Audio levels should be around -12db?

[Joe Riggs]

 

So, Joe, what does it sound like when you do bring up the gain?

 

Surprisingly, on the boom mic, not as bad as I would expect. Still not ideal.

 

The lavs, however, are really bad. 

 

 

You must be careful doing 32bit processing and after exporting to 24bit life.

 

Please expound on the above.

 

 

When I asked the mixer about the super low levels, he said the reasoning was headroom to avoid peaking (uhh- okay we don't need that much headroom) and that equipment/staging was in the way of getting a better mic placement for a stronger signal ( I could see this for the Boom, but the lavs, placed on the actor's body were low as well).

Obviously I was not satisfied with his answers.

[Philip Perkins]

I record peaks to -12 all the time... you get a feel scene to scene... it's never been a problem....  If it's more an Improv thing, I back off my ISO a bit more, but with a limiter engaged on my mix track.. I have never had an issue...

You are looking at much more accurate metering than that which is part of FCP7 I bet, on a system with much higher headroom.  It's fine for you to go that hot, but active dialog (ie a dramatic scene) that is hovering around -12 on FCP is an ugly clip waiting to happen.

 

philp

[Joe Riggs]

 

It's fine for you to go that hot, but active dialog (ie a dramatic scene) that is hovering around -12 on FCP is an ugly clip waiting to happen.

This has not been the case in my experience, in fact on a previous project, I was getting peak levels around-12 ro -6db, sounded great in FCP! 

[afewmoreyears]

   Funny you say so Phillip, I always ran things a bit on the cooler side, around -18.... I had some conversations with some post folks that said it was faster, better for them to have the levels hotter... Ideal was around -15 or -12.... they said it was much easier for them to drop it down a bit that to gain it up....  Not taking into account burning up a track, which never happened...  Again, some people HATE any limiting, I for one don't in the rare times it occurs... If you should get a surprise, I would rather it limit that moment than record a whole show at -20 because I was afraid of mixing a bit hotter and paying attention...

 

  I don't think  peaks at -12 is that hot... -6 is... but - 12..... for peaks... no big deal...

[Wandering Ear]

I do try to run my levels as hot as reasonable without clipping/limiting, and in post i prefer loud content that i have to turn down, but i much prefer adding gain over dealing with overload or excessive limiting. If the scene is predictable i will run my levels hot, but i don't hesitate to back off if the dynamics are all over.

I agree -12 isn't that hot on an average day.

[Wandering Ear]

" mis-using the term neverclip "

you mean NeverClip ?? that is Zaxcom's system.

I'm generic, and a 32bit sample has even more dynamic range potential than Zaxcom's NeverClip

I was quoting you, so neverclip. If not the Zaxcom system, then what are you referring to?

[Philip Perkins]

If the audio that was peaking to -6 or so in FCP was from a recording system with true peak limiting then you are probably fine (like from an audio post mix).  But very few field recording systems have this feature.  In post I see clips all the time in field audio that has been over recorded.  I for one would rather raise levels a bit (not a big deal at all--very fast) than spend time dealing with clips that the editor might not have seen on the FCP meters or heard on a mid-fi monitoring system played at low levels.

 

philp

[Joe Riggs]

raise levels a bit

 

"A bit" being key here, one should not have to raise levels 20-30db.

[Tom Visser]

"A bit" being key hear, one should not have to raise levels 20-30db.

 

I'd probably agree with you, maybe 10 to 15dB would be my best guess if I had to quantify it.  If the mixer nailed the mix, maybe little or no adjustment, or even some cutting.  Normally the ISO's would be more conservative.  For narrative, I'd tend to record ISO's peaking at -10 and the mix RMS at about -10.  For unscripted television, I'd record ISO's at -20 and the mix at whatever I'm able to manage at the time given the maelstrom of chaos going on around me.

[Christian Spaeth]

Hi Joe,

 

 

I suppose you only started your work on that project after the shoot was done. Your problems could have been quickly solved if you had been there checking on the recorded material after the first few shooting days. Seeing that the levels are too low for you to efficiently work with them, you would have told the on set sound person to set his/her peaks higher. So this is clearly a workflow problem, partly your fault, mainly production's fault, and you should point that out in order to improve matters in the future. No use trying to get the sound person fired in my opinion.

 

But I also challenge you to listen again whether the files really sound that noisy when you add 20 dB of gain. Sometimes that can be quite a subjective judgement. Recording with 24 Bit resolution gives us the possibility to mix at lower levels without risking too much. 16 Bit is a different animal. A guess: the mixer, who was apparently recording into ProTools, was just playing it safe for absence of good limiters? It is always better to get low levels than distorted tracks!

 

Good luck!

[afewmoreyears]

You are looking at much more accurate metering than that which is part of FCP7 I bet, on a system with much higher headroom.  It's fine for you to go that hot, but active dialog (ie a dramatic scene) that is hovering around -12 on FCP is an ugly clip waiting to happen.

