Challeneged! Levels for Post

[Philip Perkins]

I'd be interested to hear where you guys put your high and low pass filters.

No low pass at all ever.  High pass out (if it's a quiet set and the mic is on a stand, like an interview, esp if the speaker is adult male), or starting at 40 or 80 12 db per oct. shelving if I'm holding the boom.  Starting @ 160 if there is a lot of BG rumble and the speaker is female.  After that it's just get-what-you-can verite stuff--whatever it takes.

 

philp

[Jaymz]

I'd be interested to hear where you guys put your high and low pass filters.

 

Depends on the mic for me and whether or not is has a lo-cut of its own.

 

First lo-cut setting on the 552 mixer is 80hz 12/db octave and I rarely go above that. I rarely ever go above 100hz.

 

Never used a low-pass/high-cut on location, or a notch filter for that matter.

[Toy Robot]

I have never been asked to record at lower levels, but have indeed been asked to record 'hotter'. I have even had camera operators request that I send them tone at -16dB and -12dB which I advise against and let them know I'm a pro and this isn't going to be an issue. If post requests it, then fine I'll give them what they want.

 

I understand what is being said about S/N ratio (as we all should), but when we are getting complaints about audio that is 'too low' from people who don't even know how to gain up a recording, what do you expect is going to happen? We know our equipment very well if we are pros. We trust our ears, gain-staging and limiters. I always run tone to camera at -20dB, try and run peaks at -10 roughly under controlled circumstances, and when running to camera I set limiters at -16dB to avoid the camera crapping on my pristine sound.

 

I have never had a complaint except when someone doesn't know what they're doing in post, and those complaints were only in the beginning of my career when I tried to average right around -20dB. Please note that I am not discussing Narrative work.

 

And regarding high/low-pass filters. I don't even touch them unless there's something atrocious going on that I know post isn't going to touch or know how to fix. It's usually LF issues easily solved by a high-pass filter and never more aggressive than say 120Hz. Any more than that and they're either going to have to pay for post or I'm not responsible. Once the sound is gone, I can't put it back, but you can always take out more in post.

[Henchman]

You guys do understand that going fullscale, forces mixers to HAVE to pull llevels down, before eq's etc. as you have left us with no headroom for any processing, right?

Production sound that is recorded with an average level of -20 is more than adequate for the levels we need to reach.

In fact, with the current loudness spec of -24 , I can hit that pretty much dead on, if I maintain an average level of -20 for my dialog.

And the spec states nothing over -2.

So there is absolutely no need to be recording that hot, and having limiters kicking in, if they aren't needed.

You simply make our job harder. Resulting in a higher chance of overloading plugins that don't have an input even trim.

I wish some people, realized what we have to deal with on the stage.

And that it truly is a case where something sounds good, in spite of the sound delivered to us.

I just went through three weeks of hell, getting pisspoor dialog, to sound good enough on a movie.

The original audio a lot of times, was worse than I've had on reality TV shows I've mixed in the past.

No bottom at all. Just midrange. And noise, noise, noise. Even though it was mostly shot on a lot and sets.

No iso's to fall back on. and it had a proper budget too.

And I am simply getting used to receiving audio that ranges from somewhat useable to awful.

And this thread for me, is an example, of people absolutely not understanding what happens after they have recorded their audio. And the issues and problems that are handed off, and dropped in the laps of guys like me. Who have to make it all work. And not only have to make it work.. But are expected it make it sound great.

[atheisticmystic]

 

And this thread for me, is an example, of people absolutely not understanding what happens after they have recorded their audio. And the issues and problems that are handed off, and dropped in the laps of guys like me. Who have to make it all work. And not only have to make it work.. But are expected it make it sound great.

Better change your signature then, babydoll... 

...something like, "Living the dream, but still bitter as a juniper berry dipped in shit and fried in motor oil (and twice as toxic)"

[Philip Perkins]

You guys do understand that going fullscale, forces mixers to HAVE to pull llevels down, before eq's etc. as you have left us with no headroom for any processing, right?

