Electronic music

Samsung accused of cheating on hardware benchmarks – again • The Register

Samsung has again been accused of cheating in benchmark tests to inflate the apparent capabilities of its hardware.

The South Korean titan allegedly unfairly topped Galaxy Note 3 phone benchmarks in 2013, and when faced with similar Galaxy S4 claims in 2018, it settled the case for $13.4 million.

This time, Samsung allegedly falsified the results of its TVs, specifically the S95B QD-OLED LCD and QN95B Neo OLED TVs.

These accusations were made this month by YouTube channel HDTVTest on the S95B, and by review site FlatpanelsHD on the QN95B. The claims boil down to the fact that Samsung used an algorithm to detect when the benchmarking software was running on set and adjust the color and artificially boost the luminance by up to 80% during the test to make the equipment appear more beautiful in reviews.

According to the FlatpanelsHD report, these brightness levels cannot be maintained during normal use without damaging the TV’s backlight panel.

An algorithm to detect and trick benchmarking software is exactly what Samsung has been accused of employing in these earlier examples. The chaebol has never denied or admitted to any wrongdoing, arguing instead that it was under no obligation to tell consumers whether its devices included code that allowed it to get the best credentials.

Samsung’s alleged actions to cheat TV benchmarks were only noticed due to its reliance on industry benchmarking standards, which test HDR TVs by focusing on 10% of the screen. When FlatpanelsHD testers changed their baseline size to 9%, they were able to bypass the alleged algorithm and get what they claimed was an accurate score.

Based on his findings, Rasmus Larsen of FlatpanelsHD wrote that the Samsung TV he tested “has an overly bright picture overall. Also… peak brightness is much lower when tested with a 9% window.

What does Samsung have to say?

Samsung Korea told FlatpanelsHD that it will “provide a software update that will ensure consistent brightness of HDR content across a wider range of window sizes beyond the industry standard.” To cynical minds, it sounds like Sammy making sure he doesn’t get called next time.

When The register contacted Samsung for comment, a representative said today, “Samsung Electronics does not use any algorithms for the purpose of providing specific test results.”

Samsung also told us that its own test, and those of independent third parties, “show that HDR content is displayed accurately across various window sizes, not just 10%”. He also said peak brightness is maintained at similar levels on window sizes below 10% “without damaging the panels.” ®