Slater the Smart Clapboard.

I’m building a “smart” clapboard slate called, “Slater” – logically named after A.C. Slater from Saved by the Bell. It’s a purpose built digital slate that logs all of the takes as I’m shooting. It’s specifically designed to help organize media for crews that are very small.

As I’m preparing to shoot a feature as a story test with 0 crew (More Info), I’m thinking about ways that can help make the process smoother.

I won’t have a 2nd AC, or a Script Supervisor. I will be doing those jobs. So I need some tools to make sure that I am organized throughout the day.

For background, a clapperboard (aka clap slate, slate, clapboard), is a simple physical device that can be visually and audibly clapped to help sync video and audio. Additionally, it contains specific metadata written on it about the shot so that when someone is looking at the video, they know exactly what they are looking at.

In a feature shoot, thousands of clips are created. There needs to be a good way to know exactly what the clip is. Otherwise, it’s going to be a nightmare figuring out which clip is which. Because that information is written on the clapperboard, it’s burned into the video of the clip, and when you are organizing your clips, you can always see it at the beginning of every take.

On typical shoots, there is someone called a Script Supervisor who is in charge of making sure that all the shots for the movie are actually shot, organized, and additionally ensuring there are no errors. So typically, a Script Supervisor has a script and a prepared shot list, and as each shot take is completed, they log it and take notes that can be passed off to the editor later.

The person who claps the board in front of the camera is usually the Second AC (Assistant Camera). This person makes sure all of the metadata on the board is correct. This usually involves using dry erase markers to put that information on the board clearly. The scene and shot number is dictated by the Script Supervisor.

I don’t have any crew members, so I have to do it all myself.

The movie I’m shooting is fully storyboarded. I have a shot list for every scene. In fact, this was generated for me automatically by using Storyboarder (an unreleased custom feature). Therefore, I’m prepared to know at least what I need to shoot to get my scene shot. Like a Script Supervisor, I just need to go through the list of every shot I need until I’m done with the day.

Because I’m going to personally clap each take, I can automate a lot of the information on the slate. I already have all of the metadata in the form of a shot list, I should be able to use that data.

Also, because we live in a magical time where you can easily make a piece of hardware yourself using incredibly cheap electronic components from China, I can make my own “smart” clapperboard.

How it works:

Shooting

When I open the clap, numbers light up. On the first row, it shows the Scene, the Shot, and the Take number. On the next row, it shows the Timecode synced from an audio source. The next row shows me 2 E-Ink displays. The first display is a storyboard image and other metadata pulled over WiFi from the Shotlist Server. This is for my sanity to make sure that the Scene and Shot are correct. The next E-Ink display shows vanity production information. To the right, there is a blank cross where you could put custom info with a dry erase marker. On the bottom row, the time is displayed in 24 hour format, and the date in any dd/mm/yy format you want.

When I clap the board, the Timecode display pauses and shows the TC at the exact moment I clapped the board. Additionally, one of the E-Ink displays a large QR code. This code contains the TC, Scene, Shot and Take number. Additionally, and most importantly, Slater sends out a message to the Shotlist Server that logs the TC, Scene, Shot, and Take number over WiFi.

After the board is clapped and the take is being shot, only one LED number is lit: time. This shows the elapsed time in seconds from the clap. When the scene is cut, you can click cut and the timer stops. Slater sends a cut event to the Shotlist Server, and the server logs the cut. Additionally, if the take was bad, you can click a “NOPE” button which will log that it was a bad take. There are 2 other options for “GOOD” and “GREAT” takes. If you forget to click “CUT” on the board, it really doesn’t matter. It just never logs the cut.

When you are ready for a new take, you open the clap, and the Take number has automatically incremented. But if you are done with the shot, you can click the button to increment the shot. The Take number will automatically set to 1, and the board information will be retrieved and displayed on the E-Ink display. This is nice as a one man band director, because I know exactly what shot I’m supposed to be shooting.

So the great thing about Slater in this case, is that it’s logging all my takes as I clap, and can even record some basic feedback about the take, which will be super valuable when I’m editing.

Shotlist Server

The Shotlist Server is pretty simple. It contains the data of the schedule, the ordered list of shots, and all the metadata about the shots. This has been prepared ahead of time.

When I’m shooting, I need to get through each shot, logging every take. It receives events from Slater and logs the information to the appropriate take. This is super simple because the server can look up the shot by Scene and Shot number. The Shotlist Server also takes a live camera screen grab 3 seconds after a clap.

What’s also nice, is that I can rearrange the shot list on the server, as I’m shooting, and Slater obviously doesn’t care.

Shotlist Server Displays

If space permits, you can display information on large displays. If not, you can look at this information on your phone or tablet.

The first display shows an overview of the shooting day. It’s a list of every shot in the day with the metadata including storyboards. Someone can quickly glance at the screen and know everything that needs to be done. As shots are completed, they are displayed differently on the list. It shows how many takes and how much time a shot took. It also makes a rough estimate of how ahead or behind schedule you are.

The second display shows a looping animation of the current scene’s storyboards. Below the animation, it shows a contact sheet of each shot’s storyboard. As shots are completed, the frame grab from the best take of a shot is displayed next to the board on the contact sheet, and displayed animated side by side of the storyboard animation. This gives you an idea of what the scene will look like edited, and can be inspected for continuity. This also displays the shots in correct story order, so you can shoot your day out of order.