Production sound that is recorded with an average level of -20 is more than adequate for the levels we need to reach.

In fact, with the current loudness spec of -24 , I can hit that pretty much dead on, if I maintain an average level of -20 for my dialog.

And the spec states nothing over -2.

So there is absolutely no need to be recording that hot, and having limiters kicking in, if they aren't needed.

You simply make our job harder. Resulting in a higher chance of overloading plugins that don't have an input even trim.

I wish some people, realized what we have to deal with on the stage.

And that it truly is a case where something sounds good, in spite of the sound delivered to us.

I just went through three weeks of hell, getting pisspoor dialog, to sound good enough on a movie.

The original audio a lot of times, was worse than I've had on reality TV shows I've mixed in the past.

No bottom at all. Just midrange. And noise, noise, noise. Even though it was mostly shot on a lot and sets.

No iso's to fall back on. and it had a proper budget too.

And I am simply getting used to receiving audio that ranges from somewhat useable to awful.

And this thread for me, is an example, of people absolutely not understanding what happens after they have recorded their audio. And the issues and problems that are handed off, and dropped in the laps of guys like me. Who have to make it all work. And not only have to make it work.. But are expected it make it sound great.

I agree with all of the above, but would point out that the point of contact with post for most production soundies is the editor or their assistant--they are the next folks in line after production.  There are often requests from these folks for production sound to "record fat", ie loud, even when it isn't necessary or is even problematic for audio post.  I'm not sure how to fix this except for re-recording mixers to let editors know that their requested "fat recorded" tracks are making problems in your mix.  Producers generally require production soundies to follow edicts from the editors, bucking this often necessitates a protracted discussion.  But worth having, if they'll talk to us.

 

philp

[Max Hirtenstein]

Who ever uses low pass filters on location?

[PTA]

You guys do understand that going fullscale, forces mixers to HAVE to pull llevels down, before eq's etc. as you have left us with no headroom for any processing, right?

Production sound that is recorded with an average level of -20 is more than adequate for the levels we need to reach.

In fact, with the current loudness spec of -24 , I can hit that pretty much dead on, if I maintain an average level of -20 for my dialog.

And the spec states nothing over -2.

So there is absolutely no need to be recording that hot, and having limiters kicking in, if they aren't needed.

You simply make our job harder. Resulting in a higher chance of overloading plugins that don't have an input even trim.

I wish some people, realized what we have to deal with on the stage.

And that it truly is a case where something sounds good, in spite of the sound delivered to us.

I just went through three weeks of hell, getting pisspoor dialog, to sound good enough on a movie.

The original audio a lot of times, was worse than I've had on reality TV shows I've mixed in the past.

No bottom at all. Just midrange. And noise, noise, noise. Even though it was mostly shot on a lot and sets.

No iso's to fall back on. and it had a proper budget too.

And I am simply getting used to receiving audio that ranges from somewhat useable to awful.

And this thread for me, is an example, of people absolutely not understanding what happens after they have recorded their audio. And the issues and problems that are handed off, and dropped in the laps of guys like me. Who have to make it all work. And not only have to make it work.. But are expected it make it sound great.

I agree with what you are saying again, but I think you are blending two different things here. A lot of mixers here are commenting on ENG styles of mixing where there isn't a ton of, if any, processing that goes on in the edit while you are commenting more on post for movies, etc. (Correct me if I am wrong) I think it is important not to assume these are the same thing, especially for post and what they want from a location mix. ENG and reality gigs that I have worked on and that I have had comunication with post have told me about past mixer's levels being too low, not too hot. Mixing dialogue for movies, I assume, is a bit of a different animal and one that I originally was not commenting on, although it has raised an interesting disccusion.

[Max Hirtenstein]

And Henchman I do believe we're on the same side. -20 avg and -10 peaks with perhaps the loudest going closer to my limiters which are set at -2. Furthermore all ISO's are attenuated with neverclip. Palmer's original post specified specified -20 PEAKS.

[Henchman]

Btw, what is neverclip?

[Henchman]

Who ever uses low pass filters on location?

