Personal View site logo
Make sure to join PV on Telegram or Facebook! Perfect to keep up with community on your smartphone.
Please, support PV!
It allows to keep PV going, with more focus towards AI, but keeping be one of the few truly independent places.
Post Production Sound Tutorial (+ a bit on location sound)
  • 63 Replies sorted by
  • 8) This is where a lavalier mic comes in handy. They allow you to mic the talent and capture dialogue where the use of a boom is difficult or impossible. There are a few downsides to using a low-budget lav setup, however, which relegates them, at least in low-budget narrative filmmaking (or where you want to recreate reality as closely as possible), to worst-case-scenario status. Lavalier microphones have an omni pickup pattern which means they capture sounds coming from all directions; they don’t exhibit the same frequency response curve as a shotgun/cardioid mic and because of their proximity to the mouth sound very dry and clinical which makes them a challenge, but not impossible, to match in post; they have to be wired to a separate recorder (more gear/batteries + syncing in post) unless you use a wireless transmitter which also implies more gear/batteries - plus transmitters also have their own issues notwithstanding the fact that they don’t have as wide a frequency response as a hard wired setup unless you’re using very high end gear = $$$. Within a low-budget filmmaking context lavs are best used to complement a shotgun (for intelligibility) or when you have absolutely no choice.

  • 9) A slate costs about $20. Get one and use it for each shot. Not only does it force you to be more organized but it's use helps you to identify your takes and offers a visual/sonic point of reference to perform audio/video syncing in post. This is particularly important where the camera is much farther from the action (i.e. using a telephoto lens) than the mic. Sound travels at 340 m/s so if the camera is 7m (21ft) from the character speaking the camera’s audio will be delayed by 20ms making your in-camera audio an unreliable sync reference. The slate also acts as a safety if your in-camera's audio track is too low or gets corrupted.

  • 10) OK people, this is THE most critical component to ensuring smooth, flowing dialogue transitions in your film. In a perfect movie, dialogue doesn’t explain WHAT your characters are doing but WHY they’re doing it. That makes the dialogue mix crucial to the story. And anything that takes your audience out of the story, like a fumbled/mumbled word, set noise, rumbling stomach, sneeze, sniff, etc. needs to be corrected. Roomtone is the sonic glue that allows you to slice/dice/fix/nix dialogue issues transparently. If you have not recorded roomtone during your shoot you will be forced to create it from whatever snippets of silence you can find in your dialogue tracks. This can be a very long and extremely aggravating procedure; and you will need to do it since roomtone is absolutely essential to post-production dialogue editing. You cannot simply go back to the location and record ambient sound as the sonic character of the set will most probably have changed significantly by then making any such recordings useless.

    What is the easiest way of acquiring roomtone? Easy. If you are the director, make sure for each shot that you WAIT AT LEAST 3 SECONDS BEFORE CALLING ACTION after your gear has started recording. Five seconds is even better, eight fantastic. Ensure that everyone (actors, crew, producers, investors) involved in the production on the set knows from the outset that for the few seconds prior to action after the camera starts rolling everyone has to be dead silent, i.e. no whispering, moving, sniffing, scratching, etc. For example, when you’re ready to do your first take shout “Silence on the set”; once everyone is quiet tell the soundman to “roll sound” then the camera operator to “roll camera” (they should each confirm by replying “rolling”) then count a few seconds in your head, visually queue the slate person to call the shot and once he/she’s out of the frame call “action”. You’ll regret not having made this simple but time-saving effort during post.

  • Thank you, @spacewig. Much appreciated.

  • @spacewig Thank you for all the effort

  • 1, 2, 3, 4, 5... SUPER Comandante Espacioooo!!! +)

  • Thanks!

  • Great stuff @spacewig ! You should put this together on its own blog page somewhere for filmmakers to read since this covers many perennial questions really, really well. (I'd love to have this as a guest post on my filmmaking blog if you don't have better options.)

    A couple thoughts: When one is running sound, it is crucial to listen. Sounds simple, but I think that's where a lot of DIY filmmakers get in trouble by paying more attention to the levels meters than the actual sound. Listening to rehearsals and prior takes really helps in setting proper headroom (knowing whether a big outburst is coming up in a quiet scene, for example). At the same time, it's frustrating in post when the sound dynamics change significantly between different takes or angles of the same scene--which makes cutting these together a challenge--so maintaining consistency is equally important. Sometimes it's better to "ride the levels" or just record a second track at a lower level to cover the yelled portions.

    Labeling/naming tracks and using the "Slate mic" functions of better recorders is such a nice treat for post. As is keeping a good log of tracks to give to the editor (this is often done in conjunction with the script supervisor). These are the organizational grace notes that make an enormous difference on large projects like feature films and series.

    There's also a lot of experience involved in selecting and placing lavs and wireless packs so they don't rustle or appear visibly. I'd love to hear your advice on that.

  • Your recommendation to record a section of room tone with the takes is a great one since a dedicated room tone pass often gets missed in the hustle to move on. I've been doing this for years since it's such a great compromise between the need for great sound and the need to move quickly.

