Audio Test & Post Workflow

[trueindie]

Here is the audio (and video) test. I uploaded a first version with just noise reduction and some dialog leveling for you guys to see what I'm working on. Please ignore the crackles here and there (I did not fade the audio files in this version). Also please ignore the sudden rise in volume at 4:49. I don't want to correct this at this point. I just want to upload and go to sleep. The color also, is all over the place. I'm experimenting. If there is any color scheme anybody likes, let me know. Around what minute.

https://vimeo.com/117028317
password: audiotest1

For you audio experts:

As you can see above, I took a little bit of my edit to practice the audio process today and this is what I did:

1. Exported OMF
2. Exported Video
3. In Pro Tools:
- Separated Dialogs in two tracks for different characters
- Separated ambient noise such as chair creaks and footsteps into two separate FX tracks
- Separated music to music tracks

Then when I was going to start noise reduction, I realized that I have a bunch of broken up tracks, so I would have to do NR on each separate portion.... . So Now I feel stuck.

So now I'm beginning to think that I should change my workflow. I should
1. First do NR/DeClick/Dehum
2. Separate DIAL/FX/MUS
3. Create DIAL / M&E
4. Create Combined Track
5. Level Combined Track
6. Export back to NLE

Kindly advise. Please let me know the gaping holes that my amateur mind cannot see through. I'd love to know your own workflows.


Some other general inquiries from a baby sound person.

A. When I export to OMF am I supposed to do Broadcast Wav, or AIFF
B. Why Interleaved option?
C. Around minute 00:29 and 1:29 (among other places), you can hear the microphone away from the guy as his pitch shifts. How can I correct this? Also at 1:46, suddenly the audio sounds different. Will just lowering the volume minimize the apparent difference? Or is this the right time to learn about EQ Match? Seems like overkill in this case?
D. Around minute 2:02 (if you can ignore the crackle), how can I make the train noise more menacing? Exactly what should I do with the EQ? or is there some other method?
E. When I level dialog, after NR, it raises the noise floor in those areas. Am I supposed to do a second NR run after leveling? How to solve this problem?
F. Is there a theory on how loud the music should be in a movie with respect to other sounds?
G. Any other general advice for an Audio Post Baby

I will appreciate all comments, of any nature. If you think it sucks, that's okay. I mean, I'll be okay with it,... in about a week . But I'd like your fiercest criticisms please.

Much thanks for your time.
aveek

 

[AudioPostExpert]

Wow, to answer your questions in detail would require a book! In fact, I most strongly advise you to buy and read this book: Dialogue Editing for Motion Pictures - John Purcell. If you read no other books on sound/audio read this one. It's well written, easy to understand and contains so much filmmaking knowledge and wisdom that I consider it to be essential reading for directors and pic editors, let alone audio personnel! Make sure you get the second edition (as per the link). It's also available as an iBook on iTunes.

Every point/question you've asked is covered in the book or at least, covered as well as can be by a book. I'll give you some quick/superficial answers now though, to keep you going:

First section #3: Normally you'd have the same number of dialogue (DX) tracks as characters in a scene, in this case 3. SFX cut from the production sound recordings are usually put on PFX (Production Sound Effects) tracks.

In section 2, your suggested workflow is a bit off:

#1 should be basic dialogue editing, not NR, which should happen much later during the mixing phase (pre-mixing to be precise). If you try to do NR during dialogue editing you will find yourself going round in circles! Performing some de-hum and/or de-click processing is OK but a good general rule until you really know what you're doing is; no processing during the editing phase.

