Audio Post-Production Tidbits

Useful tips and tricks for audio post-production gathered from around the web. Follow @AudioPostTidbit on Twitter for updates!

[audioposttidbits-blog]

I was recently checking out the special features of the 'The Bourne Identity' DVD and noticed there was a short featurette on the sound crew that wasn't advertised on the back of the box, called 'Speed of Sound'.

The interview section of this featurette, which you can watch above, is pretty interesting but what follows that is even more so: you get to hear the different layers of the car chase sound mix in solo. It gives you about 8 different layers of foley, dialogue, car sounds etc which you can listen back to on their own, an interesting insight into the different elements that make up a good sound mix. I'd recommend checking it out if you have the DVD of the film!

[audioposttidbits-blog]

This is a video following the sound team for the video game 'Far Cry 2' in Africa as they look to record some ambiences to add to the realism of the game.

It is interesting to see that at about 1:14 one of the team members shows how he is using binaural microphones and a Zoom H4 to capture the ambience of a busy African town. I would have thought that whilst binaural recording is great for playback through headphones, a more traditional stereo microphone set-up would be more suited for a game audio recording that will most likely be played back through TV/home theatre speakers. However, I have not played Far Cry 2; perhaps the ambiences turned out great regardless!

[audioposttidbits-blog]

On expanding stereo ambience to the surrounds...

Tips for making a full surround ambience when you only have a stereo one available:

"In a pinch, I'll put an M/S Processor in line with a PLII decoder and print the result. Specifically, EQuality in side mode followed by Surcode PLII. When I like what I'm hearing, bounce. Or if I'm doing on with CPU let it run in real time."

Posted by Jesse Peterson

"Take a longer stereo bg and cut it twice at even intervals. Throw the 1st third in L/R, 2nd third in Ls/Rs, and 3rd third in C (you'll only use one channel). Quick and dirty way to make a 2 track bg into 5 without boning your mixdown- also doesn't add unnecessary regions to your session."

Posted by Mister T

"I usually try to find some BG`s sounds that could have relevant "hard sounds" that could "sync" to the image. Sometimes it helps to improve the sense of realism or immersion."

Posted by rcutz

"I don't have any issue with simply panning some BGs partially into the surrounds. I usually don't want the surrounds to be super discrete anywy, because I don't wnt to draw attention to them unless it's a story point. I want the feeling of a unified field, without excessive detail in back. That means some tracks I'll pan 50-60% back, some will be hard left right, some will be 50/50 Left right and some will be mono center. That creates transitions rather than discrete zones, which I greatly prefer. Sometimes I will put a BG in the surround only, but not very often. I have used 4 mono cricket tracks, one in each speaker, but only if they are somewhat similar in frequency content, and then I always overlay a stereo track that plays in all speakers in order to glue it all together. The glue track is usually played fairly dominant with the monos enhancing the stereo effect a bit."

Posted by ggegan

Personally, I've always thought the method of cutting up a longer stereo ambience and routing one part to the front left and right pair  and another part to the surround pair works well enough. If the ambience contains any obvious, stand-out sounds however, such as a police siren or audible conversation, you need to be careful they don't repeat in the front and back of the soundfield as this will sound odd and can confuse the listener.

Original Source of Quotes: http://www.gearslutz.com/board/post-production-forum/624799-backgrounds-stereo-into-surround-techniques.html

[audioposttidbits-blog]

The SoundWorks Collection website is a fantastic resource for behind the scenes info on audio production and post-production for many blockbuster films.

The Inception video embedded here is just one example of the 'film sound profiles' they have available on the website, each offering an interesting insight into the workflow, techniques and creative reasoning behind many sound design choices in big budget film. Check it out! 

[audioposttidbits-blog]

Whilst this blog is intended to discuss matters relating to post-production sound, I think it is important to consider the role of production sound every now and then.  

This is a link for a forum/discussion group which is focused on production/location sound and is inhabited by many pro sound recordists and production mixers currently working in the industry. It is full of invaluable advice on technique, equipment and general know-how for getting on in the world of production audio; definitely worth checking out if you want technical details on how audio reaches the post stage or perhaps if you sometimes find yourself doing both production and post work.

I found the workflow section of the forum particularly interesting, with technical details of how production audio is captured in respect of what camera set-up is being used on the shoot and many insightful anecdotes of experiences on set. 

[audioposttidbits-blog]

BBC Audio Delivery Spec...