    What I've come around to as the best practice is to do this at the end of the first take on any angle (or any significant sound or lighting revision--such as a fan switching off or a new light coming on). Many actors don't like holding for room tone at the beginning of a take when they are mentally prepared for the scene--and it's important to prioritize performance. However, once the scene is done, no one minds holding for ten seconds. So at the end of the first take, I typically call "Everybody hold for tone. Camera can cut." Then give it a silent ten count (without any noises) and move on.

    The sound guy will love you because he knows you give a darn; the actors won't mind the interruption; the AD will appreciate not having to do a separate roll for room tone; the editor will always know where to look for tone. It's a fantastic system. The only gotcha is that it can be hard to remember to do it when you're a director with notes, etc. to give on the take. Once you get used to it, it's easier. It also helps to let all relevant dept heads know this will be the way you're doing it. Everyone gets used to it really quickly. I would not recommend doing it on every take. Just the first one on each angle is plenty--especially if you get a nice long chunk.

  • @douglashorn First off, muchos gracias for taking the time to respond with thoughtful feedback. Your contribution is essential to the enhancement of this topic.

    I completely agree with you regarding the importance of "listening". I would go further and argue that someone with zero experience as a sound(wo)man would benefit immensely from buying a cheap eye mask, the type you wear on a plane to sleep, and spending a day wearing it in different locations just to experience and appreciate their surroundings from a sonic perspective.

    Riding levels is definitely part of the techniques a sound person needs to employ to perform their function optimally. The complication arises from knowing when an adjustment to the gain setting is justified versus allowing a drop or increase in level due to the natural dynamics of speech in three dimensional space. My feeling is that this is a decision made largely based on experience making it difficult to impart any advice to persons whom this thread is intended for, i.e. indie filmmakers with very little or no experience with sound. This complication is further compounded by the impractical implemementation of gain settings on many cheaper recorders, i.e. access through a menu and modification by pushing buttons rather than twisting a knob. I would love to hear from seasoned pros on the subject.

    I want to make absolutely clear, in the event it wasn't in my previous posts, that I am NOT, in any way, shape or form, an authority on the topic of location or post-production sound mixing. I don't make a living from it and, compared to others actually working in the industry, have very limited practical experience. However, on top of the 20+ years of music recording/mixing experience I've accumulated combined with the multitude of shoots I've been on in the last 4 years I feel I've developed a certain "philosophy", in the pure sense of the word, that may have enough value to share with others given the dearth of information on this topic. This is a long-winded, yet important, preface to my next point which is that I cannot offer any advice on the placement of lavalier microphones as I really don't have enough experience with them to elaborate on the topic. However, I whole-heartedly encourage and urge those with relevant experience, such as yourself, to contribute to this thread so that we can expand and enrich its content. Again, I will be happy to modify the wiki with all positive contributions that add to the knowledge base.

  • In reference to point 10 above:

    'Roomtone' is often referred to as 'Atmos' recording. It's common practice to get a 30sec to 1min recording of the shooting environment after the last take is complete. This allows flexibility in sound editing when there are background sounds that may be needed to cover dialogue edits that have caused sudden changes in background sounds. Traffic sounds that change over time is a good example.

  • @caveport Thanks for the reminder. Unfortunately it's too late for me to update my post but I will add to the wiki tomorrow. Cheers.

  • This is great stuff! One thing about the gear is that I always buy three of everything instead of two, and wire everything redundantly to the mic, with a backup plan as well as a backup stereo mic. For example, using a Y cable you can split the mic into two recorders, and if a mic fails, you then have an exact replacement (the third mic you bought). If the recorder fails, since the signal is split you lose no sound at all. If you don't have that, well, then that's the very rare but total megabummer gigcrashing nightmare. If you are on a budget, use a DR 680 and mix the output down to stereo and run that into a zoom or Fostex for a stereo backup.

    I also make sure that one of the recorders will function if unplugged, like the Fostex FR2LE. Some people use a UPS instead--either way, you should have plan if the power cuts out for a minute or two. You can run purely on batteries, but batteries are crappy and if they momentarily dip voltage your gear may stop. Pulling four batteries out of a Fostex, you will usually see that just before the recorder stopped you had three good batteries and one not so good one.

    I always keep a small, low noise stereo mic on one of the cams as a third backup. Obviously, both this cam and the attached mic will run with their batteries for hours. It also helps with sync as it gives a wave form more similar to a recorder.

    When using a slate, or clapping, make sure you do it fast, and do it three times. If the stop part of the motion is trapped between frames, it can give a misleading sync. If you repeat the process at the end of a very long take, you can average the drift out as well.

    For very accurate work, in addition to the slate, a "clock pass" is a good idea. Take a numerical display clock or app with tenths (or more) of the seconds and pass it in front of the cams. You can then use these images to reconcile the clocks on the cams in post if needed.