I'm not quite sure exactly what you mean by #2. I have various templates I use in Protools which are already routed: DX tracks, Room Tone, ADR and DX effects return tracks routed to a DX (Aux) track, the output of which is routed to the DX Stem and Full Mix (audio) tracks. PFX, Foley, FX, Atmos and FX effects return tracks routed to a FX (aux) track, the output of which is routed to the FX Stem, M&E and Full Mix (audio) tracks and lastly, 4 or so Music and MX effects return tracks routed to a MX (aux) track, the output of which is routed to the MX Stem, M&E and Full Mix (audio) tracks. I also have various supplementary tracks routed only to my monitors, such as; 1 or 2 DX work tracks and 12 AAF tracks, which is where I import the OMF/AAF. As I choose what I need from the AAF tracks, I copy it to the DX (or DX work) tracks. In other words I never edit anything directly in the AAF/OMF tracks, so they're always left as first imported, for reference purposes. The editorial phase therefore takes care of all the raw materials, already organised into appropriate food groups. Once editorial is complete then I start pre-mixing the dialogue stem; audiosuite, EQ, compression, effects sends and levels on each track, EQ, NR, limiting on the DX aux track. After the dialogue I pre-mix the FX and MX stems. Final mixing is then mainly just balancing the DX, FX and MX aux tracks against each other, plus a bit of fine tweaking here and there.

Section 3:

A. Bwav and Aiff are identical as far as the audio data they contain, so it doesn't really matter which you choose, although some NLEs may happier with one of these audio formats when exporting an OMF/AAF. The only difference between Bwav and Aiff formats is the metadata. At the pro level, bwavs can contain some very useful metadata recorded by the PSM and is therefore the preferred format but without pro field-recorders you cannot record that matadata and so it doesn't matter which of these two formats you use. Whichever you choose, stick to that one format, mixing and matching can cause OMF/AAF errors.

B. If you have stereo or multi-channel tracks in your NLE, the interleaved option will export them (in the OMF/AAF) as stereo or multi-channel audio files. Not selecting the interleaved option will result in stereo files being exported (and imported into PT) as two individual mono files. While the interleaved option might seem like the best solution it can sometimes cause OMF/AAF export/import errors.

I've dealt with E above.

C, D and F are big ones because there really are no rules as such, it's about using your ears and personal judgement. The solution is a combination of EQ and levels, plus maybe some compression, reverb and other effects. There's really no easy short cut, learning what these tools are and do is relatively easy, learning to actually use them, in combination with each other, takes experience. I've been working with these tools professionally almost every day for more than 20 years and I still discover something new or unexpected pretty much every day.

G. Not really other general advice, more related to C, D and F. At the moment your mix sounds flat (two dimensional) and bland, to the point of appearing unrealistic, when ultimately you ideally want to achieve the exact opposite, a feel of realism with heightened drama/emotion. It almost sounds as if it's painting by numbers or actually worse than that, because there are only a couple of numbers! A chair creaking, a quietly spoken syllable/word/phrase, a loudly spoken phase, a train arriving, etc., all sound roughly the same volume. And by volume I don't just mean output level, volume also includes other things such as EQ. For example, a quietly spoken word isn't just a lower level than a loudly spoken word, it's also a different EQ (different harmonic structure). Auto dialogue levelling and EQ matching tools sound like the perfect solution, until you realise that a lot of the time perfectly levelled and EQ matched dialogue = bland/lifeless/boring/unrealistic!

I know this advice is a lot more vague and less precise than you were hoping for but as I essentially said to start with, precise answers which actually progress your understanding are just not possible in the time/space available here. I'll try to add something a little more useful when I've more time.

G

 

[trueindie]

Hey APE. You're a Rock Star man. I f'kin luv you.

Just ordered the book. My sound guy is busy, but he'll be with me in a couple of weeks to go over post basics. I hope I'm armed with enough questions.

Yeah, you're right about the "flat" and "boring" part. Every time I hear it, I find something unacceptable about the audio that I think I have to correct in the next round. Also, I'm listening to the sound mostly on my headphones. This time when I heard it for the first time on my speakers, there were all sorts of low frequency rumbles all over the place.

This is too much work. Too much go**amn work. You audio guys deserve medals.

