FS7: Audi dial delay
- Thread starter truelife
- Start date
morgan_moore
Major Contributor
And there u was thinking my fingers are too fat
epicdream
Well-known member
morgan_moore
Major Contributor
on a side note does anyone know how to get the little door off in a non destructive manner?
Run&Gun
Veteran
Why are people trying to adjust audio levels on the camera during shooting? I've been doing this for almost two decades and I don't try to ride audio levels manually on the camera while shooting. My opinion, you either have an audio guy who does it with a mixer OR if you're solo, the cam is in auto for audio(wireless and nats). And a lot of my shooting is solo in one of the WORST audio environments on the planet: racetracks.
That would be like posing a similar question on an audio forum and someone answering, "you're nuts, why do you bother with manual iris and focus? Stick it all on Auto..."
We ride the pots because audio has this awful habit of changing amplitude - so we adjust the pre-amp gain to reflect that.
Racetracks might be challenging acoustically - but one could argue that most listeners don't have high expectations of audio fidelity at those sound pressure levels - it's just a full-on wall of noise.
But if you want to record a voice, for example, any AGC circuit is going to pump up and down, lift the residual noise floor up and then slam it back down when the signal (voice) hits in. Up and down. Up and down. The result would be horrible noisy pumping audio.
A bit like auto iris and focus...
Alan Dater
Member
If the audio control door is hard to open, as mine was, take a nail file or fine sand paper and ever so slightly file down the plastic catch on the door. It will now be easy to open and stays closed.
basspig
Veteran
If the audio is really as good as 24 bits, volume should be set at the level where optimum gain staging is obtained (similar to native ISO as a video analogy) and the sensitivity set via attenuator settibgs (assuming the FS7 works like the EX3). Normalize in post as needed. I have set my auio levels once and not touched them since.
Run&Gun
Veteran
Apparently you missed my point, which was that the audio controls/dials on board cameras aren't intended to be used to ride the levels and adjust in real-time like on a mixer.
Run&Gun
Veteran
If I'm solo with no audio guy and we have to get sound(stick mic/wireless) like in the locker room, on the field, garage, etc. audio is on auto, because I'm in uncontrolled acoustical environments with subjects they may go from yelling to barely audible whispering and anywhere in-between at anytime. NOW, if I'm in a more controlled situation(but still solo) like when shooting a stand-up or cutting trax, I'll flip over to manual and 99.999999% of the time, I never have to adjust the levels.
I didn't miss your point - I think your view is based on having a separate sound recordist - ie, calibrate the audio mixer to the camera and then leave the pots alone.
If you're solo shooting, you need to be able to ride the pots. It really isn't that radical an idea - I've got a 302 mixer, for example, but I'm not going to carry it around, slung over my shoulder, just to add another audio pre-amp to the chain.
Some of us cheeky upstarts are even turning the lens focus ring ourselves - frightfully vulgar, I know, when there are perfectly good ACs out there
And don't get me started on carrying my own tripod...
Ben.
basspig
Veteran
This is a recording method that comes from the old days of analog tape with 40dB s/n ratio, where you had to gain ride to stay above the hiss, but keep the distortion under control too.
FS7 has 24-bit audio. That suggests it should have a dynamic range of up to 144dB (in practice, more like 136dB, as is the case with my MOTU 896 Mk III). The way I set levels is I find the loudest sound I can possibly expect to record. Set the level so that's just below clipping. Then leave it be. The audio I record of a bumble bee buzzing might be 78dB below that set maximum level, but if I want it to be louder in the final production, I'll add 50 or 60dB of gain/normalization to the audio. In 32-bit float, the quantizing errors are surprisingly infinitessimal.
Another thing that is necessary is to find out the sweet spot where you get maximum dynamic range and minimum noise. For the EX3, the audio level is '5' on the dial. Don't know about the FS7 as I don't have one to test yet. But the idea is that level aligns the preamp gain with the DAC. Too high, and your noise floor increases. Too low, and your preamp will clip before the DAC is all bits on or at 0dBfs. Find that sweet spot where the gain staging is aligned and then use the attenuator (FS7 has an attenuator menu, right?) to adjust the levels, controlling how much signal comes into the front end. In a noisy environment, -24dB for the atten, if using a Rode NT4, gives you plenty of headroom for shouting. Under no circumstances should you use AGC in a locker room. The live acoustics will be made even more 'live' and the echos will be accentuated. Fixed gain and directional mics to minimize room noise/reflections.
All sound advice (excuse the pun.)
Where I find myself having to occasionally ride the pots is with inexperienced interviewees. I allow what I think is a decent amount of headroom, based on a "what did you have for breakfast?" soundcheck and, sometimes, they still manage to get over excited and push the signal into clipping. I'm also often recording in rooms with residual noise - so I try to push the signal as far as I can for optimum S/N.
