Preparing for DVD

[Todd Beabout]

Don’t know how your workflow is, but when I get audio from the audio post facility we use, I always pull it down 12dB (based on our in-house mastering standards) for output to tape. Digital tape needs a little head-room, but for DVD you can master right up to 0dB. So, in my case I export the audio at 0dB (the way I got it) for DVD and that works. If you don’t use an audio facility, I suppose you could bring the master audio level up in your FCP sequence until the audio is PEAKING at 0dB and then export via Compressor. Be sure that you listen to the whole sequence first, and make sure the audio doesn’t clip.

BTW, I used to leave the audio down when exporting until I realized my error, but I did not receive any complaints regarding the audio levels on those DVDs.

-Todd Beabout
Vazda Studios

[Alexander Kallas]

[Todd Beabout] “BTW, I used to leave the audio down when exporting until I realized my error, but I did not receive any complaints regarding the audio levels on those DVDs.”

So Todd,
Why change to 0dB?
Surely running so close to clipping is unecessary.
Cheers
Alexander

[Todd Beabout]

Actually, in my case I don’t have to worry about clipping. I just leave the audio exactly as it was prepared by the audio post facility. Part of their job includes ensuring just that. The other reason is that no one wants their DVD to not sound as “hot” as the next one.

Just my $.02

-Todd Beabout
Vazda Studios

[Alexander Kallas]

John and Todd
I would caution against 0dB audio, it’s just too close for comfort.
For DVD AC3 sound is the standard.
In A-Pac you must set the “Dialog Normalization” to -31dBFS (from the default -27dBFS) to avoid this volume drop on your DVD.
Cheers
Alexander

[Samuel Frazier]

I’ve been using Sony Vegas for a while and there’s a well known tip in its DVD authoring program to prep your audio for DVD to -31 something or other. Perhaps this is the same thing you’re talking about.
At any rate, 0 is cutting it a little close. But, I can’t think of a reason to put it too far off 0. The safer you get from 0 the more audio fidelity you lose. And, unless I’m wrong, there are no set rules or standards for DVDs, so close to 0 is advisable. Think someone once suggested -.3 unless that’s just in my head.

[Alexander Kallas]

This is an Apple forum, I don’t know Vegas.
What I’m saying is, setting -31 in APack keeps the audio to the level you have in FCP, so the latter must not be too hot.
Cheers
Alexander

[Samuel Frazier]

So, do you not want me to mention my old editor? That would be kind of strange, especially as I was only pointing out what is a universal trick.

[Alexander Kallas]

[mus man] “So, do you not want me to mention my old editor? That would be kind of strange, especially as I was only pointing out what is a universal trick.”

??? Too cryptic for me.
Cheers
Alexander

[Gary Hughes]

What you’re dealing with here is a question of standards, not quality.

dBFS stands for dB full scale. It is the digital equivelant of a dB measurement. It simply means all bits on, or all bits off, or somewhere inbetween. All digital audio devices and media support up to 0dBFS and no more, in it’s final stage (more on this later). Therefore, in truth, headroom is not actually required by any digital media. In other words, it is possible to record audio at 0dB on DVD, miniDV, DV-Cam, iPod, etc…, without clipping (distortion). Before I go on, let me say that I don’t advise it in all cases because it isn’t standard. Industry standards say to drop the level of your audio if you are outputing to tape, regardless of format. You should abide by the standards, because that is what keeps us compatible with each other. PLUS, it is what allows a straight dub of a digital tape to an analog tape without worry about overdriving the audio levels (that’s the true reason the standard was invented to begin with). It keeps you backward compatible.

In the case of DVD, it follows the same standards as CD. 0dB is actually the standard for PEAK audio levels on CD or DVD.

A-Pak, Vegas, and other software that allow or suggest normalization during the digital compression stage, are assuming that you do not have your audio properly set to begin with. This is ok if you aren’t an audio person, or don’t understand all of this. They are “dummy-proofing” the audio settings for the users who do not know better. (Please don’t take that personal!) So why so far from 0dB (like -31dBFS)? My guess is that they are using the “RMS” value of audio, which is completely different from “Peak” values, and they really can’t be discussed in the same argument. It’s like saying “my apple is redder than your orange”, “no, my orange is oranger than your apple”. I will say this about RMS, be cautious when using an RMS scale to set the levels of your audio because it is possible for a negative RMS value to end up with a level that tries to exeed 0dB, therefore causing clipping (distortion). That’s why you like -31 better than -27. (If -27 allowed a +1dB level, then it would clip, but at -31 on the same file, it would only be at -3dB, which is safe from clipping.)

Ok, above I said “All digital audio devices and media support up to 0dBFS and no more…” Let me prove this point. We all know that digital signal is made up of ones and zeros and nothing more. 0dB is simply all ones and no zeros. For a tape to require headroom, it would have to be analog. The tape itself can’t know whether it is holding ones or zeros. It doesn’t say “oops, there’s too many ones here”. All digital tape formats can hold all ones without zeros, it’s just that simple. Digital tape can’t clip or distort. For there to be distortion in audio on digital media, the distortion had to be there before the file was copied to the media. The tape, or other media, is simply a container, like a shelf. It holds the bits, like books. No audio is actually being recorded to the tape, just bits of data, and data can’t clip.

So if all this is true, why do we need to be so cautious? Because the preparation of the file is what is so destructive, not the media that it’s on, or the file that contains it. Software that allows manipulation of audio levels will allow you to set the level so that the peak will try to exceed 0dB. FCP, unfortunately doesn’t allow precision metering, so you have to set the level and watch the entire program to be sure it doesn’t clip (not fun). Sound Track Pro will scan the file for the peak level and report it back to you (much more fun), allowing you to know how far from 0dB you really are, so you can set the level accordingly. There is a big “gotcha” with STP though, like FCP, it has a “master volume control”. The setting of the master volume control isn’t taken into consideration when it looks for a the peak value of a file. If you have your master volume control set to anything other than 0dB, you still don’t know how loud the actual level will be, so just be sure it’s at 0dB unless you are in a multiple track project and need to set it to something different.

Here’s my workflow for DVD. I set the levels in FCP to tape standards. Then I export my sequence to an aiff file. Then open the aiff in STP or similar audio sweetening software. Then I use the normalize process and set it to 0dB peak. Make sure the master volume control is set to 0dB. Then I save the aiff and import it into FCP. Then I duplicate my final sequence and label the duplicate as “DVD Final” and open it. Put the aiff on 2 new tracks, mute the other tracks, make sure the master volume control is at 0dB, then, I export using compressor to go to DVD. I do not allow compressor or anything else to touch the audio level after FCP. I know my levels are at 0DB and not going over and all is right with the world.

To sum it all up, keep using your usual standards for tape. Set your audio to 0dB peak, or very close, for DVD. That’s what Hollywood does, so it’s good enough for you too.

Thanks,
Gary

[Alexander Kallas]

Gary,
Interesting post, but some questions.
1. ” I set the levels in FCP to tape standards”-what tape standards? There are audio peak markers in FCPHD and up, is this what you mean?
2. “Make sure the master volume control is set to 0dB”-what/ where and how please?
Thanks
Alexander