200 Days with Shooting with RED and post workflow

Discusss your Cinematography hardware, Cameras, Lenses, Grip, Rigs etc here
Post Reply
Karthik Ganesh
Newbie
Newbie
Posts: 11
Joined: Tue Sep 29, 2009 3:21 pm

200 Days with Shooting with RED and post workflow

Post by Karthik Ganesh »

Hi, this is an article I had written for the cinematographers combine meet for the Digital Camera Assessment Series on the 26th and 27th of June. Hope this is useful....

SHOOTING WITH RED AND POST WORKFLOW

The New Colour in Town

When I joined the ‘Powder’ team, a TV series from YRF, I was faced with a pretty weird situation. I had to finish reading the 26 episodes of the series (yeah, we had a bound script), do recce, sit on pre-prod over and above the task of reading whatever material was available on the RED One camera. All this in less than two weeks!!

What’s more, I couldn’t ‘buy’ time by way of reading the initial episodes that were to be shot and read the remaining episodes during shoot, because we were shooting by locations instead of by episodes. (For instance, on the third day we were shooting parts of the 26th episode and on the 7th day the opening sequence of the series.) As if all these challenges were not sufficient, I was hearing lots of scary mumbo jumbos about the camera. Luckily I decided to give my instincts a try and just concentrated on reading the story. Actually that was the sanest option available and after 204 days of shoot I don’t regret that decision.

Setting the exposure

Before commencing shoot, I took some pointers and read a bit on how to start shooting with the camera. Things like how to set your exposure, what ASA to shoot etc…

As shooting progressed, I realised that different people were adopting their own methodologies for working with the camera. There were 4 serials that were being shot on RED. And each one of us had a different take with the workings. Even basic thing like the exposure was set differently.

There are different modes of shooting on RED. We can either shoot RAW, or REC 709 or Redspace. I started shooting on RAW for a while, where everything gets recorded in the cameras’ native setting (320 ASA, 5000k). But after a point I realised that I was not going to get too much of time while grading to make it look the way I wanted it. So I tried REC 709 – a color space/ gamma set according to HDMI standards for TV and then dabbled for sometime in Red Space also – the native color space of the camera itself, before deciding on REC 709 as I found it more comfortable with regard to setting the exposure as compared to Redspace.

Anyways back to setting the exposure.
The recommended way of setting the exposure is by the histogram. Go to RAW mode, and then set the aperture in such a way that the peak in the histogram is more to the right than left – ideally in the middle and then you can shoot in a mode of your choice – Raw, or Rec 709 or Redspace.

Then of course we have the different approaches adopted by various members of our fraternity. The camera has an inbuilt meter akin to a centre weighted meter in a still camera which gives the reading in IRE. And one of the cameramen was using both the histogram and this meter and was exposing for a reading of 45-55 IRE in the scale.

Another general method adopted has been to light up as per the camera ASA rating of 320 and expose accordingly.

Give a lazy man a task and he will find the easiest way to do it. I used to set the exposure by eye. After a little bit of research and some friendly tips I found out that as with most digital cameras, what we see in the eyepiece is what we get. For indoors I usually light up for a meter reading of 200 ASA and then set the exposure by eye and that most often matches with the meter reading. If I am shooting on REC 709 mode I set the exposure by eye and open up almost to a third of a stop for the final exposure.

For outdoors it is important to note that when we set the exposure we don’t clip the highlights. The shadow details can be brought up in post by at least 1 ½ stops, but the same doesn’t hold true for highlights. So here a meter reading of the native ASA rating of 320 for RED One works.

So that brings us to the question of ASA rating. The ASA rating or to put it practically the camera’s sensitivity, changes according to the lighting situation and the quality of light used.

For a low contrast setting the RED camera can be compared to a 200 or 250 ASA daylight balanced film stock while in a high contrast setting it behaves like a 320/400 ASA daylight balanced film stock.

Basically, we have to take into consideration what is possible in the post when it comes to trying to understand the usable ASA. During a test I got 49 IRE on a gray card for a rating of 80 ASA, but when viewed on a standard monitor the exposure for a rating of 200 ASA worked as well and in some cases better with regards to the whites.

Another key factor to account for while setting the exposure is the quality of light that is used. A highly diffused source makes you want to open up the exposure compared to even a slightly specular source.

In general the camera is more comfortable with a little bit of contrast when compared to a totally low contrast situation. In a night interior scene – once I tried keeping the character 2 stops below exposure level for a while until he opens a door and gets enough light. It was lit up with a meter rating of 200 ASA and the source was pretty much diffused. In the post I had to face quite a few problems in trying to get some detail and improving the image. In another situation I had lit up using sources which were a mix of diffused and specular light. But due to a error what came out in the end was pretty underexposed. But this time I could bring out the details with a fair amount of ease. It was like the camera could actually sense the specularity in the air.

Work Flow

This is another term which I got hit with pretty regularly before the shoot and even afterwards. But the work flow is definitely one of the most important things to take care of. This is because we might actually end up getting a totally different footage from what we have shot, especially if we are shooting for TV – commercials or serials.

The camera records the images in RAW format (with the native camera settings of 5000k, 320 ASA) and stores the individual settings (for e.g., a color temp of 3200k) as metadata. You further have the option of setting the saturation, contrast, exposure, hue, ASA etc… to suit your aesthetics and all of these are stored as metadata. The advantage of such a system is that in post, we have the option of either working on the RAW files afresh or incorporate our personal settings which might end up saving a lot of time.

