I am using cs3 and am confused about setting the audio level on my clips. I know that digital distortion is a big NO NO but when working in Premiere sometimes the audio goes into the red. I am not hearing any digital distortion, so what do the meters represent? I am sure there is a simple answer to this question but just how concerned should I be about clipping my levels in this application?
Levels
- Thread starter sshafer
- Start date
I'm not a CS3 user, but there are some universal basics for digital audio. You'll most likely see meters hit red before absolute clipping... it's a warning that you are dangerously close to 0dB. You shouldn't be clipping, or anywhere close to it.
Depending on the format you're working with, your mix should peak at -10dB (formats like DBeta) or -12dB (formats like DV, HDV). Now, that's typical level for broadcast masters. When it comes to Internet or DVD distribution, there are fewer standards. Especially for Internet, you can push your mix to -6dB peaks or even -3dB.
Depending on the format you're working with, your mix should peak at -10dB (formats like DBeta) or -12dB (formats like DV, HDV). Now, that's typical level for broadcast masters. When it comes to Internet or DVD distribution, there are fewer standards. Especially for Internet, you can push your mix to -6dB peaks or even -3dB.
Richard Crowley
Veteran
Need more details
Need more details
What does "when working" mean?
What are you doing? Recording? Mixing down? Setting levels for final rendering? What is the workflow? Are you making DVDs? Clips for web viewing? Writing back to tape?
What does "audio goes into the red" mean?
Where are you seeing this? Does "goes into" mean hitting the top? or just approaching it?
You probably need to be more explicit about the details around the question for best results.
Need more details
What does "when working" mean?
What are you doing? Recording? Mixing down? Setting levels for final rendering? What is the workflow? Are you making DVDs? Clips for web viewing? Writing back to tape?
What does "audio goes into the red" mean?
Where are you seeing this? Does "goes into" mean hitting the top? or just approaching it?
You probably need to be more explicit about the details around the question for best results.
What I am trying to say is that whenever the audio levels go into the red I am not hearing any noticeable distortion. Usually the situation is when I am mixing music, sound effects and dialogue that will push the levels into the red. In an example project I increased the music 6dB so that it would constantly be peaking and I am not hearing much of a difference in the quality of the sound. Sorry about not giving enough information.
Again, you should not be peaking. It's about safety margin, and not about "it still sounds good."
Digital audio mixing standards can be a bit confusing.
For commercial CDs these days, the levels peak at about -0.1dB but the mix has a brick-wall limiter on it that will keep it from overmodulating.
For broadcast and video production, mixes should peak at -10dB (or -12dB), with the average running anywhere from -18dB to -10dB for dialog, depending on the network's technical requirements. The dynamic range is usually squeezed quite a bit for broadcast as well.
Motion pictures for big screen have dialog that averages around -22dB with loud effects (crashes, explosions) peaking at -12dB or so. Anything mastered to a higher average/peak level will be uncomfortable for a movie theater audience.
DVD has a looser standard, though I know lots of folks who output their -12dB (or -10dB) broadcast standard mixes straight to DVD as is. Others crank it up a little more to about -6dB.
Internet is the Wild, Wild West. Different folks have different marks they hit. Some keep to -12/-10. Others push it to -6/-3.
For you, unless you are mastering for a CD, I'd edit within the standards of your format, which I'll safely assume is -12dB. What are you editing, by the way (TV program, indie film, etc.)?
Richard Crowley
Veteran
Thanks C2V- My background is teaching Premier Pro to high school students. I teach them that we don’t want the meters to peak out in the red zone that we should keep the peaks in the -6 range. But because there is not an audible difference when they do go over 0 they are not buying into why they should be concerned with their audio. I will share your numbers with them and keep telling them to turn it down.
Steve House
Veteran
I think your problem is that kids listen to music so loud and overdirven that they actually think severe clipping is the way music ought to sound! Turn it up till it crackles!
Here's a brief explanation for your class:
It's like a runway at an airport. The runway itself may actually be a mile long, but the airplane will not touch the last 1/4-mile except in an absolute emergency.
In audio, we call it headroom. It was different in the analog days, because analog could pass 0dB quite gracefully. Analog can go up as high as +18dB, or higher, depending on the circuits in a given device. Add to that, mild analog distortion is, in some cases (music recording, especially) considered pleasing and even sought-after. Analog deciBels (dB) are rated as actual voltage measurements, though there is no solid voltage-to-dB value as the dB is relative to a given reference voltage, known as 0dB. That 0dB value can vary depending on the system used, so all other dB values (+ or -) are relative to the given 0dB.
That 0dB value is also considered the "sweet spot" in the signal-to-noise ratio. In other words, analog audio at 0dB has the most ideal saturation of signal with the most headroom and the lowest amount of self-noise in it. Analog circuits add their own noise to the signal, though the better the device the less the self-noise.
Enter the digital world. We are now dealing with what is called "Full Scale" audio. It's now all 1s and 0s instead of a measured voltage. And because it is a string of 1s and 0s, it has a limit. You have sample rate (the number of times per second that a word is recorded, such as 48,000... aka 48k). And you have word length, the number bits (each bit is a 1 or a 0) that are recorded at each sample, such as 16-bit. So for 16-bit/48kHz audio, 16 binumeric values are recorded 48,000 times per second.
This scale has an absolute maximum capacity. When it is full, it cannot hold anymore information. Try to feed it more, and it's going to cop an attitude. That maximum capacity is called 0dBFS (FS being Full Scale). Since there is nothing at all above 0dB, you have to have some headroom. That's why mix levels peak below 0dB. And because digital audio has no inherent self-noise, one can operate below 0dB without the fear of compromising the signal-to-noise ratio.
There is still plenty of analog circuitry in play (cameras have 'em, field mixers are typically analog, etc.). And in order to interface analog and digital, there has to be a standard. So 0dB in analog has been given a sibling value in the digital world. In most high-end production, a 1kHz tone at 0dB analog = -20dB in digital.
That's the quick and dirty version of it, anyway.
And if there is no crackle (digital clipping), then they aren't actually crossing 0dB. Most NLE systems have audio meters that hit red within 2dB of 0 as a warning that you're about to have serious problems.
EDIT: One other note... if any of them are seriously considering this industry after graduation, they need to know why we keep our levels where we do. If they try to get work editing and the employer sees that they're pushing the audio levels, and then they explain "it's fine 'cause it sounds fine," they'll have a hard time continuing in that employment.
Last edited:
Painful truth.