The black box videos collected from vehicles loaded inside the Sewol ferry’s cargo had indicated time on a corner of the screen. However, the time setting was often incorrect, creating some time gap with real time.

Checking the accurate time of recording was the first step in the long journey of finding clues to the cause of the ferry’s sinking.

Different types of information appeared on the videos was used in the process, because figuring out accurate time setting helps drawing meaningful analytic results by comparing it with the ferry’s Automatic Identification System (AIS) tracking records and testimonies from sailors and passengers.

In its September 2014 reports, the Korea Center for Investigative Journalism (KCIJ)-Newstapa had already unveiled that the time appeared on the CCTVs inside the Sewol ferry was running late by 15 minutes 21 seconds compared to accurate time. In April 2016, it published an article that time setting of the CCTV videos gathered from Incheon Port was also inaccurate, a minute and 17 seconds faster than the actual time.

This time, already-unveiled information like these were utilized to prove that the black box videos’ time setting was inaccurate and to calculate the accurate time.

Black box from a Matiz vehicle parked in the twin deck : (Time on the screen) - (11 minutes and 6 seconds) = (Actual Time)

The black box collected from a vehicle loaded in the Deck C twin deck contained a video, which shows a scene ahead, recorded while driving into the Sewol ferry’s ramp from Incheon Port.

KCIJ-Newstapa analyzed this video and a CCTV video shot from Incheon Port, and identified this vehicle as a yellow-green Matiz sub-compact car.

Both videos included one same scene, which captured the moment when the Matiz was driving through the ferry’s ramp and the internal space.

When compared the two scenes and time settings appeared on the screens, Newstapa confirmed that the time setting on the black box was 11 minutes and 6 seconds faster than actual time.

Newstapa went back to the scene when the ferry started tilting, after reflecting the time gap.

Vehicles started leaning sharply to one side at 9:00:49 a.m. according to the black box screen. When applying the time difference of 11 minutes and 6 seconds, the calculation ended up with a conclusion that numbers of cars in front actually started sliding at 8:49:43 a.m. on April 16.

Black box from a 1-ton truck parked near the Deck C’s starboard : (Time on the screen) + (1 hour 21 minutes) = (Actual Time)

KCIJ-Newstapa carried out the same calculation method for the 1-ton truck loaded inside the Deck C near the starboard.

This truck’s black box also had its time set inaccurately, an hour and 21 minutes slower than actual time.

Using the time difference, Newstapa succeeded re-calculating correct time for crucial scenes.

This truck started inclining towards the right wall at 8:49:44 a.m., and the metal box -- which was sitting on another truck in the front left -- fell at 10:14:17 a.m.

A sub-compact vehicle, which was initially parked in the front left, fell off the floor to hit the ceiling at 10:16:07 a.m. Lastly, seawater started to flow in from back of the twin deck at 10:17:19 a.m.

Black box from a 1-ton truck parked near the Deck C’s port : (Time on the screen) + (43 seconds) = (Actual Time)

KCIJ-Newstapa carried out the same calculation procedure for another 1-ton truck parked right next to the left wall.

This truck’s black box also had an inaccurate time setting, 43 seconds slower than actual time.

As the ferry tilted to the left, the truck (which was shooting the video) was also pushed to hit the left wall at 8:49:49 a.m.

Black box from a Starex van parked in the Deck C’s center near the bow : (Time on the screen) + (35 seconds) = (Actual Time)

The last black box was collected from a Starex van parked in the Deck C’s central part, near the bow.

However, KCIJ-Newstapa was unable to calculate accurate time using the same method as it did to other black boxes, because the forensics company failed to recover the scene of the van driving into the ferry.

So, this time, Newstapa tried a sound analysis to compare one certain sound in this video with the same sound recorded in other videos.

The Starex van was parked in the ferry’s central line. Same with the Matiz.

Therefore, Newstapa supposed that both Starex and Matiz vehicles must had been slid to the left almost simultaneously, creating colliding sound at the same moment.

First step of the sound analysis was to look for colliding sounds from both black box videos.

In both videos, high-frequency sound -- of course inaudible to human ears -- were detected twice at similar points of time. The intervals between two sounds appeared to be identical as well.

By matching the two videos by the moment when the colliding sound was detected, Newstapa was able to find accurate time setting for both.

Then, based on the time appeared on the Matiz’s black box, it succeeded calculating an accurate time for the Starex van’s.

The time of recording remained in the Starex video file’s name, to be 8:48:58 a.m. of April 16.

The sound test proved that the time was inaccurate, 35 seconds slower than actual time. The actual moment that vehicles started shifting to the left was 8:49:33 a.m.

By carrying out video and sound analyses, Newstapa succeeded in finding out the actual moment when the Sewol Ferry started tilting and when the cargos in the Deck C and twin deck above began sliding off.

This also allowed careful examination and analyses on what was really going on at 4 different spots inside the Deck C.