DVD Audio Levels/A.Pack Encoding

I am encoding a 16min :39 sec digital short shot on DV from the Panasonic AGDVX100A (old dated model), using FCP 4.5 HD to edit, QT Player 7.0.2, and Compressor, A.Pack 1.5 for DVD Sp 3.0.2 and finally authoring on DVD SP 3.0.2.
I'm editing my editing in FCP not Soundtrack or ProTools at this point.
Question: After I've encoded the audiophiles from my m2v (using QT PLAYER) and import into A.Pack 1.5 the audio is very low, or at least when I play the disc in my player I have turn up the volume on my DVD Player. Phasing also occurred, where the audio would be audible on channel and then the sound would drop out or there'd be no sound at all on any of the speakers. (Sound meaning dialogue, music, fx tracks, etc.) Why is this? Yet when I import the .AIFF philes into DVD SP the audio sounds much better, even though it's 2 Ch Stereo on my DVD Player with the exception of the subwoofer.
Do I need to turn up my audio levels any higher in FCP, export as .AIFF files with my m2v encoding and then import into A.Pack and reencode? Or do I need to apply any digital deemphasis/LFE low pass filter, etc.?

Please let me know if this is the correct way to do it or how to improve the process.

Robert

Hey, Robert.
You're likely experiencing Dolby's dynamic range compression.
While it's a long story, the following will surely take care of turning off Dolby's Dialog Normalization.

In A.Pack, make sure to switch the compression profile to "None" (From "Film standard". Then set the Dialog Normalization Value in the main window to -31.

Not only you don't want to raise the volume in FCP, but (on the contrary) you will probably want to turn it down a little. For some reason, Dolby's dynamic range technology seems to kick in also when your audio is too hot. Bear in mind that DVDs have much lower reference tones than TV mixes, for instance.

Exporting to AIFF and giving that to A.Pack is also a good idea.

Upgrading to DVD SP 4 is a better idea yet. Dolby encoding is now handled by Compressor, which works much better than A.Pack (which was really unreliable on dual processor rmachines).



Adolfo Rozenfeld
Buenos Aires - Argentina
www.adolforozenfeld.com

Adolfo,
This sounds like a great solution. I tried it and the sound was good. However, what is the best way to edit sound in FCP? I know soundforge and ProTools and Soundtrack are a few ways, but if I'm just editing in FCP with DV quality or even uncompressed 10-bit, how do I differentiate and make "sound" judgments in post? In terms of monitoring audio from my speakers and from a TV what is the best accurate gauge without turning up my speakers too loud and then finding out the audio levels are either too low or too high, while I was adjusting audio levels in the FCP post process? What I mean, what is an accurate gauge without having to adjust on a false accuracy rate? Sometimes I'll be editing with my speakers and they sound just fine but when I go through and output to tape or DV and encode in DVD SP the audio either sounds too low or too high? What is the best solution for this?

Thanks

Robert



Robert Rodriguez

Robert: I am by no means an audio mixing expert. Far from it. But you see, the whole problem you discribe is the very reason why we have meters. Basically, before you even begin your audio mix, you want to choose an audio refererence level. I don't mean placing it on the timeline or using it any concrete way, just choosing it. Then you want dialog to be around that reference level.

Notice that meters like FCP's have two "areas" - a solid area and the little lines above them. The perceived, average level is the top of the solid area, not the semi-random peaks above. So, basically, you want your dialog to be around the reference level, and then there's no problem if the little peak lines are -6 to -8 db higher. That usually is the case. If you want maximum confidence, there's a free audio units plug-in called Inspector, which provides RMS (averaged) metering. You will be surprised to see how much lower the average level is compared to the ephimeral peaks! That plug-in works with Peak and, maybe, there's a chance works with SoundTrack Pro too. I haven't tried that.

Two very common reference levels are -12 db in prosumer/dv environments, and -20 in more "pro" environments. The only difference between them is that -20 db can have a bit more headroom or dinamic range. One simplistic way of looking at the reference level is: since digital audio uses a different scale (db fs) than analog audio, your reference level is what will match zero on analog equpment. So, as you see, the levels we're talking are not that low as it may seem.



Adolfo Rozenfeld
Buenos Aires - Argentina
www.adolforozenfeld.com

Adolfo
Thank you for the info on audio leveling with FCP and Peak. So let's say I do mix the audio based on a reference that makes "sense" and is consistent with the soundfx of everything else, and I make sure the levels don't peak too high or low. The question is can I still keep that consistency when I go to encode the audio in A.Pack or even when I export the file as an AIFF and just want 2.0 channel PCM 48K audio? (Yes, I did peruse your other previous e-mail with normalization and dbfs, etc.)



Robert Rodriguez

Absolutely. PCM audio will be just the same. Dolby too, unless you do something that screws it and DNV/compression profiles get in the middle. It's important to understand that Dolby's dynamic range control is not really a modification of the audio information, but rather something the player executes in real time.



Adolfo Rozenfeld
Buenos Aires - Argentina
www.adolforozenfeld.com