DisplayPort Adaptive-Sync: When Fixed Refresh Rate Is Not Enough

Displays are central to nearly every modern computing experience. Whether we are watching a video or playing games, we expect motion to appear clear and continuous. However, when the rate at which a Graphics Processing Unit (GPU) produces frames does not align with the rate at which a display refreshes, visible artifacts such as screen tearing and stutter can occur. Adaptive-Sync addresses this mismatch by allowing the display's refresh rate to follow the timing of frames driven by the GPU.

Understanding the Artifacts that Can Occur without Adaptive-Sync

A conventional display refreshes at a fixed rate. For example, a display operating at 120 Hz begins a new refresh approximately every 8.33 ms, calculated as 1000 divided by 120. The GPU, however, may not produce every frame in exactly 8.33 ms. One frame might take 8 ms to render, another 12 ms, and another 20 ms, causing screen tearing or screen stutter.

• Screen tearing can occur when a new frame is transported while the display is still scanning the previous frame. The resulting image may contain portions of two different frames.
• Screen stutter can occur when a frame is not ready for the next fixed refresh interval. The display may repeat the previous frame, momentarily pausing motion.

Adaptive-Sync reduces these artifacts by allowing the GPU to vary the interval between transmitted frames within the refresh rate range supported by the display.

Fixed Refresh Rate vs Variable Refresh Rate (Adaptive-Sync)

Let's assume the monitor is configured for its maximum refresh rate of 125 Hz, which makes its refresh period set at 8 ms. If the GPU renders frames at intervals of 8 ms, 12 ms, and 20 ms, the monitor continues to refresh every 8 ms, regardless of when the GPU makes each frame available. This creates a mismatch between the GPU and monitor, which produces undesirable visual artifacts.

With Adaptive-Sync, the source can dynamically vary the frame interval based on the application, allowing the display's refresh rate timing to more closely track the GPU's output. For example, an 8 ms frame interval corresponds to a 125 Hz refresh rate, a 12 ms interval corresponds to approximately 83.3 Hz, and a 20 ms interval corresponds to a 50 Hz refresh rate. Synchronizing the GPU and display reduces artifacts while improving the overall experience.

The link rate and lane count do not need to change. See Figure 1 for a comparison of the GPU-to-display timing relationship before and after Adaptive-Sync.

Figure 1: Adaptive-Sync aligns the display refresh interval with the GPU’s frame interval

How DisplayPort Frame Timing Works

All frames have a blanking region and an active video region. The DPTX, such as the GPU, transmits Main Stream Attributes, or MSA, describing the frame. The key MSA timing parameters include:

• HTotal;
• HStart;
• HSync Width;
• VTotal;
• VStart;
• VSync Width;

During fixed timing operation, the DPRX (display) uses the MSA timing parameters to interpret the transmitted video timing and reconstruct the displayed frame.

What Changes with Adaptive-Sync?

Under Adaptive-Sync, the duration between frames may change. The active video resolution can remain the same while the vertical blanking interval is shortened or extended.

Figure 2: Adaptive-Sync varies frame duration by adjusting the vertical blanking interval

Because the actual frame duration is now variable, certain fixed MSA timing parameters can no longer be relied on for the frame's timing and reconstruction.

Before Adaptive-Sync can be used, the DPRX must advertise support for ignoring MSA timing parameters during the discovery phase before link training. Once link training is complete, the DPTX can enable this functionality by setting the MSA_TIMING_PAR_IGNORE_EN bit (bit 7) in DisplayPort Configuration Data (DPCD) address 00107h. After this bit is enabled, the DPRX ignores the applicable MSA timing parameters as defined in the DisplayPort specification. The Adaptive-Sync SDP communicates the information needed to coordinate variable frame timing. The DPRX uses this information, together with the transmitted video stream and MSA, to control its refresh timing.

Adaptive-Sync Secondary Data Packet (SDP)

Modern Adaptive-Sync operation uses an Adaptive-Sync SDP to communicate frame-related timing information. Depending on the SDP version and operating mode, its payload can identify the Adaptive-Sync operating mode and carry information such as:

• Target refresh rate
• Target refresh-rate divider
• Frame duration increase constraints
• Frame duration decrease constraints
• Panel Replay timing information

This allows the source and sink to coordinate variable timing without changing the active frame.

AVT and FAVT Adaptive-Sync Features

More advanced Adaptive-Sync features add greater flexibility.

• Adaptive Video Timing, or AVT, allows the source to vary frame duration dynamically within the capabilities advertised by the sink. The DPTX may increase or decrease the interval between frames while respecting the supported timing boundaries.
• Fixed Average Video Timing, or FAVT, allows individual frame durations to vary while maintaining a specified average frame duration or target refresh rate over time.

Conclusion

The key capability provided by Adaptive-Sync is the ability for the DPTX to vary frame duration while communicating the associated timing information to the DPRX through Adaptive-Sync protocol mechanisms. This enables smooth operation without requiring link-rate or lane count changes.

Cadence has a very mature Verification IP solution. Verification over many different configurations can be used with eDP 1.5, DisplayPort 1.4, DisplayPort 2.1, and USB4 DP tunneling designs, so you can choose the best version for your specific needs.

Learn more about Cadence’s Verification IP solutions.