Working Resolution: Considerations & Best Practices

INTRODUCTION

In this article, we take an in-depth look at the concept of a working resolution and outline best practices for image scaling throughout the production lifecycle. While not a requirement for Netflix productions, establishing a working resolution can help eliminate image scaling confusion between departments. This can reduce operational friction, resulting in more efficient workflows, a diminished risk of image artifacts, and a simplified path for final delivery and archive.

Netflix is available to assist in navigating production-specific workflow decisions in collaboration with key production stakeholders. Please reach out to your Netflix contacts with any questions or concerns specific to your production.

For a more general overview of these concepts see: Framing & Working Resolution Calculator: User Guide.

WHAT DOES WORKING RESOLUTION DO?

Even on a single camera production, multiple resolutions and/or varying aspect ratios will likely be used throughout the production’s lifecycle. The resolution of Original Camera Files (OCF) will likely differ from the resolution that a production frames for on set. The VFX team might ask for a safety area for image repositioning or tracking effects, and a production’s delivery resolution could differ from the final active image resolution.

The scaling workflow becomes increasingly complicated after image capture when considering editorial media, creative review dailies, VFX pulls, and archival needs.

Establishing a project’s working resolution will help define, clarify, and simplify scaling needs, unifying different resolutions into a single container resolution from which minimal image scaling operations may be performed during the lifecycle of a project. Furthermore, establishing a working resolution allows for a single scaling operation between OCF and final delivery. This single scaling operation reduces the possibility of image artifacts created by uninformed scaling operations in post-production. The working resolution itself helps anyone involved with the image pipeline do their jobs without fear of inconsistent image scaling, multiple VFX pulls, or friction between departments.

Once a project’s working resolution has been established:

Dailies teams will know how to map the image for their deliverables.
Editorial can be confident that all provided media will show the intended active image.
Final Color teams can use a single timeline or project resolution for their work, which easily maps to different render and delivery formats.
VFX teams will receive and return all VFX pulls in a uniform resolution, eliminating any doubt about what is supposed to be seen on screen.

It’s best to think of working resolution as both a number per project and a methodology that can be applied across projects. Discussions around establishing a working resolution should begin during pre-production.

The best way to define the method for establishing a working resolution is to look at real world examples, but first we need to define some fundamental terms.

WORKING RESOLUTION GLOSSARY

The following terms will be used when talking about specific areas of a captured image:

OCF: OCF stands for Original Camera Footage. The resolution of the OCF can also be referred to as the Capture Resolution or Sensor Resolution.

Active Image Area: The part of the OCF that is being framed for on set and ultimately delivered to the viewers at home. This is where the action happens. In some cases, the Active Image Area matches the full OCF, but more often than not, the Active Image Area is cropped in from the OCF.

Delivery Resolution: The resolution of the primary screening deliverable. In Netflix’s case this would typically be a UHD (3840x2160) IMF.

Safety Area: When a production captures their OCF at a resolution higher than their deliverable resolution, they have the option to add a safety margin around the active image area. This safety crop can be helpful for stabilization, reframing and other post-processing later on in the filmmaking process. In the examples further in this article, you will see a 5% safety area applied in most cases. This is also commonly referred to as “Safety Crop” or “Surround View.”

Final Active Image Resolution: The resolution of the active image contained within the delivery resolution. If the active image aspect ratio doesn’t match the final delivery aspect ratio, letterboxing or pillarboxing will be applied.

Now that we have a foundational image mapping vocabulary established, we can demonstrate the method for arriving at a working resolution using examples.

Click header to expand glossary.

START WITH THE FRAMING CHART

Before establishing a working resolution a production should start by examining the framing chart for their primary camera and mapping the individual elements of the image.

For more information on framing charts and how to use them see: Framing Charts Best Practices.

EXAMPLE A

Let’s start with a simple example in which the OCF resolution matches the delivery resolution:

Imagine we are working on a single-camera project that plans to shoot on Panasonic Varicam Pure, UHD sensor mode, for a 1.78:1 output aspect ratio. The studio requires a UHD IMF at 3840x2160 as the final deliverable. In this case the captured resolution in the OCF matches that of the IMF delivery, and thus no scaling or cropping needs to be applied.

Thus the working resolution for this production would be the same as the OCF.

EXAMPLE B

In our next example, the Director of Photography (DOP) has chosen to shoot at a resolution slightly larger than UHD, allowing them to include a safety area.

Our production will still shoot on the Panasonic Varicam Pure, in the 4K DCI sensor mode and the final aspect ratio will be 2.00:1.

This time we will frame for a UHD center crop within the OCF. By doing so, we include a ~6% safety area (4096 x 0.94 ≈ 3840).

To get from our OCF to the final delivery of the IMF, we would center crop to 3840 width and then letterbox to a 2.00:1 aspect ratio. Since we want to have the entire image available in post for reframing and stabilization purposes, the working resolution in this example would include the active image area plus the safety area: 4096 x 2160.