    I don't use close micing, but most people do (well, almost everyone, really), and you certainly can't go wrong with it, except for popping and sibilance. So don't do it my way :)

  • Good stuff guys - thanks everyone for tips.

  • @spacewig, I can't thank you enough for this.

    I can vent about my current project because nobody else I know is on this forum. It was shooting video of a play performance in a theater with a huge echo. Previously, I'd used a camera track and a Rode Videomic, and had decent results. But this was a special project, so I hired a sound person to rig up a professional-level sound recording. He spent hours rigging up four Sennheisers around the main stage, and a portable recorder and a few smaller, lapel-sized mikes up on an upper stage area. He gave me the tracks he recorded.

    Thing is, the tracks he gave me had problems. They seemed to have been recorded at a very low level. Some of them cut out into silence for short periods. The upper-stage areas... well, when they did record, their quality was actually far clearer than the sound from the main floor. (There was still room echo, but that was never going to be eliminated anyway.)

    I couldn't get that sound guy to help with the mixing, so I hired someone else. He took the first guy's tracks and managed to assemble a final that... well, it was more uniform, but it had its problems. Lots of drop-offs when the sound levels got really high, i.e., screams, bangs, etc. A personal issue came up and I had to stop working with the guy. (In his defense, he'd never mixed for video before, and his system didn't include the ability to watch the video: he was strictly sound-only. And as I said, the tracks I gave him were disappointing.)

    So now I have to learn sound mixing to try to salvage something out of this. It turns out that two of the cameras captured fairly decent tracks. One camera was decent all the way through, except that it was a Canon, and had to be stopped-and-started every ten minutes or so. The other's sound levels were VERY high, with lots of peaking and cutoffs, but it got some good portions of the upper-stage area. (The end result will probably be messy as hell. But at this stage, I can live with the mess.)

    Sorry to go on at this length, but the bottom line is that I have to train myself in doing location sound and be able to do it myself. Obviously, I can't just leave it up to a sound expert. So tutorials such as yours are a HUGE help, and at times like I'm facing, an encouragement. Thanks a lot; I'm looking forward to more.

  • I've never heard of clapping a slate three times. SOP is to quickly say the scene and take number, then "Marker" and whack it. If you're on a close up and close to the talent's face you say "Soft sticks" and do a soft clap. If your video is being manually synched, the Asst Ed should be able to quickly find the clapper peak on the waveform. The word "Marker" right in front of it confirms that it isn't some random sound on set.

    Of course a timecode slate with a digital bleep sound at the open is easiest but these days I only use these on big shoots. (Clients and producers love to see them, though they don't add THAT much anymore with timecode sync, PluralEyes, and the rest.)

  • @spacewig @all What do you recommend for eliminating echo from spaces that tend to produce echo? (I'm interested in techniques for both when recording and in editing.)

  • Best way to remove echoes on set is to use sound blankets on as many surfaces as you can. (Generally ANYTHING not on camera.) Echoes are really the hardest things to remove. It's like the audio equivalent of a big camera flare--darn tough to get rid of in post. Though, perhaps there's now some smart software that can do this.

  • @Douglashorn Thanks. Any mics better than others for minimizing echoes?

  • @Brian_Siano I don't think your comment "Obviously, I can't just leave it up to a sound expert." is justified. I worked as a sound operator for some years and all I can say is, you hired the wrong guys. Quote: "(In his defense, he'd never mixed for video before, and his system didn't include the ability to watch the video: he was strictly sound-only.)" These guys were not "Expert". Good luck learning how to do sound.

  • @matt_gh2 Microphone pickup pattern is what to look for. Hypercardiod shotgun mics for recording at a distance or boom pole. The only other thing you can do is get the mic's as close as possible.

  • @DouglasHorn well, now you have heard of it :)

    If you use the slate and film the clock in each cam, one clap should do it, then you can decide if you want to compensate for drift. I often do not bother with the clock, so a few clacks, which takes zero time and zero effort works for me. If you are running more than a dozen cams they may not all be framed on the clapper, and then with that many cams I definitely use the clock. I would never use a slate with a clock and a beep, the clock has to be fed to each can, with confirmed focus, plus the wave form of a beep is not as sharp.

    Also, the slate is not really necessary since you can clap your hands, but it looks more professional (so I don't use it, but I can see why some do)

    One more thing about the backups--if your costs are really high, you have to have it. If your costs are low you can consider a smaller rig. But either way, I recommend calculating the costs of reshooting a day if your audio gear fails. If it's six figures, well, there's your answer.

    Whatever works! What I like about spacewig's plan is that he has a real plan and thought about all the steps.

  • @caveport Thanks for advice - much appreciated.

  • @matt_gh2 it depends on the space. A cardioid is my first choice for eliminating echoes, a hyper/super/shotgun will have echoes from the rear lobe in either a live hall or a church, which is accentuated by the nulls on the sides of the mics. A cardioid has no rear lobe. However, in a narrow hallway, you want the hyper like the Schoeps MK 41.

    You can also make a really great baffle for your mics to help with echoes. I also would not use blankets without a bass trap.