Brian De Palma’s editor was the Artist in Residence at The Manhattan Edit Workshop. We’d just watched The Black Dahlia, and we had the opportunity to ask questions.

The film has a lot of voiceover – in the style of the 1940s movies that De Palma wanted to emulate. Bill gave us a couple of tips. Even if you know you are going to have voiceover in a scene, make it work without. It’s important not to introduce the voiceover too early in a scene. Establish location first. Make sure the audience has taken in the scene’s pictures and sound. Then let them see the characters – understand their initial roles in the scene. Make sure that when you do introduce the voiceover that it doesn’t distract from what is on screen. The voiceover overpowers most images and sound from the scene, so be careful.

Start the voiceover at the point the audience becomes curious as to what is going on.

In the Black Dahlia and Assault on Precinct 13, Bill put some fades to black between scenes. We asked him if that was due to act breaks. In some cases it was, in others he said that fades can be used for another reason.

If you do not want the emotion from one scene to be carried into the next, fade to black between scenes.

Bill said that filmmakers know that the effect of one scene on the next might be too strong – it is a good idea to give people time to think on what’s happened, and permission to start a new emotional line in the following scene.

Once editors have worked with directors for a while, they should be able to read the dailies. They can tell how the director planned the scene to be put together. That’s the version you should show the director first – even if you think that there are better ways of doing it. If you omit dialogue, stay in the master when there are many cutaways; start distilling the scene, the director will want to see footage the way they planned in the first place. There might better ways of making it work, but they need to see the way they planned to do it in the first place.

The first version of the film should show all the scenes that the director shot. That is, even if not all the scenes need to be in the film, the director and producer need to see a version of all that was captured. The editor can also create their version in parallel, but the first version to look at will be the ‘script assembly’ version.

After refining the film, fixing scenes, changing the structure, Bill likes to go through each of the scenes again, looking at all the dailies for each scene. It may turn out that once the scene it cut into the film, an alternate reading of a line may work better for the film. This can be seen when the scenes are in context.

That was some of what he talked about that wasn’t specific to the films we saw.

On Friday, Bill Pankow, Brian DePalma’s editor came to visit the classroom. He talked about The Black Dahlia and Redacted, Brian’s next film.

We asked him questions about specific films and about editing in general.

I asked about the reel change marks that I saw during the new Die Hard film. Films used to be delivered in a series of 20 minute reels. Cinemas used to have two projectors for each film. The first reel plays on the first projector. The second reel is threaded into the second projector. Ten seconds before the first reel runs out, a mark appears in the top right corner of the screen. The mark looks like a cigarette burn in the film. That prompts the projectionist to switch the projectors over when they see the second cigarette burn, which appears a few frames before the end of the reel.

The reason why this is important for editors is that it isn’t advisable to have a reel change within a scene. We have to allow for a few seconds of the film after the reel change not to be shown. The projectionist may switch too late, and the first few frames of the next reel may be missing – they are the frames left dangling on the outside of the reel if the leader isn’t re-attached properly.

That means we can’t have any important frames in the first few seconds of each new reel. Also the score and sound shouldn’t carry across reel boundaries or there will be a jump in the soundtrack.

This means that you need to bear in mind reel breaks when putting your film together. You need to ‘balance the reels’: make sure that each reel is as close to 20 minutes in length as possible, but not over.

These days films are played using a single huge platter connected to a single projector. So why are there the same marks in modern films? Bill said that films still need to be projectable in reels for executives and for festival showings.

Yesterday the Manhattan Edit Workshop class went to Splash Studios in Chelsea. Pete Levin gave us a tour and told us a great deal about the sound component of postproduction. At places like Splash we are known as ‘picture editors.’

The cliché is that sound accounts for half the movie going experience: If the picture is good, but the sound is bad, then the movie is bad. But so-so pictures can be saved by great audio.

It’s a pity that producers and directors do not spend as much time selecting their sound recordist as they do their picture recordist (director of photography).

Just as picture editors solve production problems, so do sound editors, sound designers and mixers.

The production process for sound is divided into two stages: preparing the sound elements, mixing the sound elements.

Preparing the elements starts of injesting in the correct format, editing the dialogue, sound effects and score to make them ready for the mix.

A dialogue edit is needed to make sure that each transition between sounds is not noticeable by the audience. It is the nature of a hunter and hunting creature that our brains are especially attuned to a change in sound. When we were hunter-gatherers having a good sense of hearing and a brain to interpret sound meant that we could eat instead of being eaten.

It is the nature of production sound – sound recorded on set or on location at the same time as the picture is recorded – that each recording setup sounds slightly different for the any other. Take location recording for example: a scene between two people talking in a diner. The scene starts off in a wide master shot. We then cut into a two shot, then further into over the shoulder shots, followed by close-ups. Each time the camera is moved, the microphone needs to move. Closer to traffic, away from a wall that reflects sound, closer to clothes that absorb sound. This gives each recording a different ambience along with a different set of background noises. The general background sound heard on location is known as ‘tone.’ Tone is different in different sorts of locations. The size of the room, the materials that go into its construction, the location of the building that it’s in. That’s what makes a mortuary sound like a mortuary, a café like a café and a police station like a police station.

