communicating in creative contexts-pt 2

In part 1 of this series I laid out a few factors that can cause problems when clients communicate with audio people about their needs. Here in part 2 I'd like to address a specific example:

When the client asks for something that you belive he or she will not like.

I'll begin here by saying that most of what I'll be talking about is less listening to what your client is telling you and more interpreting what your client is telling you.

My baseline premise here is that your client is going to be more tuned-in to the overall creative context of the thing that you are working on, and that you're being brought into the process at some point after it has already been set in motion. This premise can fit whether you're doing sound design for a film or VO recording for a radio spot, but it's an important starting point because it forces us to realize a simple thing - the client knows what he wants better than we do - especially initially.

Now, even though the client knows what he wants and we know what we think he wants there are times when communication requires interpretation.

This is because clients will at times use certain vocabulary and concepts that can lead you astray if you're not seeking out the meanings that motivate the words. Also not all clients use confusing vocabulary, and some clients only confuse certain things so working knowledge of the relationships is very important. In addition, some clients like to manage creative concepts tightly, while others will trust us to make the best creative decisions we can, so interpreting direction can get very tricky very quickly depending on the context of the relationship.

----------

An example - I recently did a sound design project for a major convention that had a fair amount of thick low end impacts among other things that were going on. It was pretty dense, but I was happy with where I was at.

After review the client asked for more low end throughout.

Now, many clients ask for "low end" or "bass" when they really mean they want other things. "More low end" can mean some or all of the following:

- more low end
- less low end
- more midrange
- more high end
- more compression
- more distortion
- fewer elements
- I'm listening on laptop speakers
- I'm listening on my iphone speaker
- I don't know what I want to hear yet

That specific request almost always invites follow up questions from me, not only because I like low end and sometimes have to restrain myself anyway, but also because of how often that request is a red flag for something else being the problem that the client is hearing.

In this case my response was "there's already quite a bit of low end in there. We have to be careful not to lose impact on the big moments by oversaturating the low end."

My client then repsoned "well, what I really need is a big long drone for about 30 seconds before the big reveal. Don't mess with the rest if you think it'll hurt it."

With one simple prod my client was able to focus his request and I was able to give him exactly what he wanted on the next iteration. If I had just rolled my eyes and said "OK" we would have gotten nowhere.
---------------------
So with a good outcome fresh in our minds, lets look at strategy for dealing with troublesome requests:

1 - Ask followup questions

sometimes clients will give specific direction that can be interpreted in a number of different ways, or that on its face sounds like it will do more harm than good. Dont shrug your shoulders and do it, get further inside of the client's head.

2 - continue to talk specifics

If your client's request doesn't exactly add up, start offering specific alternatives. Talking in generalities is really only useful if you're completely lost. Discuss your opinion of the specific options that can help the client get to where he wants to go, and offer things that you think will get you moving in the right direction. If you start veering out into the wilderness your client will let you know pretty quickly and you can adjust course. Remeber that your client is more tuned in to the context of the project than you are. They have been living with it for a longer period of time.

3 - Don't correct language without a specific reason.

If your client is says he's hearing a "tempo change" when the tempo is constant but the instrumentation changed, then just make whatever adjustment she's asking for and leave it at that. You know what they're asking for, so don't bother spending time and brainspace discussing language - just do what they're asking.

4 - communicate technical things with simple language

I'm a big fan of simplifying language in general. Jargon is for keeping people out, common speech is for letting them in. Don't fear letting your clients into your thought processes and techniques. They're smart enough to get it, and they appreciate knowing what you're thinking. There are simple ways to describe every single thing we do, so just speak in English and not AudioJargonEese and you'll be fine. Discussing technique and methodology has often moved me quickly into where my client needs me to be or otherwise moved clients towards something that can be accomplished within their timeline and budget.

