Audio levels...oh boy!

R

ronin99

Active member
Somewhat a continuation of my most recent posts.

The hits just keep coming; let's talk levels for a moment. I'm going to be talking to production tonight and want to make sure I have my facts straight and I'm not talking out of my behind. I've gone through the audio for the first couple days and the levels are pretty low around -42 to -30db and occasionally peaking around -24db. This is occurring for both the boom and lavs. Shouldn't the levels be around -12db?

I'm not just talking about Wide shots where the boom can't be near the actors either, the same thing is happening in Mediums and O/S, and it's not framed in some artistic way, the players heads are near the top of the frame, so the boom mic if placed properly should be giving a pretty strong signal and I imagine the lavs should be as well.

The boom mic, sounds okay when boosting the signal, but clearly with the boost, noise is being added. The lav on the other hand sounds pretty bad with a boosted signal.
 
Well is there a line up tone and is it referenced to the same scaling you are looking at the audio on?

There are numerous different scales and they can be different for analogue and digital sources so it is always important for a reference tone to be recorded.

This is normally a 1khz sine wave tone that has a voltage of .775 volts and is then relevant to the scalings that are pictured below!

Peak levels above the reference tone are usually +8db for broadcast and that then corresponds to the peak levels also shown on the diagram.



So for 1khz sine wave line up at .775v the meter scalings from left to right correspond to 0dBu, 4PPM, -4VU and -18dBfs.

Just to confuse matters the USA and other regions such as Japan will read 0dBu, 4PPM, -4VU and -20dbfs.

So the normal +8db peak levels for broadcast will be 1.946 volts = +8dbu, 6PPM, +4VU and -10dBfs (or -12dBfs for USA)

So back to your levels do you have a reference tone that you can line up the material to? I assuming you are using dBfs but if you have no reference to the original levels it would be that they have been transferred low or that the scaling on your system is not calibrated the same as the recorder.

Impedance termination can also affect levels if you are coming into your system via analogue inputs so I hope you can understand that just quoting a number and saying the audio is low is not the true picture as you are just looking at a meter that has no reference to the original audio.

Yes you are right that -12dBfs is the normal peak level as described above but if the audio has been transferred low then that level will not be the peak and it may be that all the audio is 10db down.

If the original audio has been recorded digitally and is still low then adding digital gain should not increase the noise as the dynamic range of digital is around 100db and noise floor should not be an issue but once again unless you have a common reference to the original other factors may explain why your audio is low.

Hope that all makes sense but if you I am sure me and others will be able to answer any further questions you may have.
 
Last edited:
The big problem you're going to run into with low audio levels is the noise-floor.

Professional recorders like the Sound Devices models have a noise-floor that exceeds -100 dB, so boosting your recording by 20-30 dB won't be a problem.

But, consumer level recorders like the Zoom H4n has a maximum noise-floor of -66 dB RMS, so boosting any signal more than 10 dB will introduce a very obvious noise problem.

Personally, I consider -60 dB RMS to the noise-floor threshold for high quality audio.

...BTW, what audio gear are you using to record with?
 
It is common for people with DAW's such as pro tools to record low as once you start adding plug ins it can very quickly overload the mix buses.

As said get him to record you a 0db reference level at -20dBfs and then advise that he should peak to around -12dBfs if he is using digital meters, he will still have plenty of headroom for peaks but I suspect he is following music recording practices that allow for the mentioned addition of plug ins!
 
GaryNattrass said:
It is common for people with DAW's such as pro tools to record low as once you start adding plug ins it can very quickly overload the mix buses.
Click to expand...
But we are talking about TRACKING individual mics. No processing. No mixing. Does the recordist understand the requirements of his job?

As said get him to record you a 0db reference level at -20dBfs and then advise that he should peak to around -12dBfs if he is using digital meters, he will still have plenty of headroom for peaks but I suspect he is following music recording practices that allow for the mentioned addition of plug ins!
Click to expand...
Maybe that will give him a clue what kind of levels you need.

ARE you getting proper slate? Sync marker? Reference tones?
 
Yes Richard but if I am tracking on my pro tools rig for music I tend to keep the levels low, it may be that the Operator does not know about broadcast or film levels and is just tracking as he normally would for music.

As I said a reference tone will assist and give indication of what levels are being recorded.

On a personal level I always check my workflow for any project I do from mic all the way thru to camera or recording device any production would be sensible to do this before they start and as Richard also says defines how things are going to be idented and marked for sync reference.

You can even go to the extent of putting a tone generator source on the XLR that plugs into the mic at -50db and check that the levels and any limiters are doing what they should be from source all the way thru any mixing stages to the final recording device or camera!

Even the behringer cable tester that costs less than £30 here has a tone source that can output a 1khz or 400hz tone at +4 pro level, -10 prosumer or -50 mic levels and can be really useful for defining the signal workflow to recorder and assist setting gain structure through the signal chain.
 
Last edited:
GaryNattrass said:
On a personal level I always check my workflow for any project I do from mic all the way thru to camera or recording device any production would be sensible to do this before they start and as Richard also says defines how things are going to be idented and marked for sync reference.
Click to expand...

