am missing?

Rehearsal? The boom operator and the production mixer should know what the dialog (and delivery) will be for the shot before hand.
Adequate crew? Are you the dedicated production mixer? Do you have a decent boom operator? Does the boom operator have headphones to monitor what they are doing?

Here is my typical set-up;
MKH-416 shotgun mic AND TR-50 lavs with if needed G3 Transmitter/Receivers (once Lectrosonics)
-mic in- with48 phantom power, to-
Sound Devices 302 mixer (three channels)
-Line out to-
Tascam DR-100 Mkii Recorder (two channels) WITH a confidence return to the MIXER (which I stay on when I am recording)
Recording 24/48 quality onto two channels with a scratch track to the camera typically using a 1/8 cable.

Seems reasonable, except not enough mics/channels for productions that have many speaking parts (which seems to be what you are describing)

At the beginning of each set-up I check my levels sending a 1k tone through the loop to calibrate the mixer to the recorder

OK. Or perhaps calibrate the recorder recording level to the mixer reference level.

For whispers or soft Dialogue I try to have levels between -30dBFS and -20dBFS
Average Dialogue between -20dBFS and -12dBFS
Loud Dialogue between -12dBFS and -6dBFS (placing the limiter around -3 on the SD 302 mixer

IME, recording levels should always be optimized for maximum signal-to-noise ratio (SNR).
Dialog levels in the final mix should be handled in post-production editing, not by trying to second-guess the mix during shooting.

 

[Jaws]

<This is a Re-Posting of my above opening text with a change of font color for ease of reading. Thanks to everybody for the recommendation>

Hi Fellow sound guys! I am just getting into Location Sound Recording, and have done a few short low budget films. I have some questions that I am having trouble figuring out on my own, and Iapologize in advance for my long winded question's/ramble.

My two main questions focus on recording methods and appropriate dialogue levels.


1) When recording a master shot in a dialogue based situation where the DP has made the frame too wide to be able to get close boom coverage providing low S/N ratio, and their aren't enough wireless radio mic's (or channels on the mixer(SD302)) to cover all the talent. What is the appropriate response and practice to still record quality usable dialogue, or is this just not an option based on the provided set up?


As a general question, what do people do when recording master shots with dialogue?
Typically when recording more than three actors I have put two wireless mic's on the talent and tried to boom from up high, but this obviously only covers some of the dialogue with the wireless and then the rest of the tracks are echoey and with a high noise to signal ratio. Do other mixers turn up their gain and just deal with the recorded noise in post? I have tried planting mic's but this is difficult to count on when concidering available production time, sporadic actor movements, and the physical reality of a location. When communicating with the picture department, we have agreed to just use the sound from the Master shots as a Guide track for intercutting the mediums and close ups or cutting out the recorded close-up dialogue and editing it to fit the master later. Although while considering my limited experience with editing sound, when using just lapel mic's in a Master and then trying to mix it in post the shotgun and lavs are quite difficult to match, even when adding reverb to the lapels. Additionally, I know it is important to maintain constant recording levels throughout a scene, when considering that a mic can't be placed close to the actor's mouth in a master shot, do mixer's turn up the Pot to make up for the difference to a medium shot regarding actors levels?


2) I also tend to have dramatically fluctuating signal levels when recording dialogue. In general I find that they are too low. In my last shoot I made a serious effort to request that the talent speak louder when their levels were coming up to low on the mixer. Is this a simple reality of doing sound mixing, or is there possibly something I am missing?


Here is my typical set-up;




MKH-416 shotgun mic AND TR-50 lavs with if needed G3 Transmitter/Receivers (once Lectrosonics)


-mic in- with48 phantom power, to-


Sound Devices 302 mixer (three channels)


-Line out to-


Tascam DR-100 Mkii Recorder (two channels) WITH a confidence return to the MIXER (which I stay on when I am recording)


Recording 24/48 quality onto two channels with a scratch track to the camera typically using a 1/8 cable.