Here is another sample delivery specification, this time from the BBC:

http://www.bbc.co.uk/guidelines/dq/pdf/tv/tv_delivery_to_network_programmes_v1.0-2011.pdf

It is interesting to note that they give example volume levels for different parts of the mix when metered using the BBC standard PPM meter:

This gives you an idea of what the levels of different elements of the mix should be set to when delivering for the BBC.

For a software, plug-in version of the PPM meter, check out zplane's PPMulator+:

http://products.zplane.de/index.php?page=ppmulator

The BBC also offer an open source version of the PPM meter, however it is not as user friendly:

http://sourceforge.net/projects/baptools/

It is also interesting to note that the BBC spec is set to be updated to reflect the new EBU recommendation R128. This could change how loudness is metered in broadcast post-production and is worth keeping an eye on.

Here are some further tips gleaned from the web:

"SD stuff is line-up at PPM4, no peaks above PPM6. The bit about stereo in here applies to all BBC channels, even though this is a document for HD delivery. My understanding is Dorroughs are only useful in the US and have been made pointless by loudness metering. For PPMs, get something like PPMulator and do not reply on the "BBC" setting in Digi Phasescope - utterly useless.

It's worth adding, that although you could do it just reading dBFS, it's worth getting PPMs for the integration time, as peaks could actually be momentarily above -10dBFS, so you can go a little louder, rather than if you were just doing it with digital meters."

Posted by tom lowe

"ref level -18dbFS = PPM 4 = 0.775v Peaks 8dB above ref level (i.e -10dBFs), PPM 6 on a BBC PPM (integration time of 40ms I believe) which is a quasi peak meter as defined by PLoud, the EBU Loudness experts"

Posted by MIKEROPHONICS

Original Source for Quotes: http://duc.avid.com/showthread.php?t=298881

[audioposttidbits-blog]

On tackling sibilance...

Today I was working with a dialogue track that had some distracting sibilance issues. I was having trouble taming this sibilance with the Pro Tools De-Esser plug-in so I looked for some tips on the net. Here is some of the advice I found:

You have to listen for what frequency range is hissy, though, and that can vary widely from singer to singer. Sweep with an EQ and find it if need be. It's there somewhere, hissing away.. I still keep de-essers early in the signal chain because they're more of a quality control mechanism than any type of specific shaper. I'm usually past -15dB (further from zero) in terms of threshold, but that equates to between 2 and 5 dB worth of attenuation at the de-essing stage, which is not drastic. I don't like getting rid of esses altogether since they're a very important part of speech and coherency. A helpful tool is sidechain listen, like on a Waves DeEsser. Then you can really hear the band of frequencies you're pulling back and it makes pinpointing the "hissyness" easy. Even barring that, you can still find esses with a narrow-Q parametric and either pull those down via EQ or set a frequency range for the de-esser. Multiband compression works in a similar fashion and it can help some, but the frequency bands are much larger than that of a de-esser. I still notice myself using less de-essing after multiband, however.

Posted by Synonym Music

There is often an EQ point somewhere between 3500 and 5K that seems to calm some voices down, or make them less harsh anyway.

This is particularly true of sibilant people recorded on lesser condensor mics like the Senn.ME66 onto mini DV type cameras--this combo of gear makes sibilant voice sound very harsh. Soemtimes we just attack the offending syllables directly, cutting and reducing their volume (which is what the de-esser is doing anyhow).

A side note--often when a very sibilant person is recorded in a reflective room we find that we can get at the core sibilance, but still have the broadband reverb of that sibilance left in the ambiance around the voice.

Posted by Philip Perkins

The best and most under-control de-essing is done by drawing ultra-short volume-curves around extreme esses or other percussive hits like knocks etc that spoil the overall dialogue quality.

Posted by wildtrax

Despite these tips I could not fix my dialogue tracks as well as I would have liked; I think it is much more desirable to fix sibilance issues at source if possible!

Original sources: http://acapella.harmony-central.com/showthread.php?2347386-Sibilance-and-plosives-de-essing-lesson 

and: http://help.lockergnome.com/movies/Post-tools-dealing-sibilance--ftopict75254.html

[audioposttidbits-blog]

Sample delivery spec...

As a student, I always find it useful to look at audio delivery specifications to see what is expected by various networks. Here is one for FX/Nat Geo/Speed:

FX/NATIONAL GEOGRAPHIC/SPEED NETWORKS prefers audio to be produced and delivered in 5.1 Digital Surround Sound, prepared for AC-3 Surround Sound transmission. Audio may be delivered in several different formats as indicated below. The audio dynamic range of the material must be suitable for television broadcast. Audio levels shall be measured using the recommended loudness measurement found in ITU R BS 1770 standard, known as LKFS. The loudness is relative to full digital level (0dBFS) and represents the subjective loudness of the audio program.