Trust me. I have received stuff where the top and bottom is rolled off.

We are talking nothing below 180 and nothing above 8-10k

[Max Hirtenstein]

Neverclip is a nice feature on zaxcom recorders.

[Max Hirtenstein]

Trust me. I have received stuff where the top and bottom is rolled off.

We are talking nothing below 180 and nothing above 8-10k

I don't doubt it. There is no way I know of to consciously do a lo-pass on most recorders though. It's most likely a roll-off on cheap wireless systems (or buried lavalieres) or something went wrong in the picture edit stage.

[John Steigerwald]

You guys do understand that going fullscale, forces mixers to HAVE to pull llevels down, before eq's etc. as you have left us with no headroom for any processing, right?

 

 

It must be tortuous having to mete out those extra 4 or 5 mouse clicks from the confines of an air conditioned studio.

 

I don't run my mixes right up to the top, but I tend to kind of live around 3/4 scale depending on the material.

 

With respect to gain, I'd generally rather pad, than add.

 

Low cut usually 80 or below, again dependent on material.

 

Don't use low pass filters, though it always seems like wardrobe wants me to use theirs..

[John Blankenship]

Better change your signature then, babydoll... 

...something like, "Living the dream, but still bitter as a juniper berry dipped in shit and fried in motor oil (and twice as toxic)"

 

L.O.T.D.

[Benjamin Gandy]

 

Audio recorded with peaks of -20 is completely fine,mid that's what a client asks for.

Trust me. We do not need level peaking at -5.

 

Why? Do you mean ever at all completely whatsoever or nominally?

 

Anybody thinking they need to record to the top of the signal, needs to go back to school.

 

What's the top bit there for then?

[soundslikejustin]

Why? Do you mean ever at all completely whatsoever or nominally?

What's the top bit there for then?

Headroom for post-processing.

[Benjamin Gandy]

Headroom for post-processing.

Not going to argue with that. 

 

I did go to lower my limiters to +16 and it turns out they've been at +16 all along, not +18 as I thought.

[Joe Finlan]

It still surprises me at the number of misconceptions and misunderstandings concerning recording levels that exist within our industry. While it is true that there are differing operating procedures between film and eng/tv work, as production sound recordists/mixers we should be well aware of the whys and wherefors of what is good practice when recording for a particular end medium and how these affect the post production chain. Having had a background in music recording and audio post production before moving into production sound, I knew what kind of levels I needed to record in the field in order to make the post process work cleanly and effeciently.

 

If I'm doing tv, eng or corporate video I'll record at 48K/16 bit with my average rms levels in the -24 to -20 dbfs range with peaks averaging somewhere in the -12 to -10 dbfs range with the occasional louder peak. Why do I do this? Because when I put my audio post hat on that is exactly where I want the levels to sit in the final mix. Every tv delivery spec I've ever seen stated that the maximum peak level must not exceed -10 dbfs and that average rms levels should not exceed -20 dbfs. What's the benefit of doing this? I, or any other post engineer, doesn't have to waste time in post either increasing or decreasing levels that were recorded too low or too hot for the project. I've never had a video editor tell me my levels were off and I've had feedback from posties complimenting me on my tracks because they didn't have to spend time "fixing them in the mix".

 

If I'm working on a film project then I will adjust the way I work. I record at 48K/24 bit with the average rms dialogue levels in the -34 to -30 dbfs range and the peaks around -24 to -20 dbfs. Why do I do this? Once again it comes back to the final mix. Mix stages are calibrated to -20 dbfs = 85 db spl at the mix position. In the real world the spl of conversation is around 65 db or about -40dbfs if you think of your meters as an spl meter. A dialogue track recorded with an average rms level of -20 dbfs is going to sound unnaturally loud on the mix stage if played back at unity gain. Tracks recorded 10 to 15 db lower tend to "sit" at a volume level that sounds more natural in the film mix. (Not only that but I've now got more headroom for that unexpected change in performance delivery). Once again in post the mixer doesn't have to spend extra time adjusting overall levels up or down prior to mixing, they're already in the ballpark.