At the beginning of each set-up I check my levels sending a 1k tone through the loop to calibrate the mixer to the recorder


For whispers or soft Dialogue I try to have levels between -30dBFS and -20dBFS
Average Dialogue between -20dBFS and -12dBFS
Loud Dialogue between -12dBFS and -6dBFS (placing the limiter around -3 on the SD 302 mixer





I would really appreciate any comments or advice that people have.
Thank you for your time and for reading my post!!

 

[Bruce Watson]

At the beginning of each set-up I check my levels sending a 1k tone through the loop to calibrate the mixer to the recorder

For whispers or soft Dialogue I try to have levels between -30dBFS and -20dBFS
Average Dialogue between -20dBFS and -12dBFS
Loud Dialogue between -12dBFS and -6dBFS (placing the limiter around -3 on the SD 302 mixer

A 302 mixer is good for about +20dBfs before it overloads/clips. So you use the 302's onboard 1k tone to set the downstream recorder to -20dBfs. Then you set the 302's limiters to +18 or +20 dBfs. This way even if your downstream devices starts clipping at 0dBfs, the mixer/recorder combination still won't (actually, it can't in a brute force kind of way) clip. So you can "set and forget" the recorder, and just deal with the mixer during acquisition. And during acquisition, you'll need to set your levels so that you just barely go into the red on the 302, which is about +10dBfs IIRC. What's nice about this is that it'll be fairly easy to see at a glance. It would be easier if SD used yellow LEDs for the region where they use orange now, but what can you do?

I know, you probably think you should go no higher than 0dBfs on the mixer. But that's not how the SD mixers are designed. If you want a good signal to noise level using SD mixers, you've got to turn the signal levels up and use 'em the way the engineers designed 'em.

I saw this exact same explanation recently on another board from none other than Ty Ford. I think it was about a MixPre-D and not the 302, but there's hardly any difference internally and the SD philosophy is the same in any case. Might have been in the CreativeCOW audio forum, I don't remember.

If you don't like this explanation, search for the term "gain staging". It'll get you to the exact same place, but said somewhat differently.

 

[Bruce Watson]

What Richard says about having enough equipment to do the job. If you've got, say, five speaking parts and a wide shot, you need five lavs and a mixer that can handle five lavs. No way around it. You can't post-process your way out of it either.

 

[Alex H.]

A 302 mixer is good for about +20dBfs before it overloads/clips. So you use the 302's onboard 1k tone to set the downstream recorder to -20dBfs. Then you set the 302's limiters to +18 or +20 dBfs.

I know, you probably think you should go no higher than 0dBfs on the mixer. But that's not how the SD mixers are designed.

dBFS is NOT a scale used by the Sound Devices mixer. That means dB Full Scale, which is the digital scale that caps out at 0dB and cannot go any further than that. The mixer uses an analog scale, and the TASCAM recorder uses full scale.

 

[Bruce Watson]

dBFS is NOT a scale used by the Sound Devices mixer. That means dB Full Scale, which is the digital scale that caps out at 0dB and cannot go any further than that. The mixer uses an analog scale, and the TASCAM recorder uses full scale.

Yes, you are correct. I was writing too fast. Again. The SD mixer shows VU metering, yes? I don't remember the correct notation for that, but it can't be as easy as dBVU can it? In any case, VU is a hold over from the magnetic tape years. But that's why you turn the gain up and record with the VU meter in the +10dB range. Mag. tapes would saturate out over a fairly wide range (say, +3 dB to +15dB or so) depending on the tape itself, the width of the track, the speed of the tape over the recording head, etc. All that is gone now, but the VU meter remains. Go figure.

 

[Jaws]

Thank you so much for your prompt reply Richard.

Rehearsal? The boom operator and the production mixer should know what the dialog (and delivery) will be for the shot before hand.
Adequate crew? Are you the dedicated production mixer? Do you have a decent boom operator? Does the boom operator have headphones to monitor what they are doing?

I do ask for rehearsals before each take (Unfortunately, i don't always get one.)
On the shoots I have worked on I am the dedicated production mixer, boom operator, and utility person. I should have included this above, I use the standard Sony 7506 headphones, and I always monitor as I am recording.

