Meeting Topic: Mixing/Sound Design/Music Composition for Heavily CGI'ed Projects
Moderator Name: Rob DiVito
Speaker Name: Steve Gamester - Producer: Entertainment One Television; Daniel Pellerin - Sound Mixer: Daniel Pellerin Digital Sound Production
Other business or activities at the meeting:
Rob DiVito greeted everyone.
He made a call out for volunteers to join the executive committee.
He thanked Sound by Deluxe and Daniel Pellerin for preparing this meeting practically at last minute's notice referencing a re-scheduling of this month's original meeting topic.
Meeting Location: Sound by Deluxe: Theatre One - 424 Adelaide St. E - Toronto, ON
Rob introduced Steve Gamester and Daniel Pellerin, and started to give their backgrounds. Daniel has worked in sound and music post production for film and television for the past 34 years, as recording and re-recording mixer/engineer, sound design and music supervisor/editor, specializing in digital 5.1 audio formats for film for the past 16 years. Steve is Director of Development and Specialist Factual at Entertainment One Television.
While Rob read through their very extensive accomplishments both said 'that's enough!'
Additional info about Daniel can be found here:
Daniel opened the presentation. He was joined by Steve who's producing the 6-part series "The Perfect Storms". It's interesting to note that, in addition to an extensive production background, Steve has a Masters degree in Public History and has certainly put it to good use for this production.
The presentation focused on how the elements of a series as "dense and rich" as this - which involved intense research, sound design, music, CGI and dramatic re-enactments - are merged successfully.
It highlighted the recently completed episode "The Perfect Storms: Galveston" which was about the devastating hurricane that tore into Galveston in 1900.
For most of the evening the audience was treated to before and after clips from the episode to illustrate the often dramatic difference from pre release mock ups to the final version, and to show how the overall process evolves in a series like this.
Daniel described this process as 'not as complicated' and rather intuitive. What was special about this series was that it was done differently than what is the industry norm. Musical ideas were created before the actual footage and editing were completed. This was a new way to work for Steve. So this evening was also an opportunity to see the fruits of this collaborative process.
Daniel introduced Steve who discussed the conception of the series. The whole impetus behind this series had to be epic. Scale was necessary. Sound design was an important component especially for the CGI portion. He also discussed the necessary research involved in preparing each episode in the series.
The episode shifts from interview segments to voice over to re-enactments with CGI. Daniel stated that any form of film that they mix, they depend on how good the dialogue is and how good the recorded tracks are. Actual sounds of hurricanes were used, specifically from Hurricane Ike.
All the music was composed by Mark Stewart. The music they receive is actually split and organized into 28 tracks which they pre-mix. This way, during the mix they still have full control, in case elements need to be changed or enhanced.
CGI elements take time. The challenge is it has to work in tandem with the story. They found a way in this series to communicate changes to the sound people - some scenes had as many as 25 re-writes. Time is the essential element in creative work.
Finding a narrator is another key element. They felt blessed to have Kenneth Welsh who was able, with two hours of recording time, to deliver the entire episode complete with alternate takes. His voice contributes the necessary empathy, tension and gravitas, yet "communicates" to both genders.
Responding to a question, the narration was not recorded at Deluxe but at Regent Theatre, using two mics, in a custom built booth for Mr. Welsh.
Another question addressed factual errors found later on in the production. Steve said it's inevitable. The questioner mentioned an error in the previously viewed clip incorrectly identifying longitude as latitude! Daniel welcomed these questions and observations especially before the official release.
Some mock ups were less than inspiring: Steve discussed another episode in this series where they created a Lego model shot on an iPhone to give Daniel and his team an idea of what to expect from the final CGI!
Responding to a question asking if CGI is ever complete before sound is added, Daniel mentioned there is a span from four days to a week where the CGI is locked (but not necessarily colour corrected) where things get intense because he has to work really fast. They are seeing the actual visuals for the first time and the sound has to match those visuals. Many fine details have to be articulated in a short period of time.
With a question about track count, Daniel explained it's not the count but how precise the tracks are. There aren't as many as one would imagine because there's not a lot of time to mix it. For the particular clip in question there were 32 of FX, 32 of backgrounds.
