How to translate the text version of PDF and what is the difference between it and the scanned version?

Key points of this article

PDFs that can use the mouse to select text already have text layers. Usually, there is no need to perform text recognition first when translating, and can be uploaded directly.
If Chinese characters cannot be selected and the whole page is a picture, it is a scanned copy. The text needs to be recognized and can be processed directly when uploading. To edit the original text, OCR is required first.
For mixed files, you need to first determine which pages are pictures, split them if necessary, process them separately, and then merge them for verification.
After translation, focus on checking the number of pages, missing paragraphs, positions of formula tables, proper nouns and numbers, and negative words. Machine translation cannot be exempted from inspection.

Many people search for "free online pdf translation", but the PDFs in their hands are different: some can be used to drag and select text with the mouse, and some are just pictures. The processing paths of the two types of documents are different. If you choose the wrong one, the result may be misalignment, missing pages, or even mistranslation of key figures. This article is written for readers who need to translate papers, contracts, and instructions. It explains how to translate the text version of PDF (that is, the PDF that can select text), and what is the difference between it and the scanned version.

After reading it, you will be able to judge which category the document in your hand belongs to, know what to confirm before translating, and how to check item by item after getting the translation.

How to translate text version PDF? One sentence conclusion

The text version of the PDF can be uploaded for translation directly without having to do text recognition first; as long as you confirm that the text can be selected, upload it on omnidoc.top according to the normal process, select bilingual comparison or translation only, and download the result. Scans can also be uploaded directly. The difference mainly lies in the source of the text and the common causes of errors.

your situation	Recommended practices	reason
You can use the mouse to select and copy text, such as exported papers and instructions downloaded from the official website.	Upload translation directly, select bilingual comparison first for easy checking	The text layer already exists and does not need to be identified. The original typesetting can be retained.
Only a small area can be selected, or the copy is garbled.	First find the page that cannot be selected and determine whether it is a scanned page or there is a problem with the font.	The garbled description text layer is defective, and direct translation is prone to errors.
The whole page is full of pictures, no text can be selected.	Directly upload the scanned document for translation; when you want to modify the original text, first use a PDF editor to do OCR	The text in the scanned document needs to be recognized, and the quality of recognition directly affects the translation.
The translation still needs to be submitted or published	Check first, then select only the translation as needed, export it to Word and adjust it if necessary.	Only the translation layout is cleaner, but it also relies more on pre-checking
Academic documents containing a large number of formulas and tables	After uploading, focus on checking the page where the formulas and tables are located.	The layout, tables and formulas will be retained as much as possible, but complex pages still require manual confirmation.

Before you start: Confirm these points first

1. Whether text can be selected. Open the PDF, drag and select a section of text with the mouse, and then paste it into Notepad. If you can get smooth text, it's the text version; if you can't select anything, it's basically the scanned version. This step determines whether to do more identification later.

2. Whether there are picture pages mixed in the entire document. Many documents are a mix of scanned documents and electronic text. For example, the main text is electronic, but the appendices are scanned and inserted. Quickly check by turning the page to find the page number where the Chinese characters cannot be selected. You cannot make conclusions about mixed documents just by looking at the first page.

3. File size and number of pages. large files with large pages take longer to upload and process, and are more difficult to check page by page. You can first use the split function of the PDF editor to cut into chapters, translate separately, translate them and then merge them.

4. The purpose of translation. * to read for yourself, it is convenient to choose bilingual comparison, the original text and translation can be compared paragraph by paragraph; to give it to others or publish, usually choose only the translation, but check it more carefully. At the same time, pay attention to the type of document, academic, legal, medical, technical, business, literature and other style options will affect the use of words.

5. If there is any sensitive information in the file. * contracts, medical records and undisclosed research materials should be uploaded carefully. Check the privacy instructions of the server before uploading, and upload only the necessary pages.

# steps for translating a text version of PDF in OmniDoc

Open omnidoc.top and enter the document translation function.

two. Upload the PDF file to be translated, make sure that the text version is selected, and do not misrepresent other versions.