Please let me digest what you wrote and let's see if I have anything specific that I want to ask.

Thank you as usual APE, for being so generous with your knowledge.
Aveek

 

[wheatgrinder]

from my limited experience I can fully support your love for APE! I also attest to NOT doing NR until later. As you build up the tracks and soundscape you may find you don't really need it or you don't need as much etc..

 

[trueindie]

Anybody have any opinion on the color scheme?
Between two forums, only twenty people have watched it. But I'd be most interested in knowing if anybody has any preference for any part of the footage over another. It would help me out.
thanks

 

[Sweetie]

Anybody have any opinion on the color scheme?

What in particular are you asking? There are a lot of variances between the scenes. Saturation levels, skin tones, color schemes (cool, warm, contrast and in between) were all over the place.

As some people know, I'm a big fan of the teal/orange treatment, though it doesn't seem to fit for this film. The start has a blue wash over it, pushing some of the skin tones out of the natural feel. Without keeping the skin tones in the orange range, you're failing to take advantage of the teal/orange color contrast. Then you make a cut to more natural skin tones. If that color contrast between characters is what you wanted to achieve, you did.

It felt like each shot was colored separately without taking into account the whole feeling you're trying to achieve.

 

[trueindie]

What in particular are you asking? There are a lot of variances between the scenes. Saturation levels, skin tones, color schemes (cool, warm, contrast and in between) were all over the place.

As some people know, I'm a big fan of the teal/orange treatment, though it doesn't seem to fit for this film. The start has a blue wash over it, pushing some of the skin tones out of the natural feel. Without keeping the skin tones in the orange range, you're failing to take advantage of the teal/orange color contrast. Then you make a cut to more natural skin tones. If that color contrast between characters is what you wanted to achieve, you did.

It felt like each shot was colored separately without taking into account the whole feeling you're trying to achieve.

Good question. I don't think I even know what I'm asking.

I've seen the teal / orange thing, but I think those go better with sci-fi/action films. This film is more like a quiet drama.

What I'm really asking for is if people like any point of the film in terms of color better. Yes, the color is all over the place for different shots. Some are not even color adjusted at all, like at the very beginning. I'm just trying out the colors as I edit.

Okay, before I start going off on tangents, let me ask what I really want to. I like the portion at the end of the bar scene at 5:10 for instance. That's what I really like, but it's less saturated than at 5:05 and is a little more blueish. But I like it. Does that feel natural enough. I'm not really concerned at this point about perfection or replicating any particular scheme, I just want to know if that part around 5:10 looks like a movie or not.

I still think I asked this question badly. I don't know what the hell I'm asking. Do you like it at 5:10, is what I'm asking?

 

[Sweetie]

I just want to know if that part around 5:10 looks like a movie or not.

I still think I asked this question badly. I don't know what the hell I'm asking. Do you like it at 5:10, is what I'm asking?

I don't like it. It's a low saturation, low contrasty/flat/slightly ungraded look... Some are saying it's becoming trendy... but really, that's beside the point. Whether I like it has very little relevance to the discussion. The point of grading, much like other departments is to assist in telling an appropriate story.

If it's a sad story there's very little point in playing happy music. It's very much the same in color grading. What would really help is to know what is the meaning of your story (I obviously didn't get it). Is your story a story where she is going from no hope to hope over time (Where you may start off using a lower contrast, low saturation, unnatural colors and move towards higher contrast, higher saturation, more vibrant happy color palate) or a doom and gloom story all the way through where your 5:10 clip may suit.

Does that make sense?

Without a point of perspective of what you're trying to achieve, my opinion of whether I like the shot or not isn't really helpful.

 

[trueindie]

Fair assessment. I realize now that it's tough to give this kind of advice without knowing the whole story.

And I'm not doing it because it's trendy or anything. I don't think I know what's trendy. But I take your point about the low saturation and contrast at 5:10.

Thanks. That was very helpful