The future of slating

Over the past couple of days I have been at a studio in South London shooting content for a conference.  The material consists of three actors playing five parts. It will be displayed over the duration of the conference on three screens, two portrait 9:16 ones to the left and right of the stage, and a middle one behind which is normal landscape 16:9

The actors are working off scripts, but almost every script has an alternative version we had to record.  Added to that, the actors are moving between the screens ie they walk out camera right on the left screen and then walk in camera left on the right screen.   In total there are just shy of 150 clips to be delivered.  We eventually ended up with just over 330 discreet video files at around 500GB as we were shooting 4K.

On the second day, instead of using a traditional clapperboard, we decided to try Movieslate running on an iPad Pro.  The Movieslate was synced to the sound recordists Timecode Systems timecode generator, which was also feeding the camera.

We set Movieslate up so that the Title field was part of the script we were working on. i.e “my story - start”, “my story - middle” “my story - end” and things like “into break” and “out of break”.  

Using Movieslates snippets meant it was a) quick to change and b) there was consistency across all the clips.  We used data fields for Character name, Actor and which screen we were shooting for.  We also had a “Alt script field” which we turned on and off as needed and a few clips were MOS.

Whats great about Movieslate is that you can customise pretty much how you want it and were quickly able to set it so title, character, actor and screen would come up before the board, giving the editor a very quick and easy to read description of what they were looking at. We also made sure file names matched across the board, camera and audio recorder.

They day actually went so long that by the end of it the AC was very adept at switching the info on the slate.  

After we finished filming, I simply exported the shot log out of Movieslate and hey presto, there it all was:  All the information you could need to work out what is what neatly arranged in a PDF.

But I had something else I wanted to try..

One Movieslate’s export options is FCP X, and even though I am not editing this project I was curious to see just how well the process of getting the slate info into FCP X worked.

Well it just does. And its amazing and truly revolutionary…

The workflow is really simple:

1. Export the Movieslate history to your computer (either directly or via an email).

2. Ingest the footage into FCP X

3. Once all the footage is ingested, export an XML out of FCP X.

4. Import both the Movieslate history file and the XML into Movieslates free Keyclips app.

Keyclips then matches the clips from the XML with the entries from Movieslate.  It does this matching timecode, date and frame number if its selected.

After the matching process you click on export an XML which brings up various export options - these are quite impressive and extensive so its worth going back onto Movieslate and understanding all the options depending on your production.

The last stage is to export the XML from Keyclips and import it back into FCP X.

After a moment or two, as if by magic all the information logged on set appears next to the clips in the browser. Keyclips creates a new event which contains all the keywords from the information selected with notes appearing as markers within clips.  

In my case the new event contained keywords for each actors name, each character name, script title, circled takes and which screen (L, R, M). So within a matter of minutes the logging process of the edit is done, no more sorting clips into bins or adding note information.

Using FCP X’s smart collection facility it would then be really quick to set up collections based on Character X, Screen L, Circled takes.   

And thats it.  Half the battle of the edit is done and you are just looking at the good takes.

If I was editing this I’d be laughing by now as all the leg work done by an assistant editor is done. 

Now I know FCP X isn’t everyones cup of tea, but this kind of functionality is so incredibly powerful that I cannot understand for a project like this why you wouldn’t use it. I sent a screen grab of the FCP X browser screen to the producer so we shall see what they say.

I know Movieslate been around for ages ,but for various reasons I never managed to exhaustively test it like this before but I am so glad I did. I am lined up to shoot quite a few big projects like this over the next few months and Movieslate is now firmly on the kit list.

Just a note on why we didn’t use scene numbers: The AC had been taught the British way which is not to Slate numbers as scene numbers.  Personally I prefer the US way but we were into the shoot before I noticed.

Abstract: The Art of Design

Abstract: The Art of Design is one of the best made documentary series you will see in a long time.

I am fascinated by the subject so would watch the series regardless, but it is a must watch if you are involved in documentary film production.  It is the prime example of what you can achieve with modern day production techniques.

Instead of the turn up and see what we get style of production, the programmes have been carefully thought out, researched, recced and constructed to tell each story.  The programme also use innovative visual ideas to reinforce the story instead of just simple talking heads and a bit of archive.  

And where there are talking heads they have been shot and lit properly with time and thought gone into the framing.

Its easy to see “Abstract” had a budget to die for and whilst very few production have that kind of money there are things all productions can learn. 

The myth that self shooting saves money.

In the UK, some broadcasters have whole heartedly embraced the idea of self shooting, where instead of a camera crew comprising of a camera operator and sound recordist the director shoos the program sometimes with help from a recordist but often they are on their own.