5 - prepare alt versions

If time permits, go ahead and prepare alt versions to present to the client. Always include what you believe to be the best alternative, but also allow yourself to fully indulge the client's musings as well. Sometimes they'll recognize pretty quickly that the idea isn't working, sometimes it will work and you can be the hero while hiding your surprise, and sometimes it will just move the both of you in a more productive direction. Gauge your client's imagination before going down this path. Some clients understand when they're not hearing the big picture which allows you to mockup several options quickly. Others don't have this perspective and need to be presented to in full context, which limits the number of alts you'll be able to do.

-----------

So there's a couple of nuggets from me to you. Our clients know what they want, we just have to figure out how to get inside of their heads and pull that knowledge out.

communicating in creative contexts-pt 1

Q: You are an audio person who was hired by someone that wants you to produce some form of sonic composition for a project. During the course of this project, the person who hired you starts making requests of your art that you feel are bad ideas. How do you handle this?

A few recent happenings at work along some recent internet threads have had me pondering this recently, and I'd like to share my thoughts.

Here in part one I want to get into some of what I believe to be the root causes of these types of communication problems. Understanding the causes can help us navigate to the solutions.

First and foremost, people's ears are very highly tuned in comparison to their eyes. It takes tens of thousands of audio samples per second to convince the ears that something real is happening around them. By contrast it only takes about 20 or so video samples per second to achieve the same effect. This is why film standards have 48,000 audio samples playing back every second and only 24 frames of video in that same second.

Our ears also far outpace out eyes in terms of dynamic range.

A human is capable of hearing (and usefully discerning) anything from a quiet murmur in a soundproofed room to the sound of the loudest heavy metal concert. Such a difference can exceed 100 dB which represents a factor of 10,000,000,000 in power. A human can see objects in starlight (although colour differentiation is reduced at low light levels) or in bright sunlight, even though on a moonless night objects receive 1/1,000,000,000 of the illumination they would on a bright sunny day: that is a dynamic range of 90 dB

*wikipedia

That last 10db of dynamic range perception that ears have over eyes is at the far end of a logarithmic scale, which means that the difference = a LOT.

Then there is the general function of sound in creative contexts, which is to create emotion. The visuals set the stage and to some degree tell the story, but it is the sounds that get the pulse racing or tug at the heart strings or otherwise cause the audience to feel what the directors wants them to feel. This means that sound decisions are fundamentally gut-level decisions that require technical means to execute.

Finally there is the unique relationship to time that sound has. Visuals can move fast or slow without changing intelligibility to a degree that sound cannot. Big sonic moments have to be set up with space in order to achieve full impact. Low frequency things played at high rates of speed (you know, frequencies) cease to be low frequency things. These things are part of the physics of sound and have no real correlation in the visual side of things. Also, you cannot pause a sound and examine it the way you can a moving picture. You can't just point to something and say "change that."

What all of this means is that it really takes a deep understanding of how sound works and can be manipulated for clients and directors to communicate with their audio personnel, and it truly takes years of work and study and thought to develop that level of understanding.

The end result is that in the sound world the vocabulary tends to be under-developed - especially in comparison to the visual arts. Directors have an easy time saying "brighter, darker, redder, grittier" but have an incredibly difficult time expressing parallel concepts in sound because of the degree to which sound is an abstract concept to people that don't spend all of their waking hours manipulating it. Things like dynamic range, frequency content, imaging, tempo, melody, and the effect of blank space are difficult concepts to pull out of thin air, and if your client isn't incredibly skilled in this form of communication then they are very likely to request things that are bad ideas, or to have good ideas and be unable to articulate them clearly. They're also used to being able to do things visually (like cut many things quickly for visual impact) that can have unintended and unpleasing consequences if attempted sonically.

In the next few posts I'll do my best to break down the few stereotypical "bad" requests that we sound people get, and lay out some ways to decode the client's meaning and deliver what they're looking for.

