Inspect overlap without treating every overlap as a mistake
The audit reports timing facts and rough text-density flags.
Intervals, not interpretations
Two cues overlap when their time intervals share a positive duration. A cue ending at exactly 3.000 seconds and another beginning at 3.000 seconds do not overlap in this model. A short cue entirely inside a longer cue does overlap, even when they are not adjacent in file order.
An overlap may be intentional, such as two simultaneous voices. The fictional example includes one. The audit counts pairs and marks involved cues; it does not automatically remove or merge them. The timestamp preview shows up to ten simultaneously active stored texts and tells you if more are active.
File order is a separate check
Cues can be out of start-time order without overlapping, or overlap while remaining ordered. Sorting is an explicit action. Exports use the current file order; SRT numbering is regenerated sequentially. Filtering the cue list does not change the project, audit or exported file.
Stored characters per second
The text-density measure counts Unicode code points in stored cue text, including spaces and markup characters but excluding line breaks, then divides by cue duration in seconds. You choose the flag threshold. This is a rough editing signal, not a broadcast standard, reading-comprehension measure or accessibility certification.
The preview renders stored text literally. It does not apply subtitle markup, positioning, line layout, video styling or player-specific rules. Review the exported file with your actual media and target player before publishing it.