Audio Mixdown

I believe, but can't prove yet, that we are being hosed by accidentally including 44.1 sound tracks in our 48 timeline. The hosing consists of producing garbage DigiBeta tapes seemingly at random. Playing back the devastation an hour later appears to be the only way to know we made an error. There are no other indicators or alarms.

Apple suggests that Audio Mixdown only prepares a timeline for live playback and doesn't actually render anything past that.

Suppose we included a 44.1 clip in the middle of a 48 sequence. Further supposed that we performed an Audio Mixdown. Then, instead of being good little editors and Edit to Tape, we exported a self-contained QuickTime and pushed it off to another, different Mac for Edit to Tape.

What's in that show?

If we assume that editing to tape with a mixed sample rate timeline will produce a damaged tape, why, if this is indeed the case, would either a Kona2 or an Io pass it? Both of those units will perform genlocking and certain repair functions--to the video.

Further, why would the timeline not turn colors to indicate an odd, damaged, or dangerous sound track? This leads inexorably to the question of when do the audio tracks change colors--I mean the render bars?

Koz

> Apple suggests that Audio Mixdown only prepares a timeline for live playback and doesn't
> actually render anything past that.

And I gotta take issue with Apple on that, Koz. Both of us have noted repeatedly that a movie-file export -- which is not real-time -- can also be bungled if Audio Mixdown isn't performed.

In my experience, a dropout due to lack of Audio Mixdown is inconsistent, but a movie-file export can fix that in stone -- a movie file made from this can fix the dropouts in place, making them appear at the same spot.

I do Audio Mixdown so often that it's rare for me to encounter a constant, trackable dropout when outputting to tape. I've been seeing frame-rate problems working with 60i DVCPro HD in the last week, where even after a full render and Audio Mixdown I get consistent freezes in the image. But it could be a drive-speed/bad-data issue, since if I took that flawed movie file and stuck it into a 29.97fps DV NTSC timeline, rendered, done an Audio Mixdown and then exported to a DV NTSC movie file, some of those frozen spots went away, against conventional wisdom about consistent dropped frames.


www.derekmok.com

I got three Audio 101 questions about audio mixdown. First of all, how often should I perform an audio mixdown? I thought that you only do this before exporting a sequence as a QT movie or to tape.

Second, in my sequence I have got WAV, AIFF and MP3 files, some bounced from ProTools at 44.1 and others at 48. If I understand Koz correctly, I may have problems with the 44.1 files. Should I rebounce all files to AIFF 48, or will an audio mixdown be enough?

Lastly, I have centered all my dialogue, but not music or sound effects. Should I be centering this as well? I'd like all the sound in my final audio mix to be evenly spread in both LEFT and RIGHT. Will the final audio mixdown do this for me if I have centerd everything?

> I got three Audio 101 questions about audio mixdown. First of all, how often should I perform
> an audio mixdown? I thought that you only do this before exporting a sequence as a QT
> movie or to tape.

My approach is to do one whenever I need a clean playback, not just limited to tape outputs, but whenever I want to hear everything as it should be. It's a fast render, so I do it two or three times every half hour. Nick Meyers has argued that this is too often, but not for me, and it doesn't slow me down.

> Second, in my sequence I have got WAV, AIFF and MP3 files, some bounced from ProTools at
> 44.1 and others at 48. If I understand Koz correctly, I may have problems with the 44.1
> files. Should I rebounce all files to AIFF 48, or will an audio mixdown be enough?

Sound files at 44.1kHz sample rate can cause problems in rare cases. But if I were you, I'd worry much more about the MP3 files. You can't use them in a timeline, period -- you will get dropouts. Convert to an AIFF first.

> Lastly, I have centered all my dialogue, but not music or sound effects. Should I be centering
> this as well? I'd like all the sound in my final audio mix to be evenly spread in both LEFT and
> RIGHT. Will the final audio mixdown do this for me if I have centered everything?

