Flagship Mobile Game Experience: An In-depth Analysis of Frame Interpolation Techniques

Flagship Mobile Game Experience: An In-depth Analysis of Frame Interpolation TechniquesIn recent years, flagship chipsets have consistently delivered excellent performance, with few exceptions. Qualcomm, leveraging its self-developed cores, and MediaTek, aggressively promoting its full-large-core architecture, have both achieved year-on-year increases in GPU performance while steadily improving energy efficiency

Flagship Mobile Game Experience: An In-depth Analysis of Frame Interpolation Techniques

• In recent years, flagship chipsets have consistently delivered excellent performance, with few exceptions. Qualcomm, leveraging its self-developed cores, and MediaTek, aggressively promoting its full-large-core architecture, have both achieved year-on-year increases in GPU performance while steadily improving energy efficiency. Flagship smartphones have also seen significant upgrades. While some models compromise on camera capabilities, 2K resolution and 120Hz refresh rates are now standard, and various cooling technologiesincluding larger vapor chambers (doubled in size in just two years), air cooling, liquid cooling, and even liquid metalare being implemented to maximize performance. However, the actual gaming experience often falls short of the marketing hype. Even flagship phones may struggle to maintain a 2K resolution in demanding games, with some titles only supporting 864p, failing to reach 120fps, or even being locked at 30fps. Furthermore, issues like frame drops and overheating persist during extended gameplay, resulting in fluctuating frame rates, particularly noticeable in games like Tower of Fantasy and Shining Nikki. This discrepancy between advertised flagship performance and real-world experience is stark.

To enhance gaming experiences, some gaming-focused phones employ dedicated display chips to assist in processing game visuals, enabling frame interpolation and upscaling to boost frame rates and reduce power consumption. This seemingly energy-inefficient approach is becoming increasingly popular. However, OnePlus's recent marketing of the Ace 5 series, with its "Wind Chaser Game Engine" (an internal frame interpolation feature) and CEO Li Jie's assertion that "more manufacturers will abandon external dedicated display chips in favor of GPU-native super-frame solutions," has sparked considerable debate in the tech community. Beyond the hyperbole, a key question remains: if superior internal frame interpolation solutions exist, why do some manufacturers still rely on seemingly inferior external frame interpolation methods using dedicated display chips? What are the crucial differences between internal and external frame interpolation? Is the relationship truly one of replacement, as suggested by the OnePlus executive? And what role do external chips now play? These questions demand thorough investigation.

A Detailed Comparison of Internal and External Frame Interpolation Techniques

Before comparing internal and external frame interpolation, understanding their fundamental differences is crucial. "External frame interpolation," typically found in phones with dedicated display chips, essentially utilizes MEMC (Motion Estimation, Motion Compensation) technology, already prevalent in smart TVs. The principle is straightforward: a dedicated display chip analyzes the original frames of a video or game, identifies motion changes between frames, and generates new intermediate frames to increase the frame rate, resulting in smoother visuals. For example, if a car moves from the left to the right of the screen in a 60fps video, the chip predicts its trajectory and generates a frame showing the car in the middle, creating a more seamless transition. By inserting one or two frames between each pair of original frames, the game's frame rate can be increased from 60fps to 72fps or even 144fps.

Furthermore, the dedicated display chip can offload rendering pressure from the GPU. For instance, if the GPU renders at 46fps, the dedicated chip can upscale this to 72fps, reducing GPU load, power consumption, and heat generation. However, "external frame interpolation" suffers from inherent limitations, primarily in image quality and input lag. MEMC divides each frame into multiple pixel blocks, predicting and interpolating pixels, which can effectively handle fast-moving objects in high-speed motion scenes, but it can also lead to artifacts like mosaic, color banding, or ghosting. More critically, the post-processing nature of MEMC introduces noticeable input lag, a fatal flaw for games requiring real-time responsiveness. The lack of post-processing in the game mode of any TV highlights the significant input lag issue with MEMC.

