Remove SDH
Remove whole-line sound descriptions, uppercase speaker labels or music symbols from SRT and VTT. Choose categories and compare the automatically updated result.
The result updates automatically when you change these options.
Subtitle Preview
The converted preview is not available yet.
Choose a subtitle file to start previewing.
Choose a subtitle file to start previewing.
No subtitle result yet
The result appears automatically when your file and encoding are ready. Changing cleaning options updates the result.
How to remove SDH
- Choose your subtitle. Select or drop one SRT or VTT file. If asked, check the decoded text and confirm its encoding.
- Choose cleaning categories. Sound descriptions and music symbols are enabled by default. The result is generated automatically and updates when you change an option.
- Download or preview. Save the result using the download above the preview. Output stays in the input format. You can also compare Original and Converted, and review any repairs, skipped entries or format changes.
Format notes
Sound descriptions are removed only when a whole line is enclosed by one matching pair of square brackets or parentheses. Nested or multiple bracket groups, incomplete brackets and descriptions within dialogue are kept. Speaker cleaning recognizes uppercase English labels at the start of a line; names may include digits, spaces, dots, hyphens and apostrophes, and must contain an English letter. The colon must be followed by whitespace and dialogue. Dialogue dashes are retained. Music cleaning removes only ♪, ♫, ♬ and ♩, keeping lyrics. The order is music symbols, speaker labels, sound descriptions, then empty lines and entries. With all categories off, no SDH text cleaning runs. Output keeps the input format, remaining formatting, timing and valid VTT metadata, with UTF-8 and no BOM.
Frequently asked questions
Which categories are on by default?
Sound descriptions and music symbols are on. Speaker labels are off because removing names can affect dialogue. All three options are independent. The first result is generated automatically when the file and encoding are ready. Changing an option updates the result automatically.
Will dialogue in parentheses be deleted?
A whole line such as (sighs) can be removed when sound descriptions are enabled. Dialogue such as Hello (quietly) is kept. Complex lines such as [voice (quiet)] or [door closes] [sighs] are kept, along with incomplete or mismatched brackets.
Which speaker labels are recognized?
Labels such as JOHN: Hello or DR. JOHN: Hello can be removed. A label must contain uppercase A–Z letters, use the allowed name characters, and have whitespace and dialogue after its colon. Mixed-case names, Chinese prefixes, timestamps and colons inside dialogue are kept. VTT voice annotations are preserved.
Are lyrics deleted when I remove music symbols?
No. Only ♪, ♫, ♬ and ♩ are removed. Lyrics and other text stay. When any category is enabled, empty lines and entries left without visible text are removed; the result shows separate counts for descriptions, labels, symbols and empty entries.
Why is downloading disabled after I change options?
The result updates automatically after an option changes. Downloading is unavailable until the current result is ready; the previous preview remains visible and is marked as a previous result. Each update starts from the original input, so changes do not accumulate.
Can I clean ASS or SSA directly?
This tool accepts SRT and VTT and keeps that format. Convert ASS or SSA to SRT first using ASS to SRT, then choose the resulting file here.
Is my subtitle uploaded?
Your subtitle file is read and processed in your browser. Its content and filename are not uploaded to a server. The page still requests the resources needed to run the website.