Michal Rogowski

When generating transcription on device, sometimes final string do not contain whole string for example in trump_clintoon.mov but only on-device

Still need to think of correct design to present results, prototype used here may be much improved. . This is not a mock but working result

VideoProcessorService inside `startDetecting()` function generates participants array only for the first frame where any face was detected, so if someone will appear during a video it would not be noticed...

AVAssetExportSession crashes on exportAsynchronously ```swift return Future { promise in exportSession.exportAsynchronously { promise(.success(exportSession.status)) } } ``` When startTracking() in `VideoProcessorService` finish