Transcription guide · 4 min read
Verbatim vs Clean Transcription: Chinese Examples
Choose what to keep in a Chinese transcript. Compare verbatim and clean transcription with examples, then review original and filler-free versions.
Published
Use a verbatim approach when the way someone speaks matters: hesitations, repeated words and false starts may be part of what you want to examine. Use clean transcription when you need a readable account of the conversation, while preserving the speaker’s meaning and wording that carries information.
For Chinese recordings, the practical question is often whether a word is just a hesitation or part of the message. A repeated 那个 can fill a pause; 那个方案 means ‘that proposal’ and needs to stay. The examples below show how to make that decision before you export an interview or meeting transcript.
Choose the level of detail your reader needs
A full verbatim convention may record false starts, repetitions and nonverbal details such as pauses or laughter. Clean verbatim usually removes speech clutter while keeping the substance of the conversation. Set the rules for your task before editing, especially when several people will review the same recording.
In Hanzi Now, the controls are named Original and Filler-free. Original is the version generated without the filler-removal option; Filler-free enables that option. Use these as drafts to review against the recording. For a transcript that needs detailed pause or laughter notation, add those details during your listening pass.
| What you need | Starting point | What to check |
|---|---|---|
| Study hesitation or self-correction | Original | Keep repetitions and add the details your task requires |
| Prepare a readable meeting transcript | Filler-free | Keep decisions, qualifications, names and amounts |
| Select a quote from an interview | Original, checked against audio | Verify the exact passage and its surrounding context |
| Prepare a Mandarin reading handout | Filler-free with reading aids | Keep meaningful particles and conversational wording |
Remove hesitation without changing the message
These are invented editing examples, not before-and-after outputs from a recording. Read the whole sentence before removing a word: the same expression can do different jobs in different contexts.
| Spoken example | Possible reading copy | Reason |
|---|---|---|
| 嗯,我们明天见。 | 我们明天见。 | The opening 嗯 is a hesitation in this example |
| 我们,我们周五再确认。 | 我们周五再确认。 | Remove an accidental repeated start |
| 我想,那个,先问一下。 | 我想先问一下。 | 那个 fills a pause here |
| 那个方案还要再讨论。 | 那个方案还要再讨论。 | 那个 identifies which proposal |
| 我觉得可能要三天吧。 | 我觉得可能要三天吧。 | The wording conveys uncertainty about the estimate |
Keep uncertainty, emphasis and meaningful particles
Cleaning a transcript is an editing choice, so check what each deletion does. Changing 我觉得可能要三天吧 to 要三天 turns a tentative estimate into a firmer statement. Keep words such as 可能 when they express uncertainty, even if the shorter sentence looks tidier.
Repetition can also be deliberate. 很重要,很重要 may be emphasis rather than a restart. Listen for what the speaker is doing before shortening it. Likewise, 然后 can connect events in a sequence; removing it everywhere can make that sequence harder to follow.
Keep conversational wording when it carries the speaker’s voice. A clean transcript still records the conversation. Turning replies into formal prose or compressing several answers into a conclusion is a separate writing task.
Compare Original and Filler-free in the editor
When uploading a recording, choose Remove filler words if you want the cleaned version first. Leave it off to start with the original. Once the transcript is ready, use the Filler-free switch to select the other version.
If the other version has not been generated, the editor shows a confirmation with the credits it will use. Generate it while the recording is still available and your remaining allowance covers the task. Once both versions exist, switching between them is free, and each version keeps its own edits.
Use the public sample to explore the difference before uploading. Its original and filler-free versions are already prepared. For your own transcript, compare the passages that matter, correct speaker assignments, and replay the audio before deciding which wording to keep.
Finish the wording before adding reading aids and exporting
Check the chosen version for names, amounts and missing negatives. Keep timestamps if you want a reader to locate the passage in the audio. Name speakers consistently so that a clear sentence stays attached to the person who said it.
Then review pinyin and English translation if your reader needs them. Select Export to download the current version as Word or PDF. Choose PDF to include pinyin, or include English in either format. The result should match the version and the level of detail you chose at the start.
Key takeaways
- Choose a transcription convention based on what the reader needs to examine.
- Decide whether a Chinese word is a filler from the sentence and the recording.
- Preserve uncertainty, emphasis and words that identify people, objects or events.
- Review the selected version before exporting; generating a second version uses credits.
Put it into practice
Turn a Chinese recording into a transcript
Follow the audio, edit each speaker’s words, and read with pinyin and English.
Open Chinese audio to text →Frequently asked questions
Does clean verbatim transcription mean writing a summary?
No. A clean transcript follows the conversation while removing the speech clutter you have chosen to omit. A summary selects and condenses the main ideas, which is a separate task.
Should every 嗯, 那个 or 然后 be deleted?
No. Listen to the context. 嗯 can be an answer, 那个 can mean ‘that one’, and 然后 can connect events. Keep the expression when it contributes to the message.
Will edits in one version replace the other?
Each generated version keeps its own edits. The editor saves changes before switching, so you can work on an original transcript and a filler-free reading copy separately.