How to Build A/B Monitoring Setups for Color QC

The simplest rule for A/B monitoring is this: compare the thing you're trying to judge, not the path that accidentally created it. If you want to compare a grade before and after a trim pass, both images need to travel through the same output transform, same video I/O, same display class, and same room conditions. If you want to compare SDR and HDR, you need two intentional output paths with known targets, not one “real” output and one GUI viewer you hope is close enough. A/B setups fail when the two sides aren't actually comparable. One monitor is warmed up and calibrated, while the other is a client TV. One signal is coming from a decklink output, while the other is coming from the OS desktop. One view has the show LUT applied, while the other has a timeline transform and a display profile fighting each other. You may still learn something from that setup, but you shouldn't use it for color QC decisions. A good A/B monitoring setup is boring because it routes predictable signals to predictable displays and makes it easy to switch between versions without changing the viewing conditions. The exact build depends on whether you're checking creative changes, matching shots, reviewing dailies, comparing SDR and HDR trims, or validating deliverables.

Decide what kind of comparison you're building

Before choosing monitors or routers, define the comparison. Most post teams use “A/B” to mean several different things, and each one needs a slightly different setup. Common A/B use cases include:

Before and after a grade change on the same shot
Shot A versus shot B for sequence matching
Camera original or log image versus color-managed viewing transform
SDR trim versus HDR master
Two display paths fed by the same signal to catch monitoring problems

Those aren't interchangeable. A before/after grade check can usually happen inside the grading system with a single reference display, as long as the system toggles versions without changing the output pipeline. SDR/HDR comparison needs two output transforms and two displays that can represent their targets. A display-path diagnostic needs the same signal feeding two monitors, because the point is to see whether the monitoring chain itself is behaving.

Comparison goal	Better A/B strategy	Display setup	Main risk to control
Grade before and after	Toggle versions through the same reference output	One calibrated reference display is usually enough	Changing the output transform or viewing path while judging the grade
Shot A versus shot B	Compare shots in sequence and, when useful, side by side	Same output path and same display conditions	Matching still frames while missing motion or context problems
Log or camera original versus managed view	Generate intentional source and viewing-transform states	Reference output with known color management	Confusing a technical transform difference with a creative grade difference
SDR trim versus HDR master	Generate two target-specific outputs	Separate calibrated SDR and HDR displays, each set to its delivery target	Expecting SDR and HDR to look identical instead of preserving intent within each target
Monitoring-path diagnostic	Split the same signal to both displays	Matched displays and identical input formats	Mistaking a display, range, or room mismatch for an image problem

Start by naming the two sides in plain language. “A is Rec.709 Gamma 2.4 with show LUT v12. B is Rec.709 Gamma 2.4 with show LUT v13.” Or, “A is HDR PQ 1000 nit master. B is SDR Rec.709 trim derived from the same timeline.” That sentence becomes the guardrail for every routing and calibration decision that follows.

Keep the GUI out of color-critical comparison

For color QC, the viewer inside your NLE or grading app isn't the same thing as a reference output. The operating system, GPU settings, display profiles, scaling, HDR desktop modes, browser behavior, and whatever else is happening on that workstation affect the GUI. The cleaner setup is:

Application timeline or source monitor
Dedicated video I/O device
SDI or HDMI signal path
Router, distribution amp, or monitor input selection
Calibrated reference display
Scopes fed from the same intended signal path

Your team can apply this principle without the most expensive HDR mastering monitor on the market: keep the color-critical path separate from the computer desktop. For SDR editorial QC, a calibrated broadcast display or a well-controlled non-reference display may be enough. For HDR mastering or Dolby Vision work, you need a professional HDR reference monitor that meets the delivery target’s requirements. The practical distinction is whether someone will make an approval decision from the image. If yes, use the reference path. If the second screen is only for navigation, notes, bins, or client comfort, label it mentally as non-critical and don't let it settle arguments about color.

Route dual outputs based on the comparison type

There are two basic ways to build an A/B setup: split one signal, or generate two different signals.

A/B monitoring starts with routing: either split one known signal or generate two intentional comparison outputs.

Splitting one signal is best when you're testing displays or viewing environments. A router or distribution amplifier sends the same SDI signal to two monitors, which is useful when you want to know whether two displays match, whether one input path is wrong, or whether a client display is drifting away from the suite monitor. Generating two signals is best when A and B are genuinely different images. For example, a grading system may output HDR on one SDI path and SDR on another, or output clean feed A and clean feed B from separate timelines, versions, or transforms. This is more useful for side-by-side creative and deliverable comparison, but it also creates more ways to make a mistake. Typical routing options look like this:

Single video output into a router, then routed to monitor A or monitor B
Single video output into a distribution amplifier, feeding two displays simultaneously
Dual video outputs from the workstation or I/O device, each carrying a different transform
One output to a reference monitor and one output to scopes
Router presets that recall “A only,” “B only,” and “A/B side-by-side” room states

For pure QC, switching is often better than side-by-side. The human eye is very sensitive to relative differences when two images are adjacent, but it can also overreact to small differences that aren't meaningful in normal viewing. Side-by-side is helpful for diagnosing and explaining, while full-screen A/B toggling is better for judging whether the change actually improves the shot. If you do use side-by-side, make sure the software isn't scaling, letterboxing, tone mapping, or compositing one side differently than the other. A 50/50 split view can be great for client communication, but it's only color-critical if the software processes both sides through the intended transforms.