Another questioner asked if music stems were used: Daniel replied that the music is raw and organized on a 'grid'. He praised Mark Stewart's work perfectly complimenting the visuals at all times. The secret to the composing relationship is: "the less the producer hears 'friction and trouble', the less he feels it's a technical juggernaut, and the better off the team is". The process is gratifying because everyone is working in the same direction, including the effects editor. As a supervisor now, Daniel won't allow the 'traditional' conflicts to occur. Everything - cues and effects - is worked out before it gets to the soundstage because there isn't time to try things out.
A final question before the break concerned production and mixing notes, and the means of communication; ie: email, phone, etc. Daniel replied each of the picture editors start deciding what they want to hear right off the top. They send a list to put in a schedule, these effects get sent to whoever is editing the picture and put into the system, so as they're cutting, the effects are there as well as an accumulated library of effects and stemmed cues, creating their own cues. Mark Stewart will enhance these cues while keeping the original spirit intact.
Just after the break, Daniel began by thanking Deluxe for setting everything up for the presentation of tonight's meeting. He spoke more about Mark Stewart, and more of the history aspect of Galveston.
He also talked about the recreations in these episodes which he felt, from experience, were quite breath-taking, marvelous, and "inspiring" to work on. Steve added each episode had only 4 days to shoot the re-enactments. Documentary elements were shot first and assembled, then the drama elements are decided after that. Because they can't afford actors speaking much dialogue, sound design is used to convey emotion.
Some of the documentary visuals included stereoscopic slides from 1900 just after the storm; as well as actual film footage shot by a Thomas Edison apprentice. This footage is considered to be the 'first newsreel'. Discussion continued briefly concerning the three hurricanes: Galveston (1900), Katrina, and Ike.
Daniel spoke more about the composer Mark Stewart. Previously, Mark worked at sound design and was very good at it.
After the last clip, there was time for questions. Responses follow:
Each episode, from conception to completion, takes about 10-11 months. Steve said this is still more time than typical.
How does dialogue and music affect detail? Daniel replied it doesn't affect anything, really. He felt it was more about 'how do you build sound to wrap around dialogue'. Much of the sounds used are actual sounds. Dialogue is key. Some shows are un-successful because of lack of clarity. "If' you're going to be content heavy, you're going to have to respect content".
How do current broadcast standards affect mixing? Daniel began by stating this show is 0.1 below 23 LUFS and "clamped" there. The mix was dynamic but never overwhelming or over powering. He contains things to the point that it doesn't feel like it was compressed. The final reading was 24.9!
Considering the time and budget and the sonic resources required, were most of the sounds from a library? Daniel said not really, though not every library is going to have every sound. He continued talking about creating sound libraries and how editors trade sounds to grow those libraries. Ultimately libraries are just a tool. For future sound editors, he recommended recording your own sounds as much as possible.
He discussed recording Ken Welsh with two microphones. One mic - a Neuman - was tilted upside down pointed toward his chest, and the other was a Sennheiser 416; both out of phase with each other. Daniel doesn't use EQ: he combines the two tracks, sets the levels and the compressor. If he wants the voice brighter, he tilts one of the mics, for example. If he wants more bottom end, he'll 'mix' it in real time - again - by moving mics. It goes much faster than post EQ. And since there are two mics, in post they'll be panned a bit left and right respectively to make the narrator sound 'bigger' - not louder.
Regarding loudness, an audience member wondered if the mix's final level was where it was because broadcast standards might not have been applied the same way across the board. He continued on about how some mixers run mixes through a -10 limiter just to be on the safe side. Daniel feels that sounds squashed. He prefers an airier sound with push and pull. He clamps down on the elements that are going to create problems. He monitors with the room pinked at 84 (dB). Some sounds are compressed before going to stems if needed.
Scoring is first 100 per cent editorial and then 100 per cent to picture again. "Editorial gives us their ideas and Mark improves it". Steve thinks working with the music from the beginning and being in a constant dialogue made a huge difference editorially.
Is this relationship to the sound team typical or unique to this series? Daniel believes other groups have similar relationships. When he supervises a sound project everything has to be on the table. Planning is key.
Were live musicians used? Mark Stewart is playing all the keyboards live to the pictures. He can practically score in real time. He sends changes back in 15 minutes.
Wrapping up, Daniel thanked everyone at Deluxe again.
At the close, Rob thanked Steve and Daniel; and presented both of them with AES certificates and mugs.
Written By: Karl Machat