It can be done, if you have not seen it "Cartel Land" is an amazing movie.

But its a very contentious issue because when its done badly everyone suffers.

• The DP's suffer from less work (boo hoo).
• The directors or AP's shooting it suffer because they cannot concentrate on gathering the content - its impossible to bring out the best in a contributor if you are hiding behind a camera.  Psychologically its pretty simple: you cannot trust someone you cannot make eye contact with.
• The ingest takes longer = more storage cost.
• The editors hate it because they have to wade through hours of material to find the good stuff (is there is any).
• The producers hate it because they don’t have the elements they need to make a story.
• PM’s hate it because it costs more as the edit takes longer.  Oh and ask PMs about kit damage!
• Exec Producers hate it because it often leads to very dull TV - which is worse as the public then don't watch it.

The debate around self shooting has been going on and on for years now but last week on the Facebook group “People who work in TV who like to shoot the breeze” there was one of the most insightful conversations on the subject I have ever seen, including contributions from directors, editors and at least one series/exec producer.

We know it why it happens (money), but there is a big body of opinion that says this is wrong and why TV (in the UK at least) is failing apart.  

I should say again: there are some are very very good self shooters and some of them put some "camera operators" to shame.  But not every director makes a good shooter and there are some situations where you just need someone looking after the visuals, someone concentrating on sound and someone asking the questions.

Let hope someone can find a away to make the grown ups in charge of the money to realise there are better ways of making TV than sending out lonely figures into the wild with a camera and hoping something good comes back.

Making multi camera shooting easy

I decided to write this blog after watching one of the Sunday morning political shows.  Two weeks in a row they ran location interviews which were obviously shot by the same crew.  All three cameras looked different and it was really off putting. But this wasn't the worst:  A few weeks before a UK news programme broadcast an interview with a political leader where they used three cameras of different makes and models: 1 was 2/3" camera, 1 a DSLR (at a different frame rate) and one small sensor camera (XF305 most likely).  It really wasn't pleasant to watch!

Over the years I have become a bit of an expert in multicamera shooting after shooting and editing hours of interviews, conferences and chat style shows.

More and more productions are shooting with multiple cameras.  It saves time on location and the end result looks so much better than staged noddies at the end.   Added to that, mutlicam support in NLE's makes editing multi camera shoots much easier that before.

On paper, shooting with more than one camera should be simple, but there are a few catch outs that can have big knock on effects down the line.  Getting it all right on the day really will save time in the edit - and could easily save you money.

Here are a few main points I've learnt over the years:  (of course you can always just hire me!)

1. Matching cameras.  

The most important aspect of a multi camera production is getting the cameras to all look the same or match.  And its not that hard:  Just use the same camera, with the same settings and lenses.   There are a number of settings that affect the picture (e.g. gamma, matrix, black level) but most professional cameras can use a setup card so transferring settings between them is very easy.  

There is a useful cheat here: In some cases, you can use different cameras from the same manufacturer:  The Sony A7sMKII, matches the F55 really nicely.  An XF305 can work well with a C300 as a wide shot, or a PMW300 works great as the wide shot for a PMW500. 

If you must use cameras from different manufacturers there are a few steps you can take to make the editor/graders life easier.  First shoot a chart such as an XRITE Video colour checker.  Shoot 30 seconds to a minute on each camera - importantly under the same lighting set up.  Second: Match the cameras that can be but shoot Log or a flat look on the cameras that can't be.  There is another processs: you can use a LUT (Look Up Table) to create in camera looks.  It will get you close and needs to be done by someone who understands what they are doing!

(NB:It's worth noting that getting the Sony F55 to match the F5 or FS7 is very hard due to optical differences in these cameras. )

2. Choose your CODEC. If you are not shooting for broadcast, you have much more freedom to choose a codec that's more appropriate, so it's worth taking a few minutes to think about where your finished piece is ending up.  Remember, each camera will produce the same amount of data and whilst it's great to shoot 4K XAVC at 50p, if you don't have the support on set or the edit does not have enough disk space to deal with that kind of data, you will cause a lot of upset along the line.  1 hr of 4K 50p XAVC = 225 GB.  If you get the pictures matching and do not need to do a lot of effects or colour correction, MPEG50 is perfectly fine for a lot of work.

3. Make sure the other camera settings match.   Some edit systems like FCPX are actually amazing and will not blink if you throw material at them that is a different frame rates - but you do not get anything for nothing.  Behind the scenes your computer will be working that bit extra.   Allow time for the camera team make sure everything matches; frame rate (FPS), shutter speed, colour settings, codec etc.  If it's been a hurried set up, I like to do a verbal check with all cameras just before we roll.  

