Hi Tilman, I manually installed the two missing fonts on centos, AdobeKaitiStd, and STSong then the conversion worked. Thanks for your help.
In retrospect, it seems on centos PDFbox couldn't correctly use fallback Chinese font for adobekaitistd. 发自我的 iPhone 在 2017年8月19日，下午8:28，YE ... <[email protected]<mailto:[email protected]>> 写道： Hi Tilman, I am running the conversion on Centos, which doesn't have the two fonts installed. I have installed google's cjk fonts and in most cases PDFbox shall automatically choose the right ones for rendering Chinese characters. I will find a way to install ArialUnicodeMS and MicrosoftYaHei on centos to see if it works. Many thanks, Fangqiao 发自我的 iPhone 在 2017年8月19日，下午5:50，Tilman Hausherr <[email protected]<mailto:[email protected]>> 写道： Hello Fangqiao, I am able to render that file with PDFBox 2.0.7. You can see it at http://imgur.com/a/UOfRl In the log I get this: Warning [PDCIDFontType0] Using fallback ArialUnicodeMS for CID-keyed font AdobeKaitiStd-Regular Warning [PDCIDFontType0] Using fallback ArialUnicodeMS for CID-keyed font AdobeKaitiStd-Regular Warning [PDCIDFontType0] Using fallback ArialUnicodeMS for CID-keyed font AdobeKaitiStd-Regular Warning [PDCIDFontType0] Using fallback ArialUnicodeMS for CID-keyed font AdobeKaitiStd-Regular Warning [PDCIDFontType0] Using fallback ArialUnicodeMS for CID-keyed font AdobeKaitiStd-Regular Warning [PDCIDFontType0] Using fallback ArialUnicodeMS for CID-keyed font AdobeSongStdAdobeKaitiStd-Light Warning [PDCIDFontType0] Using fallback MicrosoftYaHei for CID-keyed font STSong-Light Your invoice does not have its fonts embedded. The messages indicate that PDFBox has chosen to use the fonts ArialUnicodeMS and MicrosoftYaHei to display. Either you don't have these fonts installed, or maybe you used an older PDFBox version? Tilman Am 19.08.2017 um 09:19 schrieb YE ...: Hi Tilman, Thanks for the quick reply. I will check for commercial solutions with font hinting you mentioned here. I have also included the links to the attachments mentioned in my previous email in case you want to take a closer look. PDF file: https://shujubiji.cn/uppv/bjPhoto/d5418a966daecda62bcf056ddc1e79c99a4c6546/1503126757317/chinese_invoice.pdf?zid=__itemtoken__fe4bd85ea612752d11b4fdb02ff43c8871f9bc1c You need to download it then use a PDF reader to open it so Chinese characters can be shown correctly. Converted image: https://shujubiji.cn/uppv/bjPhoto/d5418a966daecda62bcf056ddc1e79c99a4c6546/1503126770469/chinese_invoice_1.jpg?zid=__itemtoken__6de80b84edb217bb37a19218bcec1eb78a24bcfc Screenshot of the originally PDF displayed in PDF reader correctly: https://shujubiji.cn/uppv/bjPhoto/d5418a966daecda62bcf056ddc1e79c99a4c6546/1503126791380/screenshot_from_2017_08_17_17_12_03.png?zid=__itemtoken__a7c4267ccd9597273a7d0286cb3353b1541a9ee1 Best regards, Fangqiao ________________________________ From: Tilman Hausherr <[email protected]<mailto:[email protected]>> Sent: Friday, August 18, 2017 4:18 PM To: [email protected]<mailto:[email protected]> Subject: Re: PDFbox unable to render Chinese font correctly when converting pdf to images Hello Fangqiao, Your files didn't get through, you must upload them to a sharehoster. But I suspect that this is a known problem with chinese fonts, the cause is explained here: https://issues.apache.org/jira/browse/PDFBOX-3293 [PDFBOX-3293] Chinese font glyphs with overlapping paths ...<https://issues.apache.org/jira/browse/PDFBOX-3293> issues.apache.org<http://issues.apache.org> Font glyphs with overlapping paths may be rendered in correctly, especially when the font size is small. Sadly, the Traditional Chinese edition of Windows bundled ... How to fix it - by implementing font hinting. Which we haven't done. There is no workaround, sadly. (Except of course use better fonts when creating the PDF). There are some commercial java products (google for them). At least two of them have implemented font hinting (the others I don't know). Sorry for not having better news. Tilman Am 18.08.2017 um 11:56 schrieb YE ...: Hi, I am from China and using PDFBox to convert pdf files to images. It worked excellently in most cases. Thanks a lot for the team's great work. However recently I used it to convert some invoices in PDF to images and then some Chinese characters weren't converted correctly. Attached is a sample PDF file, converted image and a screen shot of the original PDF opened in PDF reader, which displayed all Chinese correctly. I am seeking help from the community: - what's the possible cause for the problem? I guess that in the original pdf file some Chinese characters' font wasn't set correctly. - how to fix it? If the above guess is correct, is there a way to detect correct font type and set the correct font for conversion? - or is there other solution that can fix the problem? Many thanks, Fangqiao --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected]<mailto:[email protected]> For additional commands, e-mail: [email protected]<mailto:[email protected]> --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected]<mailto:[email protected]> For additional commands, e-mail: [email protected]<mailto:[email protected]>
