Tone from mixer to camera

[Thomas Berg Petersen]

Dear forum,
I have worked on a number of short films as the sound guy, and I suddenly got nervous as to what to set the tone from the mixer if you record sound directly into the camera.

Over the years, I have done multiple things. First I have read that you should set it to -18 or -20DB on the camera. Sometimes I set it a little higher to -16 or -18 DB. Especially if I only boom and gang the mixer to output two mono signals on channel 1 and 2.

But after one gig a long time ago, I heard from the sound designer, that I had recorded the dialog a little too low.

So here the other day when I got an other sound gig, I became very particular and perhaps overly cautious. I had a lot of time to test, and I was using a SQN 5S mixer only with a boom (Senn MKH60) to an IMX camera.

I ganged my mixer's mono mic to two outbound mono channels and sat the camera levels to -20DB and -18 DB on the camera from the mixer's tone, then I tested. I could shout directly into the mic and even gain max on the mixer, and still the sound levels on the camera only got up to -20DB and -18DB. In other words. It was as if the max. level was what I had set it according to the tone from the mixer.

I tried everything possible under the sun to see if there was an easy fix, but came out empty handed. So I decided after having been editing a little and knowing that dialog should be around -12DB, I changed my strategy and sat the mixer's tone to -10 DB and -12DB. I recorded stuff on the camera and did some sound testing, and it sounded all right.

Can any professional sound people please explain what happened and what you are supposed to set the camera's levels at from the mixer's tone.

Thanks a bunch
Thomas

[Chris Swanberg]

The general consensus is to set the tone at -20db, which is zero. This gives you plenty of headroom, and with a good mixer with reliable limiters you can pretty much rest assured you will not have any clipping.

That said, some sound, that is pretty well known ahead of time not to have any higher peaks, might benefit from the increased S/N of a higher setting on the camera (like -12db.)

In that lurks some danger.

If I am recording to a camera with two independent channels, I will route the signal to both, and set one at -20 and one at -12. Post production I can pick the better of the two. If -12 clips, I revert to -20.

Overall I also think the answer will depend on your equipment and advance knowledge of the sound you will be capturing. (eg. Recently in a shot where the captured sound was to be a whisper, I went higher that -12 with some assurance...with excellent results... plus I could reshoot. In a one time take situation, even more caution is advisable.)

[Steve House]

The reason for the alignment tone is so you can look at the mixer's meters and know what the camera's or recorder's levels are doing. As a mixer you need to understand exactly how the whole system is behaving so you can control it and the standard alignment levels help you correlate the various components. The SMPTE standards call for a +4dBu sine wave signal to read OVU on a traditional analog meter and -20dBFS on a digital peak-reading meter while the EBU standards call for 0VU to equal 0dBu and align to -18dBFS. Both are "correct" because they only serve to align the two meters and the gain structures to each other and do NOT represent the recording levels for actual complex signals. If you play a 0VU sine wave and set the levels to read -20 then switch to a 0VU pink noise signal, you'll see the meters bouncing around -12 to -10 dBFS. The mixer will read 0VU and the camera's peak meters will read around -10. It used to be that 0VU indicated an average signal output level of +4dBu. Given the realities of the digital world, many manufacturers such as Sound Devices are calibrating their meters to indicate 0VU at an output level of 0dBu. If you were using an SD mixer, setting tone to -20 in the camera and then recording with the mixer's meters showing 0 to +4 VU on voice signal would result in peaks in the -12 to -8 range in the camera. You would be recording with an average level of 0VU with 20 db of headroom. One of the nice things about the SD mixers is they offer a mixed meter mode that shows both average signal in VU units and peak signal at the same time, similar to Dorrough's metering model, and they let you really see what is going on with your levels. There are some peaks that occur too fast for even a peak reading meter to register and yet can cause audible clipping effects. By aligning to -20 or -18 in the camera, then recording with average levels at 0 to +4 on the mixer and peaks at +8 to +10, limiters set to kick in at +18, you're recording at levels that should maximize your signal-to-noise ratio with enough safety to avoid peak 'overs'. If you know the material has less than 20 dB dynamic range, you can safely record with hotter average levels, if it has wider dynamic range, you need to record with less average levels but you can make those decisions working with the mixer with confidence that you know what is happening in the camera.