I don't think I succinctly described the exact problem I am having with levels. - I find that in general the levels are too low on the mixer, and then upon viewing all the files on my DAW later they are still generally pretty low (sometimes averaging around -18) . Although, with occasional dramatic peaks that will go quite high. Is that normal to have these kind of levels and to then raise them appropriately in post? My guess is no, (raising dialogue levels all so raises the noise floor). This is why I asked the actors on my last shoot to project louder. To clarify regarding possibility of mic placement problems; I try to get as close as possible, but generally when booming- I go a foot and a half away to two feet on close ups, aiming at the upper chest (for even mic coloration), positioned above the talent.

OK. Or perhaps calibrate the recorder recording level to the mixer reference level.

I am sorry, that is what I intended to write. I send the 1k signal out of the 302 mixer to my Tascam recorder and then adjust the levels of the (digital) recorder to -20dB.

 

[Jaws]

Thank you for your advice Bruce.

I do aim to get my levels into the red, or to the appropriate placement, i.e. peaking at -12dBFS on my recorder for normal dialogue. I find this difficult to maintain throughout an actors performance though, and I have read <The Location Sound Bible, by Ric Viers>, that when recording it is better to set an actors levels at the begining of a scene, rather than "ride" the pot throughout each take. Aside from levels fluctuation, I also find that in general it's difficult to achieve "good" level's in a take without turning up the Pot past the 3 o'clock mark. (which I also understand is not recommended because of introducing excessive system noise).

Thank you to everyone for their comments so far!

 

[Rick R]

Yes, the calibration tone on the 302 is (normally) 0VU. The meter setting on digital recorders (-20dBFS these days ) was once (long time ago) refereed to as digital zero.
There is also a 'full scale' method of calibrating the mixer>recorder levels, but the '0VU= -20dBFS' will suffice for most.
"five speaking parts and a wide shot, you need five lavs and a mixer that can handle five lavs".
- Preferably with a recorder with five open tracks> Running multiple mics to a single track can be trouble. If there is a RF hit or other problem, the track is trashed.

 

[Richard Crowley]

On the shoots I have worked on I am the dedicated production mixer, boom operator, and utility person.

"Dedicated"? No, I don't think so. Nobody can mix more than one mic AND wield the boom at the same time.
You simply have expectations that exceed your resources: equipment, personnel, experience.

 

[Jema]

For whispers or soft Dialogue I try to have levels between -30dBFS and -20dBFS
Average Dialogue between -20dBFS and -12dBFS
Loud Dialogue between -12dBFS and -6dBFS (placing the limiter around -3 on the SD 302 mixer

As others have touched on, but it deserves to be underlined, is that you should always record at around the "average" that you describe, with peaks somewhere between -10 and -6, regardless of the "actual" level of the dialogue.

I find that in general the levels are too low on the mixer, and then upon viewing all the files on my DAW later they are still generally pretty low (sometimes averaging around -18) . Although, with occasional dramatic peaks that will go quite high. Is that normal to have these kind of levels and to then raise them appropriately in post? My guess is no, (raising dialogue levels all so raises the noise floor).

-18 doesn't seem that low to me for raw audio. If you want it higher in post, you simply raise it - balancing the levels from different takes and making sure they're at the level you want is really basic for sound post. Raising something in post doesn't affect the snr, and if the background noise is too loud, then the problem lies in the recording. You either had a bad location, or was too far away with the mic, or the actors too quiet (or recorded it too low on the recorder, but that doesn't seem to be your problem).

Aside from levels fluctuation, I also find that in general it's difficult to achieve "good" level's in a take without turning up the Pot past the 3 o'clock mark. (which I also understand is not recommended because of introducing excessive system noise).

This depends entirely on the mic sensitivity and the level of the dialogue. SD preamps are very quiet and won't introduce much noise, but the fact that you need to turn them up high might indicate that the person is speaking too low, or you are too far away.


One person cannot be expected to boom and at the same time mix more than two people. It sounds like the expectations were unreasonable - you need more equipment and preferably a second sound person.

 

[TyFord]

BTW, using a dark colored font makes your post difficult to read

THANK YOU!