EXAMPLE C

Now that we understand the concept of a safety area, let’s move to a slightly more complex but common scenario.

Imagine we are on a single camera project that plans to shoot on ARRI Alexa LF, in Open Gate sensor mode, for a 2.00:1 output aspect ratio. The DOP has added a 5% safety area for post processing, and the studio requires a UHD IMF at 3840x2160 as the final deliverable.

As stated before, we should first examine our framing chart to ensure we understand the different image areas of our OCF.

The Red line shows the entire captured image of the OCF. This area is the maximum image that will be visible once scaled to the working resolution.
The Yellow line shows the minimum image allowed in the working resolution. In this case it is the active image plus the 5% safety area.
The Blue line shows the active image area. The active image area is the part of the captured image that is intended to end up on screen.
The Cyan line shows the 1.78:1 aspect ratio required for an IMF deliverable.

Now that we understand all of the resolution needs for our production, let’s calculate a working resolution container from which all of our scaling operations can be performed.

ARRIVING AT A WORKING RESOLUTION

VFX pulls, VFX finals, and archival delivery should generally include as much of the captured image as possible. Therefore, the working resolution will be the full OCF scaled to a resolution at which a center crop of the delivery resolution (typically UHD) results in the desired active image area.

Note: In this example, the final active image resolution (3840x1920) would be less than the IMF delivery resolution (3840x2160) due to the 2.00:1 aspect ratio. Thus the active image would be letterboxed in the IMF container.

We’ll start by calculating the width of the working resolution. Because the original framing chart included a 5% safety area from sensor image (Red) to the IMF image (Cyan), we would want to see the IMF width plus the 5% safety area included in the working resolution:

[IMF width] / (1-[safety area percentage])

3840 / 0.95 ≈ 4044

For the height, we’ll use the full OCF’s aspect ratio. By doing so, we’ll achieve the maximum resolution possible for VFX/archival renders. For the Alexa LF, that would be an aspect ratio of approximately 1.44:1. So, the formula becomes:

[working resolution width] / [OCF aspect ratio]

4044 / ~1.44 ≈ 2814

Note: For the sake of simplicity, we’ve used an approximation of the source aspect ratio, in this case 1.44:1. The actual aspect ratio is closer to 4448 / 3096 = 1.4366925… Additionally, as a digital image can’t be made of partial pixels, and as some post processing software doesn’t function properly when dealing with odd numbered pixel counts, our calculator will apply the following rounding logic to the raw results of the calculations: ROUND to the nearest integer. If the outcome of that rounding is an odd number, then ROUND UP to the nearest even number.

Now that we have established our working resolution, when we apply a center crop using our final image resolution, we end up with our intended active image within our IMF delivery resolution.

From this working resolution we can easily scale for each one of our deliverables throughout the lifecycle of the project.

Center crop to 3840x2160 and scale down to HD for editorial rushes (keeping some vertical safety area for the editors to use as they need).
Center crop to active image and scale down to HD or 720p for review dailies (letterboxed to the active image area).
Center crop to active image for our UHD IMF (letterboxed to active image area).
No cropping or scaling to render out VFX pulls and archival renders at the working resolution.

Now that we’ve demonstrated the use for a working resolution, we will take an in depth look at working resolution options and special case scenarios. If you’d rather not take a deeper dive in these concepts, feel free to explore our Framing and Working Resolution Calculator to help simplify the process of choosing a working resolution.

THE MINIMUM WORKING RESOLUTION

On some productions, depending on the camera and resolution of the OCF, the area outside of the intended active image can make up 30% or more of the captured data. In cases like these, key stakeholders might decide that the benefits of including all of that additional image area in the working resolution won’t outweigh the burden of rendering and storing extra data. In such a case, the production could decide to calculate a minimum working resolution instead.

The minimum working resolution is a slightly more cropped image, allowing for lower data rates, while still providing some room in the image for repositioning and stabilization.

If the production from our previous example needs to calculate a minimum working resolution, the formula for width would remain the same as before, because we will still include the 5% safety area:

[IMF width] / (1-[safety area percentage])

3840 / 0.95 ≈ 4044

For the height, instead of including the entire sensor image, we limit ourselves to the 5% safety area. The height formula therefore would look similar to the width:

[IMF height] / (1-[safety area percentage])

1920 / 0.95 ≈ 2022

In this scenario, the working resolution would be:

Calculating a minimum working resolution gives key stakeholders the flexibility to choose whether maximizing their working resolution allows for a more efficient workflow or simply generates unnecessary data. The working resolutions described in this article simply present best practices, not requirements or specifications.

SPECIAL CASE SCENARIOS

WHAT ABOUT SECONDARY CAMERAS?

