Function alass_core::align

pub fn align(
    reference: &[TimeSpan], 
    list: &[TimeSpan], 
    split_penalty: f64, 
    speed_optimization: Option<f64>, 
    score_fn: impl Fn(TimeDelta, TimeDelta) -> f64 + Copy, 
    progress_handler: impl ProgressHandler
) -> (Vec<TimeDelta>, f64)

Matches an incorrect subtitle list to a reference subtitle list.

Returns the delta for every time span in list.

The split_penalty_normalized is a value between 0 and 1000. Providing 0 will make the algorithm indifferent of splitting lines (resulting in MANY different deltas), so this is not recommended. Providing 1000 will assure that no split will occur, so only one/the best offset is applied to ALL lines. The most common useful values are in the 4 to 20 range (optimum 7+-1).

Especially for larger subtitles (e.g. 1 hour in millisecond resolution and 1000 subtitle lines) this process might take some seconds. To provide user feedback one can pass a ProgressHandler to this function.

If you want to increase the speed of the alignment process, you can use the speed_optimization parameter. This value can be between 0 and +inf, altough after 10 the accuracy will have greatly degraded. It is recommended to supply a value around 3.

Use standard_scoring as score function if no fine tuning is required.