 

philp

Dumb question Phillip...

 

  If I am looking at a much more accurate metering system, and my mix is a representation of that accuracy, wouldn't that mean that even though the FCP metering system may say it's HOT and in the clip area it really is not and the information is in fact NOT clipping and distorting?.. Hummm...

[Eric Lamontagne]

Joe,

 

I've been on both sides of this and had various results, gained some friends and lost some also.

 

Where is the Post Production Supervisor? Is this your job or are you the editor or assistant editor? What would be standard procedure from my perspective is to send off some files to post audio for analysis, only they will be able to tell the production if the tracks are workable or not. This is called quality control and this is one of the very valuable jobs a Post Supervisor does. That being said, many of the lower budget shows have cut this position and or don't hire them in time to help with this process during production. I'm imagining you are faced with making a call on this audio for some reason. The quality of your production audio track really not something I could begin to determine based on the information that has been provided, other than to give you some general standard practices similar to what has been said above.

 

I'm in agreement with Graham Timmer and others above. 16bit or 24bit relatively warm audio levels in general make everything downstream sound as good as it can, regardless of the playback medium being FCP, Protools, or some executive producer listing to the dailies on his/her macbook in some hotel room in Dubai. I'm also in agreement that digital gain can also be applied and could save the production audio, obviously better than distorted tracks. These are general statements though and cannot solely be used to determine if your production mixer is doing the job he/she was hired to do.

 

Good Luck!

Eric

[Constantin]

I agree with Eric abd Christian above and to that end I asked you above about the sound once you gained up. You said the boom sounded fine, but the lavs didn't. There are more reasons why lavs sound bad, gain being only one of them. So if the boom really does sound good, what's the problem?

I'm currently in a similar situation, working in post on a horrible production sound. Its levels are way too low, but unlike you case once I bring up the levels, there is a tremendous amount of low freq rumble, mostly boom handling noise. Every shot sounds different, almost no lavs, lots of overlaps, but only ever one person on mic. Awful. Never heard anything this bad

[Philip Perkins]

Dumb question Phillip...

 

  If I am looking at a much more accurate metering system, and my mix is a representation of that accuracy, wouldn't that mean that even though the FCP metering system may say it's HOT and in the clip area it really is not and the information is in fact NOT clipping and distorting?.. Hummm...

Who knows, right?   Did they gain up and render (possibly without remembering or realizing they did this), maybe not the editor but an assistant who set up the session?  Are they listening for problems like peak clips and will their DAC and moni system be transparent enough to hear them, even at low levels and with computers and drives running close by?   In any close you are close to the the max level the entire chain can deal with, so my point of view is to not live so close to the edge, where any gain up error on their part will result in a clip.  I prefer to leave them as little more headroom than that, which seems to work with the posties my stuff goes to (incl myself).

 

philp

[Joe Riggs]

I'm also a little cautious of production sound recorders, who think they know what will work in post.

 

The guy claims if exported with plugins from a DAW, the levels can be raised without adding as much noise.

 

He might be correct, but the tracks should not have to be brought up that much in the first place, or outputted with

any plug ins. Just record the strongest, cleanest, most natural sounding signal.

 

 

 

. Seeing that the levels are too low for you to efficiently work with them, you would have told the on set sound person to set his/her peaks higher. So this is clearly a workflow problem, partly your fault, mainly production's fault, and you should point that out in order to improve matters in the future.

 

How is this my fault? Especially if I come on after the production wrapped.

 

 

I'm sorry but there is a certain level of competence one expects out of a sound recordist:

 

1. They should be recording at decent levels without peaking, what those levels should be is debatable I guess

but anywhere from -18 to -12db and most everyone would happy.

 

2. They should not record below 48k

 

3. They should have an actual sound recorder and not a DAW.

 

#3 has been a big effciency hit, because it can't do polyphonic wavs, and it doesn't always label the takes appropatly.  

 

 

 

You said the boom sounded fine, but the lavs didn't.

 

I said, 

 

Surprisingly, on the boom mic, not as bad as I would expect. Still not ideal.

 

 

So if the boom really does sound good, what's the problem?

 

I would say it sounds acceptable, perhaps with a good post sound hire, it could probably sound good.

Boom was not used for every take, sometimes just lavs. Wide shots - boom can't get close, so have to rely on lavs.       

 

 

 

Where is the Post Production Supervisor?

 

 

There isn't one, It would be good to have a post sound specialist, listen to and evaluate the files.