At the end of shooting a scene, it would make sense to inspect this display to make sure everything is continuous and makes story sense. If not, you can add additional shots or shoot pickups right then and there.

Media Ingestion

At the end of a shoot, you have lots of video clips, and lots of audio clips. Without an automated system, you have to manually match video and audio, hope the TC is correct and sync them, and then add metadata for each take so you can know what to edit together.

Having organized media is the entire reason this system was built.

We built an ingestion tool that takes all the video and audio clips, and processes them automatically. The tool goes through each video clip and uses the timecode of the video to look up the specific shot metadata from the Shotlist Server. If the video doesn’t have timecode, or the timecode isn’t correct, the tool looks for the QR code in the frame. The QR code contains the TC stamp so the Shotlist Server knows exactly which shot and take it is. The tool can automatically set the video clip’s TC. Additionally, the tool can find the corresponding audio track and pair it. Lastly, the tool can render lower resolution proxies.

The output is named clips, paired and synced with audio, and appropriate metadata in XML for Premiere, Avid, or FCPX.

Additionally, each scene will have an automatically generated sequence with the best takes so you can see a very rough edit of the scene, and start your editing from there.

What if I don’t have WiFi? or a server?

It’s no big deal. The clapperboard stores a log of all the clap events internally and can be downloaded later. The only issue is that the storyboard won’t update when you change the shot.

Why not use or create an iPad app?

I want a single purpose rugged device. Slater is a tough guy. Unlike an iPad, Slater is always ready, and the battery will last for a month.

Hey! You’re missing a lot of info on the slate: Camera letter? Day/Night? Int/Ext? MOS? Sync? Roll or card? Camera/lens info? FPS? Filter info?

As far as the important information I need, the only real important information is: Scene, Shot and Take to identify the specific clip, and Timecode at the clap to use for syncing audio. All other information on the board is superfluous because I have it on the Shotlist Server.

Camera letter? I know it from the clip automatically, and confirmed by the Shotlist Server. Day/Night, INT/EXT is useless info for the editor, but again, I have it. MOS (video without sound) is known ahead of time, so the clap just occurs normally and the metadata is added upon ingestion.

Roll is almost completely useless in the case of digital video. In film, that shit was very real. If you were editing a positive you would want to know where the original negative roll was. But in this case, you have the mother fucking video clip! Why do you need to know which card it was on after the fact? Roll doesn’t need to be on a modern slate.

Additionally, you may have noticed that I don’t use the American style of shot numbering. In American productions, they combine scene and shot where the scene is the number of the scene, and the shot number is a alpha character. So the first shot of scene three would be “3A” and the second would be “3B”. I think this is to disambiguate the scene from the shot. So in conversation, you could say, “We’re going to shoot 3A again” instead of “3-1.” Additionally, before the clap, they say the scene number and word associated with the alpha character so for example: “3 Apple”, “3 Baker”. This is in case the audio is astray, you know which shot it belongs to. However, these are not real problems in modern times. And additionally, humans are prone to error specifically when doing precise tasks. This results in common mislabeling of shot letters, disparate from the shot log. The system is so dumb that they specifically have to leave out letters of the alphabet like I, L, O, S because they look like numbers. Instead, I increment shots by number, because I can easily increment and decrement in my head, it works better digitally on displays, and it just makes sense. It’s also referred to as the European System. So if you love America and you voted for Trump, I’m sure the American system is fine for you. But sorry, it’s super dumb – a real Zack Morris thing to do. Slater is better than that.

Where can I buy this?

You can’t. I built it for our own productions. And the reality is, it’s built to be improved upon, making it better or more specific with every iteration. BUT, you know, it will be open sourced! So I’ll have the hardware schematics and source code on Github. Maybe I can even make a video of how to put the whole thing together.

Notes on hardware

Since the Arduino, making hardware has been democratized to the masses. In the past few years, China has made it so you can buy any component you want, at basically cost, and have it delivered in a week (If you live in NYC). Additionally a company called Espressif makes a $2 chip called the ESP-32. For $2, you get a dual processor with a healthy amount of ram and built in WiFi, in the size of a postage stamp.

The ESP easily drives the 26 7 segment displays, E-Ink displays, takes input from 12 switches, and communicates everything via WiFi.

The TC is smash synced from an analogue audio signal that extracts the SMTPE LTC data and keeps sync using a separate RTC (Real Time Clock).

Lastly, there are 6 18650 Li cells which are connected to a simple battery charging circuit and a 5V regulator that supplies power to the system.

The clapper itself is purchased from Amazon for $20, and ripped off the crappy slate.

The enclosure for the board is 3D printed framing to attach components sandwiched in between two pieces of laser cut clear acrylic.

The buttons are simple pushbutton switches. There is a recessed hall effect sensor on the bottom clap that senses when the clap is open. There is a recessed switch next to it that will be closed on a hard clap, but not on soft claps.

The E-Ink displays are not cheap at $75 each. I needed one that had a fast enough refresh rate for the QR code to be displayed quickly. But to be honest, I’m not sure how much this is needed if your TC is synced in camera.

What do you think?

I’d love your feedback about this! If you have any thoughts, you can email me at [email protected] or text me at: 917-696-5465.