 

And to those who think they need to record dialogue as close to the top of the dbfs scale as possible I'll say this: "you've gained nothing and lost something". You've gained nothing because the level you record at does not affect the frequency response of the sound in any way and the noise floor of your mic/ preamp and recorder is already well below the signal you are recording. The days of tape hiss are far behind us. What you've lost is the headroom necessary to handle that unexpected loud exclamation or sound in the background that now causes your limiters to clamp down excessively and become audible or, worse yet, to distort. You're walking a tightrope without a net.

 

Though, as production sound recordists we are at the beginning of the audio production chain, how we do our job has an effect on the entire chain. The better we understand that chain and its' requirements the better we can provide tracks that need the least amount of attention throughout the post process. And that keeps everyone happy, even Henchman. 

[Christian Spaeth]

If you want to make your editors and posties happy then deliver well leveled tracks that average around the same levels as in the final product. That way post will have to do little to no gain riding. Partly agree with Hench in that we don't need to peak close to zero. The less fixing your audio needs in post, the better it will be considered - not primarily by how high the S/N is. Think economically.

[berniebeaudry]

There you go!  A good description of the differences between eng and film location sound.  I only do eng and record 24 bit/48 (Nomad, just because I can)  but generally record at 3/4 meter scale with occasional peaks that get limited by loud sounds.  The editors on the shows I work on have always liked it to be in the ball park of -10 peaks.  My son who is also an editor on a very tight timeline show would prefer the levels to err on the side of hot if the tracks sound good and aren't limited to death.  Maybe part of what is causing the crappy sound Henchman is dealing with are sound people who don't know necessarily know how to gain stage properly, have lavs hidden improperly, or just think that pushed, crunchy sound is good.  Over modulating wireless is another source of that nasty over limited, loud sound.  Great discussion!  I didn't know that film sound used different parameters in post.

[Jay Rose]

If I had a dollar for everytime I heard the words "it didn't sound that bad before" on the dub stage.... 

 

I've even heard it in my livingroom.

 

Wife and I were in a community theater thing. Director set up a locked camera. Wife and I volunteered our house for a cast party, and we'd look at the DVD on my home theater setup.

 

No operator, camera locked down at about 25' from apron, camera mic, AGC on so the gain was usually at max and always sucked up during pauses. You can guess how much hiss, audience noise, and reverb there was...

 

Director said "Gee, it sounded much better when I checked a few seconds on my laptop. It must be your speakers"

 

 

 

She wasn't much of a director either...

[studiomprd]

"  The real signal to noise ratio issue comes from noise on set. "

and usually way before the noise floor of the mic's and pre's.

[Tim M]

If I had a dollar for everytime I heard the words "it didn't sound that bad before" on the dub stage....

I've even heard it in my livingroom.

Wife and I were in a community theater thing. Director set up a locked camera. Wife and I volunteered our house for a cast party, and we'd look at the DVD on my home theater setup.

No operator, camera locked down at about 25' from apron, camera mic, AGC on so the gain was usually at max and always sucked up during pauses. You can guess how much hiss, audience noise, and reverb there was...

Director said "Gee, it sounded much better when I checked a few seconds on my laptop. It must be your speakers"

She wasn't much of a director either...

Omg that is hilarious...your knocking it out of the park today Jay. Does she really not know what you do for a living? How did you respond...silence?

On a "feature length" I was working I actually had to break out your book to show to the editor that lower levels are expected. They almost fired me for it, which at the time I was worried more about reputation than money! They apologized once I did that.

[PTA]

I'm glad I started this post as it has brought up many good points, useful information and knowledge from different backgrounds that I normally don't do work in.

I also primarily do work in ENG and reality where the editors have communicated to me that they would prefer their tracks to be on the louder side rather than the quite side, obviously without the artifacts of limiting and heavy compression.

That being said, the main thing that I've taken from this thread is different post editors prefer different levels for varying types of production and it is important to know YOUR line of work and the levels they require for efficient post work.