A game can render one frame in 8 milliseconds and the next in 14. A conventional display running at a fixed refresh rate does not adjust its scan timing around those changes. The graphics source and the display therefore operate on separate schedules, and the mismatch can appear as tearing or uneven motion.
Variable refresh rate, commonly shortened to VRR, changes that relationship. Within a supported operating range, a compatible display can vary the interval between refreshes so that new frames are presented closer to the time the source finishes them. The display is still refreshing one complete image after another; the timing between those refreshes is what changes.
Fixed refresh creates a timing boundary
A 60 Hz fixed-refresh display begins a new refresh about every 16.7 milliseconds. At 120 Hz, the interval is about 8.3 milliseconds. These intervals describe the display cadence, not a guarantee that an application or graphics processor will produce a new frame at the same cadence.
If a new frame is not ready at the required boundary, a synchronized system may present the previous frame for another refresh. The resulting sequence can contain uneven frame durations. If synchronization is disabled and the source changes the framebuffer while a scan is in progress, portions of different rendered frames can appear in one displayed refresh. That visible discontinuity is screen tearing.
Traditional vertical synchronization, often called V-Sync, prevents that particular form of tearing by coordinating presentation with fixed refresh boundaries. It does not make variable rendering time disappear. A late frame can still miss a boundary and remain on screen longer than intended.
VRR changes the boundary itself. Instead of requiring every completed frame to fit a rigid display cadence, the display can delay the next refresh within its supported timing limits.
VRR follows frame delivery inside a defined range
A variable-refresh display does not have unlimited timing freedom. It operates across a range supported by the display, connection, source hardware, and active mode. Resolution, interface configuration, firmware, and device settings can affect the available range.
Suppose a display mode supports variable operation across a particular interval. When a game produces frames at rates inside that interval, the source can coordinate refresh timing with completed frames rather than forcing them onto a fixed cadence. Frame intervals may still vary, but the display cadence can track those variations more closely.
This is distinct from making the graphics processor faster. VRR does not shorten the rendering work required for a complex scene. It changes presentation timing after frames become available.
The maximum refresh rate also remains a real limit. A display that can refresh at up to 120 Hz cannot present 180 unique full frames each second in that mode merely because VRR is enabled. At the other end, sufficiently slow frame delivery can fall below the display’s native variable-refresh range.
Low frame rates need additional handling
When rendering drops below the minimum VRR rate, implementations can use repeated refreshes to keep the display operating within a valid timing range. This behavior is often called low frame rate compensation.
The principle is straightforward. A rendered frame that arrives at 40 frames per second has a 25 millisecond frame interval. If that interval is too long for a display’s variable-refresh window, the same rendered frame can be scanned more than once at a higher refresh cadence. Repeating it twice would correspond to an 80 Hz display cadence while the game itself still produces 40 unique frames per second.
This does not create extra animation detail. The repeated scans contain the same rendered image. The technique keeps display timing in a usable range and can preserve VRR behavior when the source frame rate is low.
Support and activation conditions vary among display systems, so the presence of a VRR label alone does not establish how low-rate handling behaves in every mode.
Refresh rate and frame rate remain different quantities
Refresh rate describes how often the display scans or updates an image. Frame rate describes how often the source produces distinct rendered frames. VRR coordinates these quantities, but it does not make them identical concepts.
A display may refresh the same rendered frame multiple times. A source can also produce frames faster than the display’s maximum refresh capability. At a fixed 60 Hz, a game reporting 120 frames per second can be producing two rendered frames during each nominal display interval, but the screen cannot show 120 complete sequential refreshes each second in that 60 Hz mode.
This distinction also matters outside games. Video commonly has a stable content frame rate, while desktop interfaces and games can produce less regular frame timing. A display technology designed for variable timing can support several use cases, but the source, operating system, application, connection, and display all participate in the final presentation path.
VRR does not remove every source of latency or stutter
Coordinating refresh timing addresses a specific mismatch between frame completion and display refresh. Other delays remain elsewhere in the system.
Input processing, application logic, CPU work, GPU rendering, frame queues, video transport, display processing, and pixel response can all contribute to the time between an input and its visible result. VRR does not erase those stages.
Likewise, a long rendering stall remains a long rendering stall. If an application pauses while compiling work, loading data, waiting on another resource, or completing an unusually expensive frame, variable refresh cannot turn that delay into a normal frame interval. It can make presentation cleaner around variable delivery, but it cannot supply a frame that has not been rendered.
Pixel response is separate as well. A high or variable refresh rate describes scan timing; it does not state how quickly individual pixels transition between values. Display certification programs can test several properties together, but the terms themselves refer to different behavior.
The connection path has to support the mode
VRR is a coordinated feature rather than a property that a monitor can apply independently to any incoming signal. The source and display need compatible signaling, and intermediate equipment can matter.
On PCs and monitors, variable-refresh operation can be carried through display interfaces that support suitable adaptive timing. HDMI also defines VRR for compatible sources and displays. Product branding can add another layer, since vendor programs may describe compatibility, validation, or operating ranges beyond the base transport capability.
A capable display connected to an incompatible source will generally fall back to fixed-refresh behavior. The same can happen when an active resolution, refresh mode, adapter, dock, receiver, or software setting does not expose VRR end to end.
This makes specification labels most useful when read as a chain. The display panel, input port, source output, cable path, graphics stack, and selected mode all have to support the intended configuration.
A higher maximum rate and VRR solve related but separate problems
Moving from 60 Hz to 120 Hz halves the nominal fixed refresh interval from about 16.7 milliseconds to about 8.3 milliseconds. That gives the display more frequent opportunities to present new frames and can improve motion sampling when the source supplies enough frames.
VRR addresses variation between those opportunities. A 120 Hz display running at a fixed rate still has fixed presentation boundaries. A 120 Hz VRR mode can move refresh timing within its supported range to follow irregular frame completion.
For that reason, maximum refresh rate and variable-refresh capability should not be treated as substitutes. One describes the fastest supported refresh cadence in a mode; the other describes whether cadence can vary with frame delivery.
The practical result is easiest to see as a timing relationship. A graphics processor produces frames when its work is complete, while a display has limits on when it can begin each refresh. VRR gives those two schedules room to meet instead of forcing every frame onto a single fixed clock. It cannot repair every rendering delay, but it can remove the fixed-cadence mismatch that produces tearing and some forms of uneven presentation.