If every sound file is centered, then there is no left/right distinction.


www.derekmok.com

- HOW OFTEN:

yeah, i'd do it just before export
no need for me to mixdown my audio as i work.
my system copes just fine with previewing it.

one benefit of the mixdown is that FCP doesn't have to search al over the drives for the many bits of audio needed for the sequence.
mixing down creates one simple file
it may actually be producing a series of files, but they will all be in one location on the drive.

to my mind, doing many mixdowns throughout the process would negate the value of this:
creating many files all over the drive that FCP has to hunt for.

but if it works for Derek, then it works, i guess

- 48k?

agree w Derek.

- PANNING

"If every sound file is centered, then there is no left/right distinction"

by which Derek means
leave you music and SFX as Stereo.
as it will sound better that way, and "leave room" for the dialogue in the centre.


nick

Thanx!

<<<mixing down creates one simple file
it may actually be producing a series of files, but they will all be in one location on the drive. >>>

Just to cover the bases, wouldn't you get an auto render if you exported a Self Contained QuickTime (formerly known as Final Cut Movie)? That certainly produces one file, gathering all the bits and pieces from the show clips.

If you don't, then there are some very messy conclusions......

Koz

> Just to cover the bases, wouldn't you get an auto render if you exported a Self Contained
> QuickTime (formerly known as Final Cut Movie)? That certainly produces one file, gathering
> all the bits and pieces from the show clips.

If computers always followed logic, that would be the case. But I've observed since around 2001 that your movie file can have dropouts that disappear the minute you do an Audio Mixdown. As for why the export process renders everything else but leaves the audio, I have no idea. I just know that Audio Mixdown is not negotiable any time you do a tape output or movie-file export.


www.derekmok.com

And one more point, It's a better than even chance that we're trying to send 44.1 audio to a DigiBeta machine resulting in a very unhappy machine. Does "Audio Mixdown" force conform all the 44.1 clips to 48, the project settings?

Everyone keeps saying "Audio Mixdown is essentially an Audio Render. How essentially?

We've been accidentally spraying 44.1 clips throughout hours of 48 final work. Nobody noticed because we never go to tape. Until now...

Koz

What the?! I'm totally floored by this thread.

I've never known about the need to do an audio mixdown on anything. (OK, I've only been using FCP for less than a year, but still...)

I've done plenty of QT exports, as well as recording to DVCAM and Digibeta.

How can it be that doing this is so important if it's not mentioned anywhere. The manual only refers to using the mixdown command to create a preview file to increase real time playback performance. Nothing about NEEDING to do it prior to an export or edit to tape.

I've got three training books (including Larry Jordan's excellent FCP 5 book), and I've taken two classes so far, and again, no mention of this anywhere.

And for god's sake, if an audio mixdown is necessary to an accurate QT export, than why doesn't it happen automatically as part of the export QT movie command?

And can someone tell me what is the difference between choosing Render - for playback, and Render - mixdown?

That's because problems caused by a lack of Audio Mixdown don't always happen. You could have a perfectly fine QuickTime movie one day, and then a week later you might hear dropouts. So I've decided to work the process into my output procedure rather than waste 15, 20, even 45 minutes waiting for a movie file that might not be perfect, in lieu of the eight to 10 minutes that it would take for a long sequence to go through Audio Mixdown.

> And for god's sake, if an audio mixdown is necessary to an accurate QT export, than why
> doesn't it happen automatically as part of the export QT movie command?

I've always believed that it's not the application's job to do most things automatically. If FCP were to start applying Broadcast Safe "automatically" to my movie-file exports or tape outputs, I'd start to get very worried. Learn the command and use it.


www.derekmok.com

I guess I'm just a little perturbed at only learning about this by accidentally reading a thread on this forum. I mean heck, I've even got the almost 900 page "Optimizing Your FCP System", which has an entire chapter on outputting your project - no mention. Makes me wonder what else I'm missing...

And Derek, I must respectfully disagree with your point, because I don't see the comparison. While there are plenty of reasons I wouldn't want a broadcast safe filter applied to my output or export, there are zero reasons I'd want audio dropouts.

-Ross

I've also never done an audio mixdown, but I do render everything before any kind of output. Doesn't that achieve the same result?

> I've also never done an audio mixdown, but I do render everything before any kind of output.
> Doesn't that achieve the same result?