Truck record-part 4:cataloging and databasing

In part 1 of my truck record retrospective I talked about the prep work required to get ready for a full on vehicle record. Part 2 covered the actual recording date, part 3 went into the editing process, and now we'll put the whole thing to bed by cataloging and databasing.

-----------

I usually set a certain number of hours aside each week to deal with the ongoing task of cataloging and databasing the various sounds that I record and collect day in and day out. In the case of the trucks recorded for Benavides Born this process didn't really start until after the film sound was edited, which meant that there was a few week lag between recording and databasing.

Sometimes this kind of a lag can be detrimental to the process, but for a straightforward soundset like this the process went pretty smoothly (though still pretty time-consuming).

As a side note here, even though I work in a facility with 5 total engineers I do all of the sound effects databasing myself if I can. By doing the databasing I get to hear, varispeed and otherwise become familiar with all of the sounds as they go into the library. I also get to pick the metadata and can input it in a way that is both relatively standardized and as thorough as I need it to be. Filtering all of the facility's recordings through my databasing process plugs me in to everything going on with regards to the library, which makes me a more creative and efficient sound designer.

In the case of the truck recordings my labeling system is pretty straighforward and thorough. In soundminer I list the word "auto" then include make, model, type of move and fill in the mic used. Then I tag the files with the appropriate photo.

On this specific vehicle the whole process produced about 2.5 gigs of data, 66 unique files, and about 20 photos that got tagged to various moves. Here's a screenshot of my general output:



Once all of the metadata is in place I use soundminer to embed it and the photos into the wav files themselves.

The final step is to move them to their final resting place on the main sfx server, scan the embedded files into the main sfx database, and redistribute the database to the facility.

-------------

After much planning, work and editing we now as a facility have immediate access to pretty full coverage of something that was needed for and recorded in context with one of the many projects that run through here. This process is incredibly important to keep up with during the course of a project so that your internal bank of sounds grows and becomes more unique with each passing project.

Failing to edit, catalog and database unique recordings like this is just not an option for me.

--------------------
That was the whole process. Benavides Born was a true pleasure to work on, and I'm always grateful when I get to work on interesting projects with creative and interesting people. We have pretty cool jobs in the sound industry. :)

Truck record-part 3:editing

In part 1 of my truck record retrospective I talked about the prep work required to get ready for a full on vehicle record. Part 2 covered the actual recording date, and this part will go over the post and editing process with the final recorded files.

I've thought a bit about how to approach this particular blog post, and I think I'm just going to go mostly with words here and less with screenshots and audio or video.
-------------

Once I got all of the files back into the studio I dumped the raw tracks and photos into a folder on a drive called 1993 Ford F150-raw. I needed to jump right into editing the sounds into the film, which mean I really didn't have time to go through a formal library edit/metadata tag process at first.

I had tons of stuff, so I started listening to the audio slates I had given at the top, and used those to roughly rename the files. This was mostly a working naming system that looked something like Truck-onboard-tire-start stop by-20mph.

Once I had my roughs labeled I opened a 24 bit 96k protools session and imported everything. My first order of business was to come up with something from the onboard mics that I liked. Each onboard truck mic angle was put on a track, the tracks were routed through an aux and out to a mono print track. I started playing with balances and EQ until I was getting a moving vibe that I liked.

My initial impression was that I was very happy with my control of the wind during the shoot. I was also a little surprised at how much rattling and craziness was going on in the engine compartment when monitored soloed out on the big speakers. When I put together a comp track like this I'm looking for which mics and textures are working in which frequency range. For example, the tire mic was very good for mid range whirring and high end grit so I pulled back the low end there, while shelving back the top end of the exhaust mic a bit to replace that part of the spectrum with a better texture. The engine mics were used pretty sparingly, and the end result of each move was a balanced and mixed mono track. I knew I still had the split out iso's if I needed them in the edit, but I had made the determination that I'd go to them only as needed.

