OmniDocHow to preserve the format in document reading and translation
Misplaced tables, garbled formulas, and missing pictures are common problems after translation of foreign language papers. This article explains how to choose the file type and output method, and provides a step-by-step verification method after translation to help you retain the original text format.
Key points of this article
- The digital version of the paper can be uploaded directly to PDF, and the scanned copy can also be uploaded directly without having to process it separately first.
- When you have a LaTeX source code package in hand, the source code will be uploaded first, and the formulas and typesetting instructions will be retained more completely.
- Use bilingual comparison for reading and verification, modify the translation or export and edit only the translation, and then use it with the PDF editor
- Check tables, formulas, pictures and terminology item by item after translation. The machine-translated content must be based on the original text.
People who are studying for graduate school or doing research projects often encounter a trouble: after the PDF of a foreign language paper has been translated, the formatting is completely messed up, the table is broken into several sections, the formulas become garbled, and the pictures cannot be found. At this time, the efficiency of proofreading paragraph by paragraph is very low, and it is easy to miss misplaced content. How to set the retention format for document reading can actually be broken down into several specific issues to deal with: first determine the file type, then choose the right output method, and finally check for error-prone areas.
This article is written for readers who need to read foreign literature and want to retain the original format. After reading, you will know what file format to upload, which output method to choose, where to check after the translation is completed, and what to do when you encounter a scanned PDF.
How to preserve the format when translating documents? One-sentence conclusion: Upload the original file, choose a translation method that supports format preservation, and use bilingual comparison to check tables, formulas, and pictures.
To choose according to the situation at hand, you can refer to the table below:
| your situation | Recommended practices | reason |
|---|---|---|
| Digital PDF download from publisher or database | Upload PDF directly and select bilingual comparison | Tables, formulas, and pictures are processed together with the document, and the original text and translation are arranged side by side for easy verification. |
| The paper is a scanned copy and the text cannot be selected. | Upload scanned PDF directly | The scanned document is within the supported range, so there is no need to perform separate text recognition first. |
| There is a LaTeX source code package provided by the author | Upload source code package for translation | Formulas and formatting instructions remain in the source code, resulting in results closer to the original |
| Need to revise the translation or write it into your own manuscript | Translate first, then use the PDF editor to change the paragraph text, or export it to Word and edit it later | The translation can be revised paragraph by paragraph without having to type type from scratch. |
| I just want to understand the main idea quickly | Translation only | The reading page is cleaner, but you need to go back to the original text when comparing the charts. |
Before you start: Confirm these points first
- The file is not a digital version. Try selecting a text with the mouse in the PDF. If it can be selected, it means it is a digital version; if it cannot be selected, it is usually a scan or image PDF. Both types of files can be uploaded, but the clarity of the scan will affect the translation effect, so open it first to see if the original is clear.
- Whether the source code is available. The LaTeX source code at the time of submission or the source code package provided by the preprint platform can make the restoration of formulas and formats more reliable. It doesn't matter if you don't have the source code. You can just use PDF translation. It just takes more time to check the formula.
- Whether the paper is a complete version or an excerpt. If appendices, footnotes or references are omitted, the corresponding content will not be found when checking later. Make sure the number of pages and chapters are complete before uploading.
- Use bilingual comparison or only translation. When reading a paper for the first time, it is recommended to read it bilingually to facilitate comparison of the original diagrams and professional descriptions. When you have already understood the text and just want to compile it into a translation, then consider just the translation.
- Thesis area and style. The terminology and sentence patterns of texts in science, engineering, medicine, law, etc. vary greatly. Choose a style that is consistent with the paper, such as an academic style, and the wording of the translation will be closer to the way the paper is written.
Steps to translate an entire paper in OmniDoc and preserve formatting
- Prepare documents. Find the complete PDF of the paper, or the LaTeX source code package provided by the author, and check whether the number of pages and appendices are complete.
- Open the translation portal. Work on omnidoc.top, or select document translation in the desktop client. The web version is suitable for temporary use, and the desktop client can process files locally and use functions such as a PDF editor.
- Upload files. PDF is uploaded directly; if there is a LaTeX source package, upload the source package instead of just uploading a certain file in the source code.
- Choose a style. For essay-type texts, choose an academic style to make the terminology and sentence structure more consistent with the way the essay is written.
- Select the output method. Select bilingual comparison or translation only according to the previous judgment.
- Start translation and wait for completion. Longer papers will take longer, just wait patiently and do not submit the same file again midway.
- Download the results and open. Read through the bilingual version first, and then check the checklist in the next section one by one.
- Modify the translation (optional). When you need to change some sentences, you can directly modify the paragraph text in the PDF editor, and then save or export after modification.
Check each item after completion
| Check items | How to see | What to do if there is a problem |
|---|---|---|
| sheet | Compare the original table row by row to see if the number of rows and columns, merged cells and numbers are consistent | If the numbers are misplaced, return to the original text for confirmation, and export to Word for modification if necessary. |
| official | Enlarge the page to check superscripts, subscripts, fractions, Greek letters and numbers | If the formula is garbled, use the LaTeX source code to re-translate it, or find the formula in the original text and correct it manually. |
| Pictures and annotations | See if the picture is in the original position and whether the note has been translated. | When the picture is missing, make sure that the complete file is uploaded, and then re-translate it. |
| Professional terminology | Whether the translation of the same concept is unified in the full text | Search for key terms in the translation and change them into the same translation method. |
| Key data and conclusions | Compare sentence by sentence with the original text, especially the experimental results and concluding paragraphs | In the event of inconsistencies, the original text shall prevail, and ask the mentor or people in the field to confirm |
There are inevitable deviations in machine translation, especially complex long sentences and professional abbreviations. If you quote it into your own paper, you should check it sentence by sentence according to the original text.
# Core comparison
| Output mode | Suitable for the scene | Need to pay attention to |
|---|---|---|
| Bilingual | Read the paper for the first time, check it paragraph by paragraph, check the chart. | The layout is denser and may need to be scaled when reading. |
| Translation only | Read through the main idea, organize your notes, and need to revise the translation. | When you look at the chart, you should go back to the original text, but the check is not intuitive enough. |
# frequently asked questions (FAQ)
1. Can I keep the form of the scanned paper?
OmniDoc supports translation of scanned PDF and retains forms and images. However, the clarity of the original directly affects the result, and when the lines of the table are blurred or tilted, it is necessary to focus on checking the numbers after translation.
2. What if the formula becomes garbled after translation?
First go back to the original PDF to confirm whether the formula itself is clear. If there is a LaTeX source package, it is usually safer to retranslate it with the source code. When the code is still garbled, the formula can only be found in the original text and corrected manually.
3. Paste the paper directly into the general chat robot, can you keep the format?
Usually not. Pasted is plain text, table structure, formula typesetting and picture location will be lost, and need to be retypesetting. This kind of practice is more suitable for translating a paragraph than for the whole paper.
4. Can the translation be directly used for contribution?
Direct use is not recommended. The translation can be used as a reading and first draft, and should be checked sentence by sentence against the original text before submission, and revised by the author or mentor to ensure that the terminology and data are accurate.
5. Will bilingual control make reading slower?
Bilingual control will make the page more content, so you really need to take a look at it when reading paragraph by paragraph. However, mistranslation can be found in time compared with the original text, so it is recommended to use it in the first reading. After you are familiar with the content, it will save more time to switch to translation only.
# Summary
How to set the document reading retention format, the key lies in three things: upload the original file or source code, choose the appropriate output mode, and check the tables, formulas and pictures according to the list after translation. As long as you do a good job in checking, you can not only read the literature, but also retain the original format.