The same working resolution principles apply when using secondary cameras. Productions should use their primary camera to establish their working resolution. Secondary cameras may capture more or less image data than the primary camera, but because the working resolution is based on the active image deliverable, secondary cameras will fall in line.

The methodology for choosing a working resolution doesn’t have to be rigid and can adapt at various stages of production. If production adds a new camera during principal production, the working resolution can be recalculated to include additional scaling needs.

For example, imagine we are on the same production as our Alexa LF Open Gate example above, but we also have two secondary cameras:

Sony Venice capturing at an OCF resolution of 6054x3192. Just like our primary camera, the DOP framed for a 2.00:1 aspect ratio and included a 5% safety area.
RED Komodo capturing at an OCF resolution of 6144x3240, also framing is for a 2.00:1 aspect ratio. The camera will be mounted to a drone, therefore the DOP isn’t framing for a specific 5% safety area. Instead they have chosen to frame for full sensor width.

While our OCF has different resolutions for the active image area, the deliverable active image area remains the same.

Just like our previous example, we can use the Alexa LF as our primary camera and establish a working resolution:

Because the Alexa was chosen as the primary camera, when we scale the secondary cameras into the working resolution, they will have black bars top and bottom (Venice) or even on all sides (RED).

But a center crop at our delivery active image resolution still results in the intended active image. The active image area and safety areas will still be retained, and all of the different resolutions are unified into a single scaling operation between OCF and final delivery.

The same process would apply if the Venice was chosen as the primary camera; however, instead of retaining black bars, the Alexa LF footage would lose image data when scaled from the OCF to the working resolution.

WHAT ABOUT MULTIPLE PRIMARY CAMERA FORMATS?

Some projects might decide to use two or more "primary capture modes" and will need to find a common denominator resolution. In these cases, there are certain technical considerations that are worth talking through and arriving at a working resolution can take multiple steps. Please reach out to Netflix to discuss your options if choosing this path.

WHAT ABOUT WORKING RESOLUTION FOR ANAMORPHIC CAPTURE?

When working with anamorphic lenses, productions should calculate the working resolution using unsqueezed dimensions. The process is roughly the same as when shooting with spherical lenses, we just need to take a few extra things into account.

As an example, let’s imagine we are on a production that plans to shoot on RED Helium at 8K 6:5 sensor mode. Our production also plans to use anamorphic lenses with a 2x squeeze factor for a 2.00:1 output aspect ratio. The DOP has added a 5% safety area for post processing, and our final delivery resolution will be an IMF at 3840x2160.

To minimize scaling operations, we want to desqueeze while scaling down to our final delivery resolution. For this example, our deliverable will be a UHD IMF at 3840x2160.

After desqueezing the framing chart we can use it to calculate our working resolution.

Typically, anamorphic capture retains more of the active image in the height than the width, which is the opposite of our previous example with the Alexa LF with spherical capture. Therefore in this case, we must begin with the height when calculating our working resolution.

Note: If your OCF aspect ratio (after desqueezing) is wider than the active image aspect ratio, calculate the height first. If the OCF aspect ratio is taller than the active image aspect ratio, calculate the width first.

The desqueezed framing chart includes a 5% safety area from sensor image to the final image. Thus we would want to see the final image height plus the 5% safety area included in the working resolution:

[IMF height] / (1-[safety area percentage])

1920 / 0.95 ≈ 2022

For the width, we’ll use the full OCF’s aspect ratio. This way we’ll achieve the maximum resolution possible for VFX/archival renders. For the RED Helium in this sensor mode, that would be an aspect ratio of 1.20:1. But, when working with anamorphic lenses, we need to account for the squeeze factor of our anamorphic lenses, and thus multiply the aspect ratio with the squeeze factor:

[working resolution height] x [OCF aspect ratio] x [squeeze factor]

2022 x 1.2 x 2 ≈ 4852

Or, to get to the minimum working resolution, we would once again limit ourselves to the 5% safety area:

[IMF width] / (1-[safety area percentage])

3840 / 0.95 ≈ 4044

And the working resolution options would be:

The Framing and Working Resolution Calculator can help calculate working resolution options for lenses with different squeeze factors, and the Anamorphic Minimum Capture Resolution Calculator can assist with planning for anamorphic capture.

In some cases, the VFX vendor may choose to work on the squeezed image. In such situations, the VFX pulls would typically be at source resolution. VFX returns could be either at source resolution (still squeezed) or at the working resolution (desqueezed during the compositing process).

WHAT ABOUT USING THE FULL SENSOR RESOLUTION FOR VFX?

The working resolution will generally provide enough image data for the majority of the work on a production. But, in select cases, some VFX work may require more image data than the working resolution provides. When this happens, simply use the OCF to generate plates at the full, originally captured source resolution.

TRANSLATIONS

العربية‬
Deutsch
Español (España)

Español (Latinoamérica)