I sometimes find it difficult to allocate time to read (edit: academic) papers. However, I sometimes have long commutes so have plenty of time to listen to audio books. That said, is there any accurate (key word) software that converts PDF text to audio? Is there such software tailored to various academic publication formats?
Edit: Since I'll probably use this while commuting, Android solutions are preferable, or at least Mac OS X converters with output that can be downloaded to Android.
Preview in Mac OS X can speak selected text or a whole PDF document. To save the speech to a file, you can copy the PDF text into a plain text file, edit out any erroneous bits, and use say -f plain_text_input_file.txt -o audio_output_file.aiff in Terminal.
Tables, equations, non-dictionary acronyms, Greek letters, and other such "oddities" are not read very well, though the tool deals well with regular sentences. Perhaps there is a more science-orientated text to speech program available, but I haven't heard of it.
pdftotext on my Linux box is a bit out of order. Grabbing the LaTeX source from arXiv and running detex is another option. Every solution I can think of will still need a little manual tweaking for each paper to remove page numbers, running headers, etc, while getting even uglier when equations and symbols get involved.
say seems to work fairly well for my text to speech. The only issues are conversion (AIFF to MP3) and formatting the .txt output for say. Copy-paste can grab some irrelevant footers, and LaTeX isn't always available (or is hard to find). I was hoping for a more unified, less manual solution to save substantial time, but for now, this seems like the best possible option. Thanks!
afconvert is a command-line interface to Core Audio to compress to AAC or HE-AAC, though the documentation (afconvert --help) is a little obtuse. You could install ffmpeg with Homebrew to convert to MP3. There are also plenty of GUI converters out there (iTunes included) that can do the job for you too, if command line tinkering isn't your thing.
On Android, there's e.g. Moon+ Reader Pro supporting PDF and text-to-speak:
Moon+ Reader Pro: Read aloud, ColorDict, Highlight/Annotate (Source: Google Play; click images for larget variants)
As the screenshots show, it's pretty configurable: not only can you adjust volume, but also speed. A bunch of additional features make it very much fit for your purpose: you can highlight/annotate text (for study material, this is probably a must). With PDF files, these annotations/highlights are directly stored into the .pdf file and additionally in "Moonies" database for faster access (with other formats, it's only stored in the database) – so you can later process your annotations/highlights on a different computer. Furthermore, it supports different dictionaries like Fora or (my recommendation) ColorDict – both capable of online and offline dictionaries and using standard formats (ColorDict even offers to install some of them from within the app, while for Fora you'd need to do that manually).
For Text-to-Speech, Moon+ relies on the TTS engine provided by the system. Depending on the language used (and the quality of speech support for it), quality varies; for English it seems to be pretty good, for German I wouldn't use it to read me a novel (but that's a general thing: TTS cannot detect "moods" of the text it reads). Words are pronounced accurately.
Though not specifically tailored to academic text, I think it's a good match (especially considering you might want to use it "on the road", where a tablet/smartphone is preferable to a laptop) – and currently offered for 50% off at Google Play. I'm a long-time user of Moon+ and really love it. Support is great, and the app very reliable.
There's a free (ad supported) version available for a first look, but that doesn't support PDF nor TTS, unfortunately.
PS: As an afterthought. I rarely use PDF (mostly only to proof-read print variants of my books sent to me by my publisher), and almost never TTS – so I cannot say how well these two features work together (though the app description explicitely states: PDF support, fast & speech compatible, so it should work or Seany wouldn't state so). Furthermore, I doubt it will work with e.g. PDFs only composed of images, as Moon+ to my knowledge doesn't contain an OCR engine to extract text from images.
You can use Adobe Acrobat Pro
The Read Out Loud feature reads aloud the text in a PDF, including the text in comments and alternate text descriptions for images and fillable fields. In tagged PDFs, content is read in the order in which it appears in the document’s logical structure tree. In untagged documents, the reading order is inferred, unless a reading order has been specified in the Reading preferences.
"Paper to Audio" is a chrome extension specifically for converting academic papers to speech (this includes PDFs). It gets rid