What do you use for your 'sound source' that is picked up by the mic? Speakers, and a sound pressure gauge?
 
j1clark@ucsd.edu said:
What do you use for your 'sound source' that is picked up by the mic? Speakers, and a sound pressure gauge?
Click to expand...

No if you read my post again what I do is plug a tone generator into the XLR mic cable at MIC LEVEL of -50dB therefore replacing the MIC with a constant level tone source at MIC level!

So rather than having a mic as the feed for the XLR cable I have a tone generator feeding at mic level!

Here is a quick video I did and hope it makes sense what I am doing: http://www.youtube.com/watch?v=pJWKLbVZVU4&feature=youtu.be

note as mentioned above on another post the camera line up for a USA model would be -20 not -18 so here is another vid with a different mixer showing -20 and -12 peak dBfs from a mixer with BBC PPM meters: http://www.youtube.com/watch?v=s7jG_djmiX0&feature=youtu.be
 
Last edited:
The short of it is YES his levels are way low, and it will probably be a problem down the road.

RE: the music side thread, I tend to record everything as hot as I can get away with. It's a lot less painful to lower a overly hot (but not distorting) track than to bust an overly low one. In a "studio" like situation that can be a problem if you don't have a monitor mix but...

But not sure that is the music norm. Im out of FX and post and I record and then mix. So I will get as many bits on the take as I can and worry about balancing them when I'm mixing. Not that anyone else needs to do it that way, just my two cents.
 
ARE you getting proper slate? Sync marker? Reference tones?
Click to expand...

If you mean is he giving a slate at the start of each day, than NO. He's also not giving tone.

It is clear that he is not experienced in film workflow and probably comes more form a music/live event background.

But we are talking about TRACKING individual mics. No processing. No mixing. Does the recordist understand the requirements of his job?
Click to expand...

Right, I had to reiterate this when he started talking about using plugins to bring up the levels.

As said get him to record you a 0db reference level at -20dBfs and then advise that he should peak to around -12dBfs
Click to expand...

I'll give it a shot.

When I asked the mixer about the -30db 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.
 
Well if you track too HOT in something like pro tools and then add a plug in such as a compressor or EQ then you may overload the mix buses.

There should be no reason to be recording up to the MAX in the digital domain and I seem to recall in my distant digital theory that under modulating can make some systems quieter but I do appreciate that getting a good signal to noise ratio is paramount.

As for too low levels then you will possibly get signal to noise problems from the analogue conversion part of the equation, Indeed on certain classical recordings they used to have to add dither to create a noise floor to mask any fades that could be audible from analogue consoles.

If the operator sticks to the broadcast normal of -20dBfs for 1khz .775v line up tone and peaks to -12dBfs that will still give 12db of head room but will give a good signal to noise ratio for post production.
 
I probably should have qualified. If you record anything above -3 dBfs you can get distortion on high frequencies, when I said "hot" I meant peaking between -6 and -3. Hotter than normal for production work.

There are lots of ways to deal with levels in PT so you should be able to fix any buss overloading. With the latest version you could even go to 32 bit float so overloading becomes practically impossible.

But I'm with you on going crazy with over hot recordings.

BTW dither is used on digital recordings to fuzz up the very low levels on fades and such to smooth out the bits.

I don't think it was used in analog, not much point. Something similar was used when doing optical tracks for film because the optical reader would get wanky if there was "silence", so editors/ mixers would always fill any "silence" with a "silence loop". The "classic" is "Yellow Sky" which is just a recording of an empty sound stage (named I believe because it was recorded for use on that film). You can still find it in some libraries.


GaryNattrass said:
Well if you track too HOT in something like pro tools and then add a plug in such as a compressor or EQ then you may overload the mix buses.

There should be no reason to be recording up to the MAX in the digital domain and I seem to recall in my distant digital theory that under modulating can make some systems quieter but I do appreciate that getting a good signal to noise ratio is paramount.

As for too low levels then you will possibly get signal to noise problems from the analogue conversion part of the equation, Indeed on certain classical recordings they used to have to add dither to create a noise floor to mask any fades that could be audible from analogue consoles.

If the operator sticks to the broadcast normal of -20dBfs for 1khz .775v line up tone and peaks to -12dBfs that will still give 12db of head room but will give a good signal to noise ratio for post production.
Click to expand...
 
ronin99 said:
...

I'll give it a shot.

When I asked the mixer about the -30db 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.
Click to expand...

Welcome to my world ;~)

BTW I believe the poster meant take slating, as in before each take you get a someone stating the scent and take and they shoot the clapper and the clap.

To be fair to the mixer, to slate or not to is generally not their decision. They can ask but it's the directors call. So if there are no slates it's not just the mixer who is not following smart film practices. It also probably means they don't have a script supervisor tracking it either. If all that is the case then strap in tight because there is a strong chance of a train wreck in your future. All you can do is push for smarter work flows and prepare the best you can. It means that probably you are pretty much stuck with what the picture editor uses because it will be a needle in the haystack finding usable alternate takes.
 
You must log in or register to reply here.
Back
Top