No. But as I wrote above, not every show has problems associated with Audio Mixdown. I just find it much more expedient to do it and then ask questions later. Also helps eliminate possibilities when you do have glitches.


www.derekmok.com

If one uses PTV (Print to Video) it does all renders as well as mixdown automatically.

Michael Horton
-------------------

> If one uses PTV (Print to Video) it does all renders as well as mixdown automatically.

But not Edit to Tape? That seems odd...

Not sure. Probably does. I never use it. ETT is more for insert editing than anything else, at least for my uses.

Michael Horton
-------------------

So then what is the difference between a render and a mixdown? How do I know when to use one or the other?

When you do a render all you are not rendering your audio, just pictures. A "mixdown" renders your audio without touching your pictures. 2 separate things just like audio and video are 2 separate things.

Michael Horton
-------------------

<<<No. But as I wrote above, not every show has problems associated with Audio Mixdown. >>>

That's what's so darn much fun about this. We are a multi-Avid, Multi-Mac, and Multi-Other shop and it's taken us weeks to nail this down. Remember Typhoid Mary? I mean, not personally, but the story? It was only after enough damage had been caused that somebody began putting the pieces together.

That and we never lay off to tape. This doesn't affect Quicktime Exports or DVDs. We turn out thousands of those. The Sony people have given our decks a perfect bill of health with the comment that, "there's something wrong with the show applied to the machine." But they can't tell me what.

<<<If one uses PTV (Print to Video) it does all renders as well as mixdown automatically.>>>

But not Edit To Tape? That's good to know.

When I get back to the shop I'm going to rip apart the last Quicktime that failed. If I inspect and find 44.1 audio at all the places that the DigiBeta failed, then I'll know. Who knew that you could have a Quicktime Movie with multiple audio standards in the same track.

Koz

Edit to tape does do an auto mixdown, if you haven't already done one. I also do them manually beforehand, though, because the PTV and ETT versions are not retained, so if you need to stop going to tape for any reason, the machine will have to redo the auto mixdown. If you've done a manual one you're ready to go again immediately.

Koz - in theory once you put your audio on your 48k timeline it will be converted to 48k. The output will not contain multiple audio standards. A mixdown will make this a smoother ride, though.

Also, here's some stuff straight from the manual regarding audio mixdown.

Use the Mixdown command to prevent dropped frames: If you have a complex audio mix
with numerous clips, keyframes, filters, and transitions, the mix may start to affect the
playback of your sequence. Also, if the Audio Playback Quality setting in the General
tab of the User Preferences window is set to High, you may get dropped frames in your
audio. You can use the Mixdown command in the Sequence menu (available by
choosing Sequence > Render Only > Mixdown) to render all the audio tracks in a
sequence, along with their accompanying transitions and filters, into a single render
file. For more information, see ?Using the Mixdown Command? on page 546.

and

Mixdown: Renders all the audio in a sequence to a single group of render files, one
for each audio output assigned to the selected sequence. This can improve
playback performance by eliminating the need for Final Cut Pro to do any real-time
mixing or audio effects playback. The Mixdown command is a nondestructive
operation and doesn?t change any of the audio clips in the sequence.

Note: This command has no effect on how your clips are edited. All audio clips remain
on their own tracks in the sequence just as they were before. The Mixdown command
only consolidates the audio for playback in an audio render file.
Once you?ve applied mixdown audio to a sequence, the menu item is dimmed and a
checkmark appears next to it to show that the audio in the sequence has been rendered
as a preview file. If you move any audio clips afterward, the preview file is discarded, and
the Mixdown command again becomes available in the Sequence menu.


There's abou another 20 entries in the manual about audio mixdown, but they pretty much say the same kind of thing.

<<<Use the Mixdown command to prevent dropped frames>>>

We don't have dropped frames. During the Edit To Tape, the process goes perfectly smoothly. The tape machine goes in with no fuss (we are using Assemble Edits into factory fresh tape). We are playing from a single show file exported by a second FCP machine for this purpose.

The show proceeds to the end and many eyes on several television monitors can find no difficulty at all. The layoff maintains perfect lip sync beginning to end.