If a dialogue edit weren’t done, you’d hear the ambience and tone change every time the picture editor made a sound edit. Picture editors don’t usually edit picture and sound at the same point, but they do need to make the edits, and the sound editor can smooth out the transitions in the sound edits.

The main technique used is to gather as much tone as was recorded for each microphone placement, and use that to extend each sound edit so that the tone can cross-fade into the tone for the new shot.

For example, our two people talking the café. It is common for editors to cut to the person reacting to dialogue so that we can see what they are think of what is being said, and so that we can see them preparing to respond. We then usually let the shot play so that we see and hear them reply. They might reply less than a second after the first person has finished speaking (in fact people in real life usually start talking during the last syllable of the phrase the other person is speaking). As the room tone being recorded during dialogue for the first person is usually different from the tone for the second if a dialogue editor didn’t hide the edit, the audience would hear an abrupt change in room tone between the first and second speaker.

Dramas are usually shot and edited so that audiences believe that the scene is playing out in real time (with multiple invisible cameras and multiple invisible microphones and invisible crews recording everything). The trick of picture editing is the hide the picture edits so that the audience doesn’t notice the transition between shots. It is the same with sound editors. They use different techniques to achieve their goal.

Dialogue editors use pieces of room tone to extend the outgoing audio track by a few seconds, and use the room tone that matches the tone of the incoming audio to make the incoming audio start earlier. That means that this dialogue is ready for mixing. When the scene is mixed the sound mixer can fade up the tone for the second shot during the last few seconds of the first shot, he or she can then fade down the audio of the first shot during the first few seconds of the second shot. Sound mixers know how long these crossfades need to be to hide the edit in the audio.

Adding tone to dialogue tracks also helps when you need to change or replace the dialogue on another track. When actors come into the studio to record replacement words, there needs to be the tone of the room the scene they are reworking underneath the new dialogue.

As well as the dialogue, sound is made up of sound effects: Foley and otherwise. Foley effects are the sounds that people make in their everyday life by interacting with their environment. The most obvious is their footsteps. It is not usually possible for sound recordists to record each footstep actors make. People make other sounds too: they open doors, handle cutlery, plates and glasses. Their clothes make noise as they rustle. These are the kind of sounds that we don’t notice until they are absent from a scene. Other sound effects include traffic, the noises that bits of equipment make in the office or home.

There are many libraries of sound that you can license or buy. A lot of the time these sounds don’t quite match the pictures you have. You can modify a lot of sounds to suit using software plug-ins, but the best solution can be record the sounds yourself. This is where sound designers ‘worldise’ sounds: if sound is coming from a TV, radio or other mechanical device it is simpler to play the sound on the device and record that. Simpler than spending hours tweaking the sound using software in the computer.

Places like Splash Studios also record dialogue that needs to be replaced. This can happen because there was a problem with the original recording, or a line of dialogue needs to change for script or censorship reasons. This process is known as ADR. This stands for Automated Dialogue Replacement (or Automatic Dialogue Replacement). The trick with good ADR recordings is to position the microphone in such a way that the recorded sound matches the picture. If you take account how far the audience feels they are away from the actor on screen, that’s a clue to how far you place the microphone from the actor.

Once you have gathered the sound elements together, it is time for the mix. For each sound, the mixer has many choices to make: position, relative volume, and EQ.

Position: where the sound should be positioned in the sound field. The choices for stereo soundtracks range on straight line from left to right. The choices for five-channel sound range from left to right and from front to back as well.

Relative volume: How loud the sound should be compared to all the other sound in the scene. A telephone ring should be quiet when being played in the background to enhance the feeling of being in a busy open plan office. A telephone ring should be very loud if it heralds the call that the protagonist of a scene is waiting for.

EQ: The set of frequencies that are used to play a sound can be modified by changing the EQ setting: how bass-y or tinny you want the sound to be.

Once these choices have been made, the sound mixer can start coming up with mixes for scenes.

The process for post-production usually has the sound house create a pre-mix with the materials supplied from the production. This will allow for a dialogue edit, and will use the sounds gathered by the picture editor to make the best mix possible before sound design and Foley work is done. The director and producer listen to this mix and give notes to the sound team who then go on add sounds not captured during production.

Of course many of the productions new editors will be part of will have little or no budget for a separate sound editor. That means that picture editors need to learn a little about how to edit sound.

Pete’s strategy is to organise the sound. Keep dialogue on the same set of tracks. Keep Foley together, keep other effects together. Create separate tone tracks. Keep score on other tracks. Get organazized. Don’t worry about learning the software: spend your time learning to listen to the audio – then you’ll start having some idea as to what you want the software to do.

If you want to learn more about sound editing, take a look at The Film Editing Room Handbook by Norman Hollyn. It was last updated in 1999, but still gives an overview of sound editing and very detailed look at the relationship picture editors have with sound editors.