If you choose the source language and the target language, the wrong source language will directly affect the quality of the translation.
Choose a style that is consistent with the type of document, such as academic essays, contract law, and optional techniques for specifications.
Choose the output mode: bilingual control is easy to check paragraph by paragraph, and only the translation is suitable for direct reading or submission.
Submit the task and wait for the processing to be completed. Do not submit the same document repeatedly during processing.
Download the translation while keeping the original file for later comparison and rework.
Check it item by item according to the list below, and then modify it if you find a problem.

# check item by item after completion

After you get the translation, don't just open it and read the first page. Here are a few things that are easy to go wrong.

Check item	What do you think?	What if there's a problem?
The number of pages is the same as the original.	Compare the number of home page, last page and total page	Check if there are blank pages or sections in the original text when there is a missing page, and upload it again if necessary
There is no missing translation paragraph.	Align paragraph by paragraph under bilingual control, spot check title, footnote, header and footer	The missing paragraphs are copied separately from the original text and made up by manual or re-translation.
The formula and table are in the correct position	Turn to the page with formulas and tables to see if the rows and symbols are complete.	Adjust according to the original text in case of dislocation, or go back to the source file and retranslate it.
Proper nouns are consistent	Several translations of searching for the same term	Unified replacement and, if necessary, retranslation of the style of the corresponding field
Pictures and annotations are retained	Check whether the picture exists and whether the annotations have been translated	If it is missing, go back to the original PDF to confirm it and supplement it with a PDF editor.
Numbers, dates, negative words	Item by item, compare the amount, time limit, dose, and the words "no", "not" and "prohibited"	These contents must be checked one by one. There are errors in machine translation. Please check them by someone who knows the profession before using them.

# Core comparison: processing differences between text version PDF and scanned parts

Contrast item	Text version of PDF	Scanned PDF
Text source	There is already a text layer in the file	Each page is an image and the text needs to be recognized.
Do you need to identify	Usually not required.	When translating, you can upload it directly; when editing the original text, you need to do OCR first.
Can you select, copy, and search	Sure	Usually not.
Common causes of errors	Garbled text layer, abnormal font, multi-column dislocation	Picture blur, page tilt, seal occlusion, recognition deviation
Post-translation revision	Paragraph text can be changed directly with PDF editor	OCR first and then modify it, and recheck it after revision.
Check the key points	Whether the layout and formula are complete	Whether the identified text is accurate

# frequently asked questions (FAQ)

1. Online pdf translation is a free tool, what to watch when you choose?

Check the file size and page limit, whether the translation is watermarked, the privacy description after upload, typesetting and whether the formula is retained. Don't just look at the word "free". It is recommended to try to translate a short document with a few pages before deciding whether to deal with the whole material.

2. If the text can be selected, why is the format still messy after translation?

The normal text layer does not mean that the layout is simple. Pages with many columns, tables and formulas are easy to be misplaced in the translation. When checking these pages, you can first look at the bilingual version, if necessary, use translation only, or manually adjust after exporting Word.

3. Can the text version and the scanned piece be translated together in a PDF?

Yes, but first find out which pages are pictures. Mixed documents are recommended to be taken apart by type, dealt with separately, and finally merged and checked, so that it is easier to find problems than the whole hard translation.

4. Can the results of online translation be submitted directly?

Direct submission is not recommended. There will be errors in machine translation, and legal provisions, medical doses and technical terms especially need to be reviewed. Please ask someone who knows the profession to read the key paragraphs again before using them.

# Summary

The text version of PDF can select the text, and you can upload it directly without additional identification during translation. More consideration should be given to the recognition quality of scanned pieces. No matter which kind of document, when translating the text version of PDF online, you should first judge the file type, choose the right output mode, and then check the number of pages, omissions, formulas, proper nouns and key numbers according to the list.

Open OmniDoc, upload the document and start translation.