[Thomas Berg Petersen]

Guys,
Thank you very much for your input.
I guess you can say that I have done it right in both instances but perhaps this one instance where the sound designer complained about low levels, the book mic must have been too far away from the dialogue.

I suddenly forgot what the vu meter says when you switch on the 1KHz tone. But I know that I will aim to adjust the gain on the mixer so that dialogue and other sounds won't ever get close to vu=0. But my point is that if indeed this happened then the camera will monitor this at the levels you have set when you aligned the tone in the first place. In other words, if you align the tone to -18dB, then no sound from the mixer can exceed -18dB (perhaps with the exception of very sudden and sharp noise). But I never really understood why the camera needed so much headroom (from -18db or -20dB to 0dB). Since my understanding is that the louder and cleaner the dialogue you get the easier it is to get rid of excess noise in post. As supposed to recording low dialogue and then if you bring the dialogue up to e.g. -12dB in post then the general sound floor will increase too and it will sound terrible.

I am still not very convinced that -18dB or -20dB is the way to go always. The commercial I just did, the were also talking very low and it was in a sound proof studio. The only noise I could hear was from the Pro35 adapter's oscillating glass, which by the way in the tight shots became unbearable and pretty much unusable if you ask me. Any tricks?

For future sound jobs I will continue to align the tone to from -18dB to -12dB. But I do prefer to record to a hard drive like a sound device 744T.

I pulled this off a website about the SQN mixer:
When interfacing to an analogue recorder with its typical slow overload characteristic, what is required is to place the mixer’s Nominal Peak Level at the correct point which will ensure that the recorder will not be overloaded by a limited transient output from the mixer. Most of the ENG recorders use VU meters, with which the Peak Level is off scale. The reference point we use, therefore, is the Nominal Line Level - the level to which the mixer’s Line-Up tone is set. Experience has shown that most of the usual ENG recorders are calibrated so that their Nominal Line Level or 0VU is only 6dB below the level at which distortion is beginning to increase. Accordingly, it has become the practice to adjust the recorder’s gain control to place the mixer’s Line-Up tone at -2dB on the recorder’s VU meter. Some recordists prefer to go further and leave themselves a little more headroom on the recorder by placing the line up tone at -4dB on the recorder’s meter.

Interfacing to a digital recorder with its much more sudden and unforgiving overload characteristic requires that the mixer's Nominal Peak Level be placed below the recorder's peak level. This is because the limiting, as carried out on the mixer assumes that the following recorder is tolerant of some degree of overload for periods up to 1ms, hence the emphasis on the measurement using a Peak Programme Meter in the paragraph above. Fortunately, digital recorders, because of their clipping characteristics, are almost always fitted with fast peak meters, reading on a sample by sample basis. It is easy to set the matching between the mixer and the recorder experimentally using transients such as sharp handclaps with the gain turned up and the limiters turned on. Typically, the Line-Up tone should be placed 12dB below the recorder peak level.

Thanks again for your expertise and quick response.

PS. Usually I am a producer, but I guess I am a little nerdy and likes tech aspects especially sound. And I really enjoy myself when on set and the only thing you have to focus on is getting good and clean dialogue as opposed to being a producer and always being two days ahead while putting out fires left and right.

Best regards,
Thomas

[Thomas Berg Petersen]

I forgot to mention a thing. I almost always switch on my limiters on the mixer. Though there are exceptions.

What I would love is to get out in the field with a very seasoned pro and pick his brain while he is working to be even better to get better sound.