Match two displays before trusting the comparison

Two monitors can measure correctly and still not look identical. Different panel technologies have different spectral characteristics, and that can create metameric differences where viewers perceive two displays with similar measured values differently. This is especially relevant when comparing OLED, LCD, QD-OLED, mini-LED, and consumer HDR displays. For A/B color QC, your team should use matched display models whenever possible. Same manufacturer, same model, same size, same firmware, same calibration target, similar age, similar hours. That doesn't guarantee a perfect perceptual match, but it removes many variables.

Matched displays make A/B differences easier to trust and easier to trace.

When matching displays, define the target in workflow language, not just monitor-menu language. For SDR mastering, that often means Rec.709 with Gamma 2.4 in a controlled dim environment. For HDR, the target may be PQ with the required peak luminance and color space for the delivery format. For dailies, the target should follow the project’s color management plan, including whether the workflow applies CDLs and show LUTs. Your team should account for:

Warm-up time before calibration or judgment
Identical input signal format, such as resolution, frame rate, range, and chroma sampling
Matching monitor picture modes and disabled consumer enhancements
Calibration to the same white point, luminance, gamma or EOTF, and color space
Consistent probe or measurement device where possible
Consistent measurement conditions and calibration records
Perceptual matching when two displays measure correctly but still look different
Controlled room lighting and neutral surroundings

The measurement discipline matters. Color measurement guidance across industries repeats the same point: your results are only as good as the measurement conditions. Your team needs to calibrate instruments, hold them consistently, and use them under defined conditions. In display work, the equivalent is using the same probe strategy, the same target, the same patch generation path, and a room setup that doesn't change between A and B. The takeaway is to use both measurement and perceptual evaluation. Measure first so the displays are technically close, and then perceptually evaluate whether they match well enough for the decision you're asking people to make.

Treat the room as part of the A/B system

The room is part of the monitoring chain, along with signal routing. If monitor A is next to a bright wall and monitor B is framed by black acoustic treatment, viewers won't perceive them the same way. If one display is angled toward the colorist and the other is angled toward a client couch, they aren't equivalent. For a serious comparison, place the displays so the viewer can evaluate them from the same seated position, at similar distance and angle. Keep surround lighting stable. Avoid colored walls, colored bias lights, bright UI panels, and windows that change throughout the day. Even if your room isn't a purpose-built grading theater, consistency beats improvisation. This is especially important for assistant editors and post supervisors doing dailies or ingest QC. The goal may not be final color approval, but you still need a controlled enough setup to catch technical problems early. If an issue is visible only on one random office monitor under daylight, that's a clue, not a verdict. Move the question to the controlled viewing path before escalating.

Use A/B for before and after grading checks

Before/after comparison is one of the most useful A/B workflows because it keeps creative decisions honest. A grade can look better after ten minutes of staring at it simply because your eyes adapted. A fast toggle back to the previous version tells you whether you improved the image or just got used to the new one.

Toggling between versions helps separate real grading improvements from visual adaptation.

Useful before/after comparisons include:

Base grade versus current grade
LUT v1 versus LUT v2
Shot match before and after balancing
Noise reduction off versus on
Legalization or gamut mapping off versus on

The key is to isolate the variable. If you're judging a LUT change, don't also change exposure, timeline color management, and output transform. If you're reviewing noise reduction, compare it at full resolution on the reference output, not in a scaled GUI viewer. If you're matching two shots, compare them in sequence context as well as side-by-side, because a shot that matches as a still can still feel wrong in motion. For dailies, the same principle applies earlier in the workflow. Your color management plan should define whether your team uses CDLs and LUTs and how they should reproduce downstream. Primary corrections and show LUTs are easier to carry through the pipeline reliably than complex secondary corrections. An A/B setup lets assistants and dailies colorists compare camera intent, CDL/LUT application, and editorial proxies before problems become finishing surprises.

Compare SDR and HDR without pretending they're the same image

SDR and HDR side-by-side monitoring is valuable, but it's easy to misunderstand. You aren't asking two displays to show the same thing. You're asking each display to show the correct rendering for its target.

SDR and HDR outputs can be compared side by side while still respecting their different targets.