For editing purposes we can use transcode the Raw files and use them or just use the proxies that come with the R3D files. The only problem of using the proxies is that they are reference files and not standalone ones; in the sense they need the R3ds to be present where they are and not moved about. So transcoding helps us to use only the quick time files for editing purposes and access the Raw files (which can be stored safely till then) only for conforming to Scratch or Lustre or…

(We can transcode R3D media to ProRes 422 or ProRes 422 (HQ) compressed QuickTime files. ProRes 422 is suitable for creating lower-quality files for offline editing, but ProRes 422 (HQ) is recommended for higher-quality mastering)

One of the common workflows – for transfer to film – transcode to ProRes422, edit it, and link it to corresponding RAW footage (EDL) for grading. So here we can either incorporate the metadata into the raw file to get what we exactly shot or take the raw images and start working on them from scratch (just an English word - not to be confused with SCRATCH – the color correction setup which natively support R3D files and reads the metadata therein)

When it comes to television a ProRes 422 (HQ) can be taken, edited and directly taken for color correction. During transcoding we have the option of taking the native setting of the camera or the settings given by us so that it’s similar to the way we shot it. Sometimes it so happens that if the transcoding is not done properly we end up with footage which is way different from how we shot it. And since going back to the raw files in some cases is impossible the transcoding stage becomes very important and it will do us good to keep the data managers happy because he has enough controls to make even a first degree color correction. It makes sense to spend time with him to work out how the final product is going to look and in between also cross check the footage that has been transcoded.

Some of us have had problems with changing color temperatures, different levels of saturation etc…etera before the problem was spotted and rectified.

Red – a still Camera

When I started using the camera I ran into lots of confusions because I was trying to understand Red One by comparing it to the way a film camera works. I wanted to see how the camera performed in under-lit conditions and so I shot something on 25 fps (almost full open) and then shot the same thing on 50 fps without compensation. Amazingly I got similar levels of exposure on both the footage. I realised that it had to do with the concept of shutter speed like in a still camera in RED One rather than a shutter angle in a film camera (with lapse in time for the movement of the film to give a shutter speed of 1/50 for 25 fps at 180 degrees).

This gave rise to two very interesting options for me while shooting in low light conditions or shooting slow motion.

First and foremost - Till 50 fps I could use regular HMI lights without fear of getting any flicker. Since the main thing while using regular HMI’s was to get a shutter speed of 1/50 to match its frequency, I could happily use HMI’s at 50 fps with a setting of 1/50 shutter speed. This was also the reason why there was no change in the exposure when I shifted from 25 fps to 50 fps. With both the shutter speed and aperture remaining the same, the amount of light hitting the sensor was also the same in both cases.

Secondly whenever the light has fallen short I have opened up the shutter speed to 1/25 instead of the regular 1/50 and thereby gaining a stop in exposures. The 7th schedule of our shoot was a night exterior action sequence. I started lighting up & soon realised that I had to rely mainly on the available street lights if we had to wrap the shoot the same night. Left with no options I tried shooting a passing vehicle with a shutter speed of 1/25 and I could get a pretty decent image without any jitters and acceptable light. And this I have used again and again when faced with similar situations.

In some cases I have used this extra stop of exposure to get more depth of field which has proved invaluable.

We have faced quite a bit of problems while focusing because of back focus issues. With the attendants usually not being technically adept in fixing this we were devising our own methods in overcoming it. Most of the times there was a difference between the tape measurement and the eye judgement. Difference enough, to give a soft image. But the difference used to be consistent – say always off by 2 inches. So we used to take the eye measurement and take it from there. This difference again was not constant. Some days it was spot on and some other days it used to change with different lenses. One bizarre experience was when this difference between the eye judgement and the tape increased to more than 4 inches when I tilted down the camera with a heavy lens on. Sounds like an urban legend and I really hope it was only that.

4K, 3K, 2K

Whether the 2k resolution is good enough for projection can only be tested and verified. But since high speed filming can be done only in 2k in Red One, we can use this setting for getting more out of the available lenses in desperate situations. What happens is when we shift from 4K to 2 K, the image is recorded in a smaller sensor area, for the sake of understanding, similar to 35 mm and 16 mm.

So when we change the resolution the effective focal length changes depending on what resolution we have chosen. So a 50mm lens at 4k 16:9 resolution will behave like an 80 mm lens at 3k 16:9 and like a 108 mm lens at 2k 16:9. I missed this trick when we went to Dubai with a single prime lens. It struck me just after I returned and it hit me really hard.

Another useful factor is that the minimum focusing distance and the depth of field doesn’t change when we shift between the different resolutions - the difference in the focal length being offset by the diminishing sensor area.

Shooting details

Shot on RED One Build 17

Lenses
Red Zooms 18 to 50 mm (this was closer to 40 mm rather than 50 mm)
50 - 150 mm (both these lenses work more like variable primes rather than zooms)
Arri Ultra Primes 20mm, 28 mm, 50 mm and 85 mm.
Angenieux HR 25mm to 250 mm

Recording Media – CF cards of 8 GB (4 mins at 4 k & 9 mins at 2k) and 16 GB capacities.

Disclaimer

"Having the proper vision for each story is what's really interesting. After that, finding the tools & technology is really a very simple continuation of the journey. The technology is only part of the photography, and it is not the most difficult part."- Vittorio Storaro, ASC, AIC


- Karthik Ganesh
Post Reply