Unlike "external frame interpolation," "internal frame interpolation" is not a novel concept. Similar to Nvidia's DLSS 3 frame generation or AMD's FSR 3, it leverages the GPU to utilize motion vectors and depth information from the rendered frames for frame prediction. Compared to MEMC, which relies solely on analyzing image changes to infer motion, "internal frame interpolation" offers higher accuracy and better results. On powerful Nvidia RTX 40-series GPUs, it can increase frame rates up to four times the original rate. In smartphones, increasing 60fps to 120fps provides a noticeable improvement in gaming experience. However, "internal frame interpolation" also has drawbacks. While input lag is lower than with "external frame interpolation," it's still a factor. AMD recommends running games at 60+ FPS before enabling FSR 3 frame generation to minimize latency. More significantly, "internal frame interpolation" requires per-game adaptation by both chipset and game developers. Currently, the number of games supporting DLSS 3 is less than a hundred, and even fewer support Qualcomm's "internal rendering" technology (around ten). Therefore, gamers cannot be certain whether their favorite games will support "internal frame interpolation."

Actual Performance Testing and Comparison of Internal and External Frame Interpolation

Let's move on to practical testing. Starting with the Snapdragon 8 Gen 3 and Dimensity 9300, many flagship chipsets now incorporate frame interpolation and upscaling features. However, major manufacturers haven't fully utilized these capabilities, a point that will be clarified by the test results. Even when theoretically supporting these features, manufacturers employ vastly different strategies; Xiaomi products completely lack internal frame interpolation support, while other brands (e.g., Oppo's "Black Factory") mix external and internal interpolation, while vivo remains committed to external frame interpolation.

This test uses the iQOO Z9 Turbo+, equipped with the Dimensity 9300+ processor and a self-developed esports chip (Q1), and the Realme Neo7, which also uses the Dimensity 9300+ but with a different frame interpolation solution. Data from the OnePlus Ace 5 is also included for comparison, highlighting differences in frame interpolation implementation between Qualcomm and MediaTek.

• Games tested included popular, moderately popular, and less common titles for comprehensive analysis. The Realme Neo7 only supports upscaling/frame interpolation in Zenless Zone Zero and Honkai: Star Rail, while the OnePlus Ace 5 supports these features, along with HDR processing, in Genshin Impact, Honkai: Star Rail, and others. The iQOO Z9 Turbo+ supports frame interpolation in five games and simultaneous upscaling and frame interpolation in Genshin Impact, Honor of Kings, and others. Notably, iQOO's compatibility list isn't strictly enforced; users can bypass the official whitelist via software to enable upscaling and frame interpolation in any application or game.

• Using Zenless Zone Zero (compatible with both phones), we tested input lag and image quality with frame interpolation enabled. Using 960fps slow-motion recording, we measured the time between finger tap and character response. Realme showed slightly lower input lag, potentially because "internal frame interpolation" allows the GPU-rendered frame to receive input commands; iQOO's generated frames might not receive these commands seamlessly, but the lag wasn't excessive. In terms of image quality, iQOO's "external frame interpolation" exhibited fewer visual errors than Realme's "internal frame interpolation," but slight blurring occurred during fast character movement, a persistent limitation of "external frame interpolation." Currently, the gap in image quality between "internal" and "external" frame interpolation isn't significant.

Concerning actual frame rates, iQOO Z9 Turbo+ demonstrated more stable frame rates than Realme Neo7, with lower power consumption and a better overall experience. With frame interpolation enabled, iQOO Z9 Turbo+ achieved 144fps (native 48fps + 1 insert 2), maintaining stable frame rates and reducing power consumption by nearly 1/4 compared to native 60fps, along with a temperature decrease. The Realme Neo7 interpolated to 90fps (native 45fps + 1 insert 1), but exited super frame rate mode after 10 minutes due to overheating, reaching a peak power consumption of nearly 9W. The Realme Neo7's 30fps to 60fps interpolation mode might be more practical.

• In Genshin Impact, iQOO Z9 Turbo+ achieved 1.5K + 72fps upscaling and frame interpolation (native 36fps + 1 insert 1). With both upscaling and frame interpolation enabled, power consumption was similar to native 60fps operation, though temperature was slightly higher. OnePlus Ace 5 achieved 900p + 120fps upscaling and frame interpolation (native 60fps + 1 insert 1), with stable frame rates averaging around 118.4fps; power consumption increased from around 4W at native 60fps to approximately 6W.

The results demonstrate that with sufficient chipset power, "internal frame interpolation" delivers a superior gaming experience, making "external frame interpolation" appear relatively inadequate. Further tests on less popular games only supported by the iQOO Z9 Turbo+ showed that enabling frame interpolation reduced power consumption by approximately 1/4, while enabling both upscaling and frame interpolation resulted in power consumption comparable to native (The text abruptly ends here).