Infer transcription progress from timestamps

[stekhn]

We have the current timestamp and the total audio duration, allowing us to calculate the progress and provide the user with a rough visual estimate of how much time remains for the transcription to complete.