Generate subtitles from audio

This is the most straight forward operation. Using the audio analysis data acquired in the previous step EZTitles automatically generates subtitles and synchronizes them with the audio.

Align loaded text to audio

Splits the loaded transcript into subtitles and automatically times (synchronizes) them with the audio.

Useful if you already have a Dialogue List, Spotting List or any kind of transcript. Just import the text into EZTitles and use this option. The only restriction for now is that the audio and the script must be in the same language.

Extract transcription for subtitles

Consider this option if you already have well-synchronized empty subtitles and need to fill their text automatically. Subtitling Assistant will not modify their timing but will replace their content with the spoken text recognized from the audio.

Generate timecode for subtitles

This option can be used if you already have well-arranged subtitles without timecodes and you need just to synchronize them automatically. Subtitling Assistant will not change the text in this case and will only set or modify the timing.

Whole File

The whole file will be processed at once. Processing the whole file requires some time to complete, usually few minutes if the video is particularly long.

From current position to video end

Starts the Assistant from the current video position.

If "Align loaded text to audio" is selected the align process will start  from the current video position and currently selected subtitle. Make sure to switch to Edit mode first, select the subtitle you want to start with then choose video position from the Timeline somewhere before the selected subtitle's text and then run the Assistant.

If the align algorithm finds significant differences between the loaded text and the audio, the whole process will be stopped. If this happens, we suggest correcting the subtitles for the mismatched part and then skip forward to the part where the video and the script are in sync and continue with the Assistant by either using From current position to video end or From current position to speech segment options.

Note that if the loaded transcript is timecoded with acceptable precision you can use "Use In- and Out-cues as sync points" option described below to re-sync the algorithm.

From current position to speech segment end

The speech segment is best described as a small part of the video with speech containing at least 50 words ending up with at least 2 sec pause in the speech.

It works for small parts of the video and almost instantly generates subtitles for them. Then you can quickly review the subtitles, navigate to any position in the video after the last subtitle and run the Assistant for the next speech segment.

It can be really handy when tuning various Subtitle Preferences and Rules. Specify the parameters and rules, press Go button and if you are no satisfied with the result use Undo and then adjust the parameters again.

Similarly to the previous option, please make sure to select the subtitle you want to start with and after that choose video position from the Timeline prior running the Assistant.

End of speech - out-cue gap

Specifies the number of frames automatically added to the Out-cue of the last subtitle of a dialogue.

Typically, the subtitles composing the dialogue are chained on the minimum interval and the last subtitle in the chain ends up several frames after the audio. The option is set to 12 frames by default but changing the value to 0 frames will switch it off completely.

CC Style

Specifies the style (Pop-on, Roll-up or Paint-on) of the automatically generated captions.

Use In- and Out-cues as sync points

Available only when Align loaded text to audio is used.

If there are significant differences between the loaded text and the recognized audio and the algorithm can't match them it stops. It usually happens if large amount of text is missing in the transcript or there is extra text not present in the audio. If this option is turned on and the loaded transcript is adequately timed the In- and Out-cues are used as synchronization points so the algorithm can continue. As always, the subtitles with possible issues are marked with comments.

Note that this will work properly only if the subtitles doesn't contain inconsistent cues i.e. In-cue before the previous Out-cue, In-cue before the previous subtitle's In-cue, etc.

Use original subtitles as text breaks

Available only when Align loaded text to audio is used.

Use this option if you have subtitles containing large amount of text in each of them and these blocks of text are logically separated. Subtitling Assistant will create subtitles in such a way that there are no subtitles containing text from two different text blocks (subtitles) in the original script/text.

Non-speech information in script

Provides options to remove character names, sound effects, on-screen text and any other additional descriptive information not included in the spoken dialogue.

More details about NSI and examples is available later in this topic.

Maximum number of rows

The Maximum Number of Rows is currently limited to 1 or 2 and it determines how many text lines subtitles can have.

Row doesn't end on

Allows word to be entered in order to prevent subtitle lines ending with them, for example "a", "an", "the", "that", etc.
Pressing the List button will open new interface where the words and phrases can be entered for each language:

Move to a new line text enclosed by

This option could be really handy when the subtitles contain descriptive text indicated by a special character like the musical note or surrounded by brackets or any other symbol. Text enclosed by the specified symbols will be moved to a separate line, if Max. Number of Rows is set to 2, or to a new subtitle otherwise.

If the text is surrounded by two different symbols it is important to enter both of them in the respective field.

Subtitle end punctuation

The specified punctuation characters will be used as split point and when found near the end of a text line or subtitle all the text that follows will be inserted into a new subtitle.

Prefer longer rows

This rule instructs the algorithm to favor creating subtitles that have more characters on each text line. Decreasing its weight will result in shorter, containing less text, subtitles overall.

Single sentences

Tells the algorithm to prioritize creating subtitles that contain just a sentence in them. Increasing its weight will result in subtitles primarily containing just a single sentence, while decreasing its priority will result into more subtitles containing text from different sentences.

Row end punctuation

The specified punctuation characters will be used as word wrap point and when found near the end of the line all the text that follows will be moved to a new text line.

Subtitle rows layout

Allows you to select preferable layout for subtitles containing more than 1 text lines:

Bottom Heavy: the bottom line will be longer than the one above.

Top Heavy: the first line will be longer than the second one.

Squared: aims to create two text lines with equal or close to equal length.

Note: the line will be split when it goes outside the safe area or contains more than the allowed number of characters in Project Settings/Safe Area.

Preferred max row length

Specifies the number of characters that each text line should preferably contain. This will instruct the algorithm to prefer text lines with this length but it doesn't mean that there will be no lines longer than that. The total length will never exceed the Max. Number of Characters set at Project Settings/Safe Area menu.

Use 0 to switch this option off.