My next move was to deal with all of the foley. Some of the foly was in mono, some was in stereo, some was pretty loud, some was pretty soft. I did no dynamics processing here, though I did boost the levels on some of the wide angle door opens and other softer moves for usability. Mostly I was dividing sorting and naming though.

The same process followed for interior driving moves, exterior moves recorded with the shotgun, and anything else we recorded that day. Listen, divide, name, and process if needed.

This whole routine took the better part of one working day, but it made the next step very efficient - cutting the sounds into the picture.



With all of the prep, this step actually went pretty smoothly. Only in a few spots did I feel like I needed more coverage, and I was able to augment my recordings with a variety of library stuff and other recordings I'd done. I got through the entire truck edit for the film in around 2 days, and was very happy with the result.

------------------

My main takeaway from this part was that the time spent prepping the whole process paid huge dividends when it came time to actually cut the sounds into the film. A lot of sound effects editing is ingredient based cooking, and when you take the time to get the recordings right and mapped out on the front end then the back end editing becomes the easy part.

Truck record-part 2:recording

In part 1 of my Benavides Born truck postmortem I discussed the spotting, scouting and planning of the record. Now on to the actual record date.

As one final bit of planning before heading out I assembled all of the gear the night before, hooked it all up and tested it all out. I then prepped my wind protection and travel setup to be as pre-made and lightweight as possible. I transported the 552 mixer and 744t recorder in the petrol bag (wired), the mics in a hardshell multi-mic case, and the cables, batteries, headphones and hd-p2 recorder in a pair of softie tote bags. The microphones that were going to be strapped to the vehicle were pre-wrapped with wind protection.

On the day of recording this was a three person shoot: myself to record, my friend Dave to drive, and our intern Mike to handle exterior perspectives. We met after breakfast and headed South to Ennis and the trucks by 10 am.






When we arrived on the scene everyone had jobs that could be handled simultaneously. Dave went to work combing through the truck and removing or securing any loose items floating around in the cab. Mike started unpacking and setting up the recorders and cables, and I went about mounting the mics.

I used zip ties to secure a pair of Sennheiser 421s to the axle and rear frame of the truck to catch the tire and exhaust perspectives. You're not looking at the final mount points (that axle mic ended up with more zipties and rotated up to allow the axle to block the wind), but you get the idea.

Those two angles ended up being incredibly useful in post and make up the majority of the onboard sound. Also, i'll note here that the combination of towels and drafting location worked incredibly well with regards to wind protection. Dealing with wind is one of the trickiest parts of vehicle recording, and I wasn't really hearing anything out of place on those mics until we passed 60mph or so.





I placed the sanken lav mic in the back of the engine compartment away from the radiator, and had Dave crank the engine to listen for the best spot to put the SM81. After some trial and error we found a good spot shooting up from underneath at the engine from right in front of the passenger floorboard. Again, you're not looking at the final mounting setup on that SM81. Many more zip ties came into play there.

Cables were carefully run along mount points under the truck and up through the passenger window, where they landed at the 552 and 788t in the bag.

Interior perspectives came via my PCM D50.


The process of mounting, cabling, and testing the mics took about 90 minutes. We had allowed time for experimentation and listening and took our time making sure all of the mount points and cable runs were secure before setting out. With everything in place, we broke for lunch and then set about covering the shot list.

For each of the moving shots I sat in the back seat with the recorders, Dave drove from up front, and Mike followed us with the boom from outside. Our moving coverage included idling (5 minutes) starts, stops, rollbys and peelouts on gravel, some bumpy roads, and street coverage of bys at 20, 35, 45, and 60 mph. Were were careful not to spend time recording things outside of the list because we only had access for the one day and needed to cover everything within our time frame.

Here's a sampling of some of the moving truck sounds:

With movement covered we shifted our attention to foley. My foley list was pretty extensive so we used multiple mics and recorders to cover perspectives simultaneously.