So, while those tiny awkward controls weren't designed to be easily tweaked, in practice they often do need an occasional fiddle - and the latency of the metre response is irritating.
Ben.
basspig
Veteran
If you're having to dive for the gain control during an interview, your levels are still 20dB too hot. Remember, there's no tape hiss to worry about. It's better to record low and bring it up in post (you have 24-bit audio here, so lots of room for normalizing levels). As long as the mic preamps are not noisy, the only noise floor you'll be dealing with is that of your mics. And the lower you set the attenuator gain, the lower that noise floor, while you gain more ceiling for loud noises, like the unexpected gunshot from a police shooting that just occured across the street from where you are shooting.
allenfacemire
Well-known member
That's fair enough but even us camera folk sometimes have to ride the Audio Auto train. I will maintain that auto video functions are more troublesome that of auto audio. Auto video while sometimes decent but really aren't as smart as the auto audio functions. We never ride the audio on the camera mic and generally it's fine.
As for me, I use a soundman 90% of the time. When I can't I depend on either auto with limiter on or no auto but still use the limiter. In the case of Lectro's, you are generally in a talking head situation so I set the tranny at 12:00 and the receiver at 10:00, preset the level and let the latitude of the system. I have to say that when I have used the Auto audio it's been fine. No searching. Not ideal but then we sometimes have to do what we have to do. It would be nice to have at least one Ch. with an addressable pot like on an F800 and I guess if you wanted to put a Beachteck type mixer or something with an XLR I/O in the equation, that might solve part of the issue. This guy looks kind of cool. http://www.bhphotovideo.com/c/product/292990-REG/Sound_Devices_MP_1_MP_1_Single_Channel.html.
basspig
Veteran
Folks don't realize that digital audio has MUCH more dynamic range than video. Imagine 30 f-stops of range and a noise floor of -90dB instead of -57db and 14 f-stops. That's audio in the 16-bit world. YOU DO NOT NEED TO RIDE THE GAIN!
If you want it to be louder, do the destructive work in post. Effects are inexpensive in terms of CPU demands. Use them. Normalize, or add gain riding AGC later.
Remember, even the best large diaphragm condenser mic (Neumann U87A) has a noise floor of -82dB. The mic pre on a good camera like the EX1 is -88dB. Gain stage the mic pre so that the noise floor of the mic is just above the noise floor of the preamp and that the clipping level of the preamp is just a tad higher than digital clipping and leave it be. If you need to handle exceptional loud sound (eg. shooting range, fireworks, cannons), use the attenuator menu to lower the input signal. The noise floor will be the same, whether you boost the gain in the camera or in post. Better to preserve all the audio info and then screw it up in post to your liking, rather than lose the information and not be able to recover it.
I made a recording a few years back which sounded like silence when played at normal levels. It was a recording of a distant owl hooting in the June night. When I normalized it +60dB, I could hear the owl clearly, but I could also hear pollen falling onto the leaves and echoing throughout the forest. Yes, even something as quiet as pollen particles make enough noise to echo through the woods. This extreme example demonstrates that you don't need to have your record levels hovering at 0dB. This is not a broadcast transmitter. This is digital audio. Use your "30 f-stops" of audio and retain ALL of the highlights and the shadows.
One day, video will be as good as audio, and you won't need to control the aperture except to control depth of field.
If you want it to be louder, do the destructive work in post. Effects are inexpensive in terms of CPU demands. Use them. Normalize, or add gain riding AGC later.
Remember, even the best large diaphragm condenser mic (Neumann U87A) has a noise floor of -82dB. The mic pre on a good camera like the EX1 is -88dB. Gain stage the mic pre so that the noise floor of the mic is just above the noise floor of the preamp and that the clipping level of the preamp is just a tad higher than digital clipping and leave it be. If you need to handle exceptional loud sound (eg. shooting range, fireworks, cannons), use the attenuator menu to lower the input signal. The noise floor will be the same, whether you boost the gain in the camera or in post. Better to preserve all the audio info and then screw it up in post to your liking, rather than lose the information and not be able to recover it.
I made a recording a few years back which sounded like silence when played at normal levels. It was a recording of a distant owl hooting in the June night. When I normalized it +60dB, I could hear the owl clearly, but I could also hear pollen falling onto the leaves and echoing throughout the forest. Yes, even something as quiet as pollen particles make enough noise to echo through the woods. This extreme example demonstrates that you don't need to have your record levels hovering at 0dB. This is not a broadcast transmitter. This is digital audio. Use your "30 f-stops" of audio and retain ALL of the highlights and the shadows.
One day, video will be as good as audio, and you won't need to control the aperture except to control depth of field.