[Philip Perkins]

I think the levels described here ARE kind of low--the location soundie is being a little over cautious, or maybe has their headphones turned up too hot re their actual record levels or etc..  But that level isn't unusable, it's just a pain for the editor.  Since most location soundies do not want to cause undue pain for editors, letting them know that the levels are sort of below the zone is a perfectly acceptable thing to do.  The 48k thing is a real standard, and someone not recording regular video audio at that SR is either very inexperienced or is making a serious, if possibly unintentional mistake.  Recording on a DAW is recording nonetheless, not what most of us location sound types would prefer to do but the results can be fine for the project.  Some DAWs CAN record in polys, if that's what you want then you should ask if the location soundies' rig can do it.  If there is pushback about this and it is really causing you problems to not have polys then you'll have to escalate the hassle with production, and expect some resistance since it might mean the soundie has to switch to a different recorder mid-project.  Based on these complaints I would say that your location soundie appears to be a video-audio newbie whose experience is in music recording, not a great situation for sure.

 

philp

[studiomprd]

" It would be good to have a post sound specialist, listen to and evaluate the files. "

I thought that was/is you...

cha-ching!

[Joe Riggs]

I am a writer, a doctor, a nuclear physicist, a theoretical philosopher, but above all I am a man. Hopelessly inquisitive man, just like you.

[Aitor Amozarrain]

I don´t participate too often, but I´ve read both threads and while agreeing with most of what has been said, I feel like I´m missing info to play the blame games. Hell, I feel like I´m missing the problem. But I could be wrong.

 

Mono to poly files ---> Wave agent ---> batch process and go grab a coffee while it gets done by the computer --> Then you can keep working as you like. No big deal, No deal at all. I would blame production for not accommodating you, not someone who wasn´t aware of your existance or any workflow preferences. Also, I don´t think that a batch process before syncing would have stopped this production from hiring this mixer or had him use another machine if his wasn´t capable of poly files.

 

Low levels ---> You might be right, but please let us know about S/N ratio if you want a clear bash to that track.  24 bit and a good preamp  do wonders ---> The built in OSX AU for gain or amplification, or whatever is called, is accessible in your NLE and will give you all the dB of gain you might need, without any DAW or resorting to duplicate tracks, hence it is not such a hassle either, even if not the most convenient.

 

I´m guessing the tracks are not too good according to your tale, but IMMO talking specs in this way isn´t going to clear what´s wrong with them. Nor help future sound quality.

[Sean ONeil]

The meters in FCP/Avid etc are only for general reference anyhow--they are not showing you true peak, and they have unknown ballistics.  

 

Hi Phillip,

 

 . . . . general reference . . . general reference . . . . seriously,  . . . . general reference?  The meters in FCP/Avid are only for GENERAL REFERENCE?  W. T. F.?

 

I find it nearly unfathomable, that Digidesign, creators of a product (Pro Tools) that has quite possibly, literally, recorded 80% of all music recorded and transmitted over our corporate controlled (FM) airwaves over the last 10 years  (OK, I admit, these numbers are coming out of my ass here) is providing only a willy nilly, arbitrary, metering system not showing true peaks and with unknown ballistics.

 

While not an experienced postie (FCP) or Pro Tools enthusiast/user, I feel a little naive here, and thank you for the wake up call in raising my awareness to this issue. 

 

1) What makes you come to this conclusion?

2) How do our Sound Devices and Zaxcomm recorder's metering compare?

 

Cheers,

 

Sean O'Neil

Brooklyn NYC

[Marc Wielage]

[Josh Bass]

Ditto.  Can you elaborate on how the meters in NLEs are inaccurate? I don't use Avid but FCP all the time and Premiere to some degree.  What do they do wrong?

[Constantin]

Sean, relax. And don't confuse Avid (as in Media Composer) with Pro Tools

[studiomprd]

JB: " What do they do wrong? "

maybe a better question is: what are proper meters doing right ?

the audio capabilities of FCsP are well known in professional audio circles,   but the metering in most of the usual suspects pro audio applications are adequate, and not particularly "wrong", but they are not detailed or precise enough, thus properly described as : a good " general reference "

[Jay Rose]

I have seen lots of NLE meters (and a few years ago, many audio app meters) written with linear scales: -6 dBFS was in the middle of the meter! Video editors who were digitizing or recording new stuff would record too hot, without enough headroom, just to get what they thought was a reasonable wiggle out of the meter.

 

I have seen a few software meters, even now, written with numbers over zero at the top.... just like a traditional VU meter. Yes, I know it's possible for a digital signal to imply a level slightly higher than FS depending on the output device. But these were exclusive of output device, just measuring the digital stream. And +6 dBFS?

 

As previously pointed out, very few software meters are written with any kind of time integration (not even the 300ms of an analog VU meter); or what there is, is random.  I won't even guess if there's such a thing as true RMS, as opposed to absolute peak, as opposed to occasionally-sampled peak... Overload lights can flash at anywhere between a single -2 dBFS sample to some unknown number of 0 dBFS samples.

 

And to the best of my knowledge, only one semi-mainstream NLE or DAW lets you display waveforms with a log scale that properly reflects audio volume... it's a switchable option in the FOSS Audacity. All other displays are voltage-linear, sometimes rectified, sometimes not.

 

R82? LKFS? Long-term integration? You gotta be kidding.