The loudness shall be measured across all channels (except LFE) using the LKFS technique and averaged across the entire segment or clip. The FX/NATIONAL GEOGRAPHIC/SPEED NETWORKS target loudness level is -24 dB ±1 dB. Instantaneous true audio peaks shall not exceed -2dbFS (BS.1771), while average peak audio levels should not exceed -8dbFS (EBU Tech 3205 or equivalent). Audio should be mixed such that the measured average loudness (using LKFS) of the LT / RT mix shall be equivalent to the measured average loudness (using LKFS) of the 5.1 mix, within ±1dB. The individual channels shall be in-phase to prevent cancellation when tracks are mixed together. Care must be taken to ensure that all elements of the audio are properly prepared, so as not to cancel out if played back in a 2-channel downmix or in a monophonic mode.

Original Source: http://forums.creativecow.net/thread/30/860629

[audioposttidbits-blog]

Walter Murch explains the concept of 'worldizing'. This involves re-recording sounds in a real location to add the acoustic characteristics of the environment to the original sound.

"This is what I discovered Welles had done in a more primitive form in Touch of Evil. What he had not done was combine the original recording and the atmospheric recording. He simply positioned a microphone, static in an alleyway outside Universal Sound Studios, re-recording from a speaker to the microphone through the alleyway. He didn't have control over the balance of dry sound versus reflected sound, and he didn't have the sense of motion that we got from moving the speaker and moving the microphone relative to one another."

Walter Murch quote from http://filmsound.org/terminology/worldizing.htm

[audioposttidbits-blog]

On getting 'that' voice-over sound...

"Have you ever tried the same mic, the same artist, same position just switching from card to figure 8? Figure 8 will sound way warmer, you don't have to put your mic in front of the glass if you're worried about reflections from it, simply move it a bit or even better place the mic with the sides facing the reflecting surface (aka the glass), sides on a figure 8 are dead. It's the nature of figure 8 polarity, it's just warmer.. Try it, for me (and for many others) has been one of those "aahhh!-moment" sometimes you see obsessing over a preamp or a compressor that will give you more bottom, more beef, more body when simply switching from card to 8 it changes the sounds 10 times more than a different pre or compressor."

          Posted by Retinal

"Anyway, maybe I still need to elaborate. I almoast daily record VO for TV-trailers. That kind of VO is very present-upfront-in your face so it survives between commercials and informs people at home. The EQ and mixing style is something completely different when you listen to Vo integrated in a fiction series, like Dexter. I tried to emulate this for a very dark trailer with a specific style, and I find myself cutting high and high mid like I never done before, a chain of 3 compressors to get is really tight with one multiband to control low end. But it still was with a test voice, so now I am curious to record it with the real VO who can give me that neutral Dexter tone and have enoough wamth in his voice."

          Posted by BrettMorello

Original Source: http://www.gearslutz.com/board/post-production-forum/618942-dexter-voice-over-sound.html

[audioposttidbits-blog]

How to record various sound effects commonly needed in sound design work using everyday items.

[audioposttidbits-blog]

Recording 'whoosh' effects with various objects, some on fire.

[audioposttidbits-blog]

On LFE use...

"I never use the built-in LFE send for this reason. Not only does PT not LP the signal, it also doesn't add any "BOOM" to the signal. Better to use a regular send to an aux that has Lowender or Recti-Fi inserted. Then assign the output of the boom aux to the .LFE sub-path. This is how most in-the-box mixers I've worked with do it. This simulates the dbx subharmonic synth box that traditional dub stages use. When I boom something, I want it to BOOM. All of my fx channels have a reverb send and a boom send as part of my standard mix template."

posted by starcrash13

"I'm using Nuendo but it's basically the same process. Each one of my stem busses has a mono LPF on the LFE channel as a catch all sort of thing. This makes it easy to just do the send method if I need a little oompf. But then I also have dedicated LFE boom channels that I can send to.. these are loaded up with sub harm generators in bypass. I'm also experimenting with other plugs on these LFE channels like GRM and realtime pitch bending. Can make for interesting automated bass stuff on the fly."

posted by Jesse Peterson

Original Source: http://www.gearslutz.com/board/post-production-forum/622341-lfe-pt-low-pass.html