Actual Sony recovery cases
Reconstruction examples from real Sony camera media.
How Anya's Sony ZV-E1 wedding video was recovered from a reformatted SD card
Anya brought us a reformatted 128 GB SD card looking for wedding footage from August 8. The original filesystem no longer told us which recording data belonged to which clips. Hex-level examination identified a Sony ZV-E1 and a surviving KLV-based stream. That gave us a real recording structure to work from instead of relying on filenames that were already gone.
First step: identify what survived
The card still contained enough Sony recording structure to identify the ZV-E1 and a KLV-based stream. That proved useful video data remained after the reformat and gave us something concrete to trace.
Next step: the first recovery method
This case became the starting point for Version 1 of Sony XAVC Recovery. The surviving Sony recording structure was used to reconstruct approximately 80 GB of video data. With the original filesystem metadata gone, the remaining job was to determine which recovered recordings were the wedding.
Final step: find the wedding
The recovered recordings were correlated back to recording dates. Anya identified August 8, 2026 as the wedding footage she needed and August 22 as the secondary date. After reviewing the output, she confirmed those were the only recovered dates required for final delivery.
The August 8 wedding footage was recovered from the reformatted card.
A wedding video is not just 80 GB of recovered data.
The technical problem was rebuilding useful recordings after the filesystem context was gone. The useful result was much simpler: the video from the wedding date could be opened, reviewed and returned again.
The full case study shows one recovered wedding recording playing in VLC. The people are blurred for privacy, while the codec information remains visible as technical proof.
How Sony XAVC Recovery added working XAVC HS HEVC recovery
Jason had missing Sony FX3 XAVC HS footage from September 9. The first useful build could find the recording and produce recognizable preview frames. It proved the video was still there, but it was nowhere near a finished recovery. After each test, Jason returned the diagnostic bundle so the next build could target the actual failure instead of another guess.
First build: found it
The scanner identified the Sony structures and recovered recognizable frames from the missing footage. That gave us a real HEVC recording to trace instead of a theory.
Next builds: close, but broken
One iteration rebuilt a 5.21 GB MP4 with the correct 3840×2160 HEVC video, PCM audio and 195.195-second runtime. It still would not play. FFmpeg reported invalid NAL unit sizes. The diagnostic bundle showed what the build found and how it mapped the recording, proving the container was right while the video samples inside it were still assembled wrong.
Version 2.0: the pieces fit
Each short build changed one part of the reconstruction, then the same recording was tested again. Tracing Sony's frame tables, chunk boundaries and HEVC sample layout finally exposed where the bytes belonged. That working method became the HEVC recovery path used by the 2.x generation.
“The clips that you found work in Resolve!”
Jason's failed builds helped turn HEVC into a working recovery.
During development, each failed build gave us a narrower technical failure to isolate. The 5.21 GB container, correct duration and invalid NAL-size errors showed that finding the recording was not enough; the sample layout still had to be reconstructed correctly.
That repeated test-and-correct cycle is what produced the working HEVC path now used by Sony XAVC Recovery 2.x.