Thomas

[Chris Swanberg]

Remember that analog and digital are quite different animals in the audio world. Exceed 0db in the digital world and there's no going back. Clipping equals audio disaster.

The equipment you are using is important. In the Sound Design world, a limiter cuts in well and maxes at 18db. Others may give different results. Therefor, using a SD mixer with 20 db headroom you are safe...meaning when you do not know the levels you might encounter. My point as that when you have some idea of the levels you can watch and set accordingly and improve your S/N. I also offer that in MY world, I am all things, so I cannot necessarily be riding a mixer and therefor I have to adjust and set in advance. Unity gain and headroom are my bywords.

That said, Steve, and others on here are the true pros, and I'd listen to them - and ask them to decipher when their answers do notinitially make sense to start with. My advice is "street" advice from an indie who has the scars to show he learned the hard way.

ps. Sound is 70% of the audience experience. Do it well.

pps. 86.7% of all statistics are made up on the spot.

[Steve House]

Originally Posted by Thomas Berg Petersen

Guys,
Thank you very much for your input.
I guess you can say that I have done it right in both instances but perhaps this one instance where the sound designer complained about low levels, the book mic must have been too far away from the dialogue.

I suddenly forgot what the vu meter says when you switch on the 1KHz tone. But I know that I will aim to adjust the gain on the mixer so that dialogue and other sounds won't ever get close to vu=0. ...

No no NO! You WANT average levels of dialog to be around 0 VU to +4dBVU or thereabouts on the mixer's meters. Remember 0VU analog = -20dBFS digital so clipping in the camera, the must-never-exceed limit, is all the way up at around +20dBu on the mixer's meters.

[Nick Flowers]

Where to set levels of tone can be confusing, especially with the varying types of meter on the mixer and on the recorder. In the UK generally speaking we don't use VU meters on UK manufactured mixers, but the BBC version of the PPM (Peak Progamme Meter). This differs from the VU meter in having a fast attack and slow decay time and zero level is shown as '4'. Note that zero level is NOT 100%, it is 8dB below and as the numbers on this PPM show an increment of 4dB 100% is '6' on this sort of meter.

When analogue ENG and Camcorders came in we had to line up to VUs, as these recorders had VU meters on them and the practice was to line zero level up to -4 on the VU. Then digital recorders came in and the practice came to line up zero level on the mixer's meter to -20 on the recorder: not ideal as -18 was what was advised by the general consensus, but most segment meters on the camcorders showed -20 and not -18. I see that HD cameras are now showing -18 on their audio meters.

I have always found VUs very hard to get used to, as speech and music behave so differently on them. I guess it's what you have been brought up on. Nordic PPMs differ again from BBC PPMs, so you guys in Scandanavia have your own standards.

[Lee Sharp]

Hi guys

Ok we all know line up is 0dbu = 0vu = -18dbfs for EBU and 0vu = +4dbu = -20dbfs for SMPTE

and that the guide is to have peaks hitting around -10dbfs....

Question is, is that just for broadcast delivery? As sound recordists can we push the peaks as far as we can in to the 0dbfs during recording (providing we know the range) thusly making a better S:N ratio?

Also what is the bare minmum acceptable level for dialogue? Ive been told -30dbfs during recording?

But if a pro can clear this up for me my main query is the levels issue, is it OUR job to keep the peaks at -10dbfs, or post production, can we go beyond this aslong as it doent clip I mean after all these mixers are desgned to go right to 0dbfs with a limiter at around 18dbu.

Also if we can have levels hotter and line up at - 12dbfs rather than -18dbfs does that not then mean on the mixer we cant go very high at all as 8dbu would push the levels on the recording deviceto clipping? It gives us less mixe gain control as we've boosted it at the camera ends, I dnt really grasp the point in that? If we want a strnger signal would we not just line up at
-18dbfs and up the gains on the mixer to acheive a hotter signal? Either way the recorder would recieve the same signal htehr the recorder gain was increased or the mixers...

It so damn confusing!