I learned from participating in Tim Prebble's Doors project the value of multiple perspectives and performances on these types of things. As such we caught near, wide and int perspectives on each move, and ran 5x iterations of soft, medium and hard. Coverage for this treatment included doors, tailgate, and hood. The doors were the really important part of this though, as people were constantly getting in and out of the truck in the film and variations on perspective and performance were imperative when cutting doors in.

Once those were finished Mike and I crawled into the cab with just the MKH60 and started covering everything else in mono. Seatbelts, locks, windows, body movements in seats, and most importantly a bunch of iterations of the dry transmission shifter moving.



By now we had lost our light, so we stripped the truck back down, packed and secured our stuff, thanked the owners and headed back.

In part three I'll go over how those sounds worked in post and what I did to add them to our library.

Truck record-part 1: prep

Benavides Born has just premiered in the dramatic competition at Sundance, and as lead sound designer I thought I'd give a little insight into one specific element of the audio design: the truck.

The film is about a South Texas girl played by Corina Calderon who embarks on a journey of self-discovery in her home, the world of weightlifting, and college. Much of this journey happens in an old 1993 Chevy pickup.

During the initial spotting session I knew I'd need to find and record a similar truck to even hope to do that part of the soundtrack justice. We were working on a condensed schedule, so time and planning were very important.

Step one for me was research. I had done some recordings of my car a month or so prior, as much to scout possible vehicle recording locations as it was to acquire the sounds and practice the technique. I reviewed my notes from that session, and re-read the thoughts of Tim Prebble and Rob Nokes on recording vehicles. That Rob Nokes article is incredible and was pretty instrumental in both my car recordings months earlier and the truck record on this go-round.

I also had to find the right truck. I tried the usual methods of poking friends and family, but that didn't pan out as quickly as I was needing it to, so I started cruising craigslist, which was both enlightening and scary. In the end one of my good friends did come through with his father's vehicles which were conveniently located at his house outside of Ennis, TX - the middle of nowhere and roughly 45 minutes from me in downtown Dallas .
View Larger Map

The location I ran my car tests before was a strip of road behind a high school in Ennis, andI ended up ruling out that location for future records because of a crazy interference I was getting from what's likely a microwave tower in that part of town.

Strange flutter distortion by Rcoronado

Unfazed, I used google maps to scout other roads from afar and talked to the owners in advance to find other recording locations nearby. While I did not end up with some perfect mecca of car recording locales down there, I did manage to get what I needed cleanly enough to satisfy the film and the library.

The truck in the film was a 1993 Chevy V8. I had a choice of a 1993 Ford V8 and a 1995 Chevy V6. I went with the Ford.


With my location and vehicle secured, I set about creating my two main checklists: performances and gear.

Performances first. I opened up the film in a protools session and separated out every shot in which the truck appeared. Playing back at 2x speed helped the process. In the end I had a 20 minute compilation of nothing but truck movements. I output this as a separate quicktime in m4v format so that I could have it on my iphone or ipad during the shoot.

Then I sat down with my notepad and watched the 20 minutes down again, noting every specific move that I saw. Door open, door close, hit tailgate, pull up head-on, peel out on gravel, etc. If moves were repeated I put down a x3 or whatever it happened to be.

With my performance list in place I began compiling my gear list. From my experience and the Rob Nokes article I knew I wanted mics on the axle, exhaust, engine compartment, interior and exterior. I went with a pair of Senn 421's for the tire and exhaust mounts, a Sanken COS11 for the engine compartment, and an SM81 for another engine compartment angle. These all ran through a sound devices 442 into a 744t. For the interior I used my Sony PCM D50. For exteriors I went with a Senn MKH60 in a rycote blimp recording to a Tascam HD-P2. For wind protection on the truck mounted mics I stole from Tim Prebble and used ripped to fit terrycloth towels as well as good drafting placement. I had tested this in my initial car record and was very impressed with the results.

I was also sure to pack lots of batteries, gaffer tape, cash, cables, zip ties, water and snacks.

With my lists and gear in place I was ready for the record date.