Regards,

Ty Ford

 

[Steve House]

It should be pointed out that because of the different energy distributions within the waveform, sinewave tones and real speech behave differently on the peak-reading meters most digital recording devices are equipped with these days. This is a source of the confusion between the "align to -20" versus "align to -12" schools. A 'real' waveform will read about 8 dB higher on such a meter than does a pure sinewave with the same average energy. So if you calibrate where 0VU = -18dBFS using the mixer's sinewave tone, then switch to speech that also averages 0VU on the mixer meter (ballistics set to averaging), the recorder's peak reading meter will indicate about -10dBFS (-12dBFS if the alignment level is -20dBFS). You can easily see this if you have a mixer like some SD mixers that have meters that can show average and peak levels simultaneously - on sinewave tones the two align but on speech the peaks are about 8dB above the average. Aligning where the 0VU sinewave is set to -12 on the recorder means you're going to be peaking up in the -4 dBFS range on speech, way too close to clipping to give you any 'wiggle room' for increases in volume. Aligning the mixer's 0VU tone to -18 or -20 on the recorder or camera's meter, then setting the mixer's faders so real speech bounces between 0VU and +4VU should give you the best of both worlds, a strong recording level to give the best signal/noise ratio yet not so strong as to tickle the dragon's tail on momentary overs.

 

[TyFord]

Yes, but that's not really the point. The point is to calibrate the mixer to the camera/recorder so you don't have to observe the camera's/recorder's meters in order to send it proper level. Once that's done, sound of any sort passing through the mixer can be properly controlled so it doesn't clip the camera/recorder.

Regards,

Ty Ford

 

[Noiz2]

Thank you for your advice Bruce.

I do aim to get my levels into the red, or to the appropriate placement, i.e. peaking at -12dBFS on my recorder for normal dialogue. I find this difficult to maintain throughout an actors performance though, and I have read <The Location Sound Bible, by Ric Viers>, that when recording it is better to set an actors levels at the begining of a scene, rather than "ride" the pot throughout each take. Aside from levels fluctuation, I also find that in general it's difficult to achieve "good" level's in a take without turning up the Pot past the 3 o'clock mark. (which I also understand is not recommended because of introducing excessive system noise).

Thank you to everyone for their comments so far!

Couple of things. First you should take into account that Ric Viers doesn't do production recording (well very little) so you may want to use it as just one opinion and less of a "bible". I like Ric and his book may be excellent but you have to take any book with a few grains of salt. What he does do is a ton of FX recording. Which is mostly what I do with some music tossed in. Production is a bit of a different beast.

Also "3 O'Clock" doesn't mean much. That would be relative to where the knob starts it's rotation. 3 O'Clock on my MixPre is only about 1/3 of the rotation. As a general rule mixers and such are at their best between 50% and 75% of their range. But use your ears. Really good mixers/ recorders can be very usable much higher than that and some consumer gear gets noisy very fast.

For the kind of work you are describing I would set the input level on the recorder to a good place (in that 50% - 75% range) then set the limiter on the mixer so that you are "clip proof". Your a one man band and you need a safety net and the limiter is it. Once that is set you don't touch the recorders levels and use the mixer to get a strong signal. If your hearing a lot of noise then something needs to change but probably it's mic placement, because you should be able to record pretty hot.

As long as you don't clip hot is not a problem for post. Low levels though are hard to deal with. If the recordings are very noise free and they are not too low raising the level in post is fine. But usually low levels are also full of noise and then there just isn't a good solution.

 

[Noiz2]

Tascam DR-100 Mkii Recorder (two channels) WITH a confidence return to the MIXER (which I stay on when I am recording)

You are probably not actually in "confidence" mode. There is a sizable time lag with most recorders in actual confidence mode since it has to write and then read the file. So you get a bit more info coming off the recorder, as in you will hear if the mixer is clipping the recorder inputs, but you are not actually listening to the recorded file so for instance you can't tell if you are in record pause or actually recording.

I don't actually use real confidence mode because the delay drives me crazy and can be very disorienting (and that can be dangerous in FX work). But you should keep the above in mind so you don't think you are recording when your not.