We just can't ship the tape because it won't play back. We get Channel Error red lights and the playback picture breaks up and the sound is always horribly distorted. It's a permanent broken tape. It won't play back on either machine--although it breaks in different places. I put that down to an unstable tape. One machine is better at covering up than the other.

It's not the tape machine. These machines are used constantly 12 hours a day and there is a line of people waiting to use them. They only have troubles with certain rooms--like mine.

Both of the On-Line Macs do this on this show. In one 30 minute layoff, certain clips break the tape playback and certain ones don't.

.....I'm playing back Friday in my head. One of the machines didn't play from the timeline. We loaded it into the Preview Window, set the in and out points and laid it off from there. Again, it seemed to go perfectly but produced a garbage tape.

There are certain common threads:

--All the failed clips seem to come from one Avid Room.
--Everything goes through the master SDI Routing Switcher.
--All three failed rooms are professionally broadcast genlocked, just not to each other.

It is a puzzlement.

Koz

They did figure out a way to get the show out the door. Edit it to BetaSP and dub that to DigiBeta.

Ich. Pooy. Yuch.

Koz

>>All the failed clips seem to come from one Avid Room.<<

Can you do a test by getting something small out of one of the other avid rooms and seeing if the problem persists? Not a likely candidate, givent hat it gets all the way to output looking fine, but you've got to start somewhere.

If you don't mixdown, then FCP uses the current Audio Playback Quality setting in the user prefs. This is especially important for mixing sample rates as when the setting is set to Low (the default), FCP does not up or down-sample audio to match the sequence settings.

When mixing down- you are assured of high quality audio rendering and playback. You don't even need to change the playback audio quality setting as the mixdown always renders to the best quality, properly up or down-sampling audio as needed.

I'm pretty sure the Optimizing book does cover audio mixdown- its been a while since I looked at it but I'm almost positive that it mentions the whole issue with mixed sample rates and mixdowns.

If you have never heard of the need to do a mixdown- I'm extremely surprised. I can't imagine a training book, video or class that doesn't cover this basic need before outputting.

BTW- doing a Render All before output does not do a mixdown- it only item level renders audio- in other words the clips are rendered individually but the sequence as a whole is not. It will still play back multiple audio files in real time. Only a mixdown specifically creates a single mixdown file, lowering the overhead on FCP for playback and output.

Andrew Balis
Lead Final Cut Pro Instructor
www.moviola.com

<<<I'm extremely surprised.>>>

We never go back to tape. Ever. So when the job came down it turned over a rock and the live playback problems crawled out. We knew about most of them, no multiple FireWire drives, export a self-contained movie, etc., but this one killed us. What do you mean you can't play the tape? It laid off perfectly.

<<<I can't imagine a training book, video or class >>>

Oh, yes. Several people have been to classes. They all work somewhere else in the building. We're going to get around to reading those books one of these days...as soon the workload settles down.

It's still possible that the audio mixdown isn't the problem. That's just the obvious first place to look. Our Sony support rep is coming by Friday. There are diagnostics inside a DigiBeta machine that might be able to tell us stuff. That and we know that one Avid is broken. We proved that over the last few days.

We also know there is a Kona2 problem. The default/presets don't "stick".

Koz

>
> > Second, in my sequence I have got WAV, AIFF and
> MP3 files, some bounced from ProTools at
> > 44.1 and others at 48. If I understand Koz
> correctly, I may have problems with the 44.1
> > files. Should I rebounce all files to AIFF 48,
> or will an audio mixdown be enough?
>

It does cause some occasional problems- I had my files playing from weird places of a piece of music other than at places where i specified the in/out points, audio skipping (cd imports). I heard that the possible cause is because of the on-the-fly conversion that final cut uses. MP3s are the worst- i had problems exporting even the omf files once because I forgot to replace the scratch mp3s with the final wav files. Beyond audio skipping (if that isn't bad enough), I don't encounter any other problems (it used to play at a different speed in final cut 4.5, i think on-the-fly conversion was one of the features of 5.1).

From the description, an audio mixdown would be good enough (it'll create a full audio file at your sequence setting), but i'll be wary and check for skips/lip syncs, etc.