4. Let's talk timecode..    

The hardest part of an editors job when editing multi-camera shoots is getting the cameras and sound synced up.    On location we have to give them a quick way of finding a common sync point that occurs at the same time on all devices.  A visual ID for cameras and a audio reference for sound.  

In the days of film this was what a slate was for.  Clap the slate together and it makes a sharp noise.  And there you have it, a very simple process.  If you are shooting more than one camera make sure every camera can see the slate and you are off.

Today we have a much better system:  Timecode.

Timecode is your best friend on multicamera shoots and if you have cameras that can record it, use it.   Having matching timecode on every camera means that the NLE will be able to automatically sync the action across all cameras and more importantly maintain that sync across the entire duration of the recording.   Timecode can either be sent down a cable or there are now several devices that can transmit timecode between each device (i.e. Timecode Buddy, Lockit Boxes, Tenticle) keeping the cabling on set to the minimum, very important if you are moving about.

If you are using cameras that do not support timecode (such as a 5D or A7s) there is still a way to use it:  Timecode is just an audio signal, it sounds horrible if you listen to it but you can record it in just the same way you would another audio source.   The trick is to feed the timecode signal into an audio channel on the camera and when the footage has been copied onto a computer, run the footage threw Davinci Resolve which has a clever facility to generate new files using the timecode recorded on the audio track.

Its also worth remembering that some of the better external recorders such as the Odyssey 7Q and VIdeo Devices PixE range can also take an external timecode signal.   If I am doing a two camera shoot with my F55 and A7s, I'll record the A7s pictures on a PixE5 feeding timecode from the F55 to the PixE.   It also gives me the benefit of having Prores files which are better anyway.

if you can't use timecode but are planning to use audio to sync (i.e. Pluraleyes), there are a few important notes:

Make sure all the cameras are recording good sound at a decent level.   Use a slate App or a real slate, make sure you roll all cameras and all cameras can see the slate. Then mark it (clap the slate).  Leave the slate in vision for a moment after.

In lieu of a slate you can always clap your hands,  I prefer two books - or something with a hard edge.  But again the same rules apply..  make sure all cameras can see the "slate" and don't stick them in clap and take out - ten seconds to do it properly here will mean your editor might make it home for dinner!

And remember:  If you cut, re-slate at the start of each take.

[NB: If your shooting off speed, most cameras don't record audio so you have to do all you can to the help the editor out.  Stick a slate in front of the camera - preferably a slate that shows timecode, ID the shots and the cameras.  Anything..]

5. Monitoring.  Looking at different pictures from different cameras really confuses some people - especially non technical corporate clients.  Again there are some very simple fixes:  Again the most obvious is to use the same make and model monitor, and make sure each monitor is set up properly.    Another increasingly available option is to use a picture in picture or quad split monitor such as the Sony PVM-A170 (side by side of two inputs) or the Odyssey 7Q (Quad Split).

6. Sound. Just think about how you are going to record sound.  If you are recording two people sitting down, I would always ask the recordist to cable to two cameras.  It makes life in the edit simpler and, if something goes wrong you have a back up.  If you are recording more than one person, ask the recordist to use a recorder to record a mix and ISOs (single channels) of each mic.  Again make sure you use timecode to lock everything together.  That way its easy to pull up a individual channel if there is a problem.    Where possible make sure every camera has some audio on it.  Just in case..

7. Organisation.  MultiCamera shoots can very easily get out of hand, and if there is too much confusion it can take hours to sort in the edit.    So...  make sure all the cameras and sound roll and cut at the same time, it makes it much easier to track down missing files.  And where possible make file names match.  Use the cameras metadata to ID camera files such as "Obama IV A cam".   When offloading make sure everything is labelled in separate folders.  There is nothing an editor hates more than a hard drive of footage with everything all over the place.  Shotput Pro 6 is very useful on multicamera shoots.  

Also, If I've been shooting material on a single camera earlier during the day, then come to a MultiCamera set up, I'll start a fresh card. 

Its all sensible professional practice but following these simple steps ahead of time and during the shoot will make the edit go so much smoother.

What do you want to know..

Unlike a lot of my colleagues, I really believe we are all part of the production team and that the more we know about each others jobs - and what the equipment can do - the better and more efficient the programe making process can be.

Modern camera technology is very complex and reading the forums, its plainly obvious that some cameramen don't understand the difference between codecs, Logs and LUTs  - and in some cases don't want to know.  Yet choosing the wrong codec, or shooting Log when you don't need to can really screw the production.

I want to know how to get the best out of my camera and more importantly want to see my pictures look the best as possibly on screen.  So I make a point of reading as much as I can, and seem to understand it. 

So if there is something that you always wanted to know, ask away and I'll see what I can do.