A typical HDR/SDR comparison may involve:

HDR master output to a PQ HDR reference display
SDR trim output to a Rec.709 Gamma 2.4 reference display
Scopes that can evaluate PQ, Rec.709, gamut, and levels
Color management that explicitly defines the HDR-to-SDR relationship
Matching display technology where possible to reduce perceptual mismatch
Room conditions appropriate for the mastering target

Dolby Vision guidance specifically warns that monitors with different display technologies can exhibit metameric issues and recommends using the same display technology for HDR and SDR targets when possible. That recommendation matters in real rooms. If your HDR monitor is OLED and your SDR display is an LCD, some differences may come from the trim, but some may come from the panel technologies. Don't mistake every perceived mismatch for a creative error. Your team shouldn't judge the SDR image by whether it has the same brightness impression as HDR. It can't, so judge whether the SDR trim preserves intent within SDR constraints: faces, contrast relationships, saturation, highlight handling, shadow detail, and scene-to-scene continuity. Judge the HDR image against the HDR target. Judge the SDR image against the SDR target. The A/B setup helps you understand the translation, not collapse both deliverables into one look.

Watch for common failure modes

Most A/B problems come from one side of the comparison being processed differently, measured differently, or viewed differently. The most common failure modes are:

One display is calibrated and the other is only factory-set
One path is full range and the other is legal/video range
One path is being color managed by the OS and the other bypasses it
HDR metadata is present on one path but absent or misread on the other
A consumer TV is applying dynamic contrast, motion smoothing, local dimming changes, or color enhancement

The fix is to trace the signal in order. Source, timeline color management, version or node state, output transform, I/O format, router, monitor input, monitor mode, calibration target, scopes. Don't start by nudging the grade to satisfy a mismatched display. First prove that the two sides are supposed to look different.

Document the monitoring state so the team can repeat it

A/B monitoring only helps if your team can recall the setup. Post teams change rooms, swap editors, colorists, and sound mixers, update software, and move between dailies, editorial, finishing, and review. If the comparison setup lives only in one person’s head, it will drift. Your documentation should capture the details that change the image:

Project color management settings
Timeline working space and output color space
Show LUT, CDL, display LUT, and output transform versions
Video I/O device and output format
Router presets and monitor inputs
Monitor model, serial number, firmware, calibration date, and target
Probe or calibration system used
Room lighting assumptions
Scope connection point
Known limitations, such as unmatched display technologies or non-reference client screens

This is also where production-to-post communication matters. Media often arrives in post after set or dailies teams have already made decisions: CDLs, LUTs, camera color spaces, viewing transforms, and documentation from the DIT. Assistant editors and post teams need to know how those teams made those choices so they can reproduce or intentionally replace them. A/B monitoring is much more useful when it's tied to that color management plan instead of your team treating it as a last-minute room setup.

Build for the decision you actually need

If you're doing editorial QC, build a simple, repeatable reference path and use A/B to catch obvious color-management or proxy errors. If you're doing finishing, keep the GUI out of the color-critical chain and make version toggling fast. If you're matching two displays, split the same signal first and solve the monitoring mismatch before judging creative differences. If you're comparing SDR and HDR, generate two intentional outputs and judge each against its own target. The best A/B monitoring setups reduce doubt. They make it clear whether a difference is in the grade, the transform, the display, or the room. Once the team trusts the comparison, review notes are easier to tie to a specific part of the pipeline.

FAQ

For color-critical decisions, use a dedicated reference output instead of the application GUI viewer. The GUI can be affected by the operating system, GPU settings, desktop HDR modes, display profiles, scaling, and other variables. It's fine for navigation and editorial work, but approvals should be made from a known video I/O path feeding a calibrated display.

Split one signal when you're testing the monitoring chain, such as checking whether two displays match or whether a router path is behaving correctly. Generate two separate outputs when A and B are intentionally different images, such as HDR and SDR trims, two versions of a grade, or two output transforms. The routing choice should match the question you're trying to answer.

Side-by-side viewing is useful for diagnosing differences and explaining changes to a team, but full-screen toggling is often better for judging creative impact. Adjacent images can exaggerate small differences that may not matter in normal playback. If side-by-side is used for QC, confirm that both sides are being scaled, transformed, and displayed consistently.

No. Two displays can measure close to the same target and still appear different because of panel technology, spectral output, age, firmware, viewing angle, or metameric differences. Matched monitor models with the same calibration target are preferred. Measurement should come first, followed by a perceptual check to confirm the match is suitable for the decision being made.

SDR and HDR should be treated as different target renderings, not as identical images. The HDR output should be judged on an HDR reference display using the required HDR target, such as PQ. The SDR trim should be judged on a Rec.709 Gamma 2.4 display or the project’s defined SDR target. The comparison is about preserving creative intent across formats, not making SDR look as bright as HDR.

Monitoring notes become much more useful when the project records which transform, target, and version were used for a given review. Aspect lets teams define fields for those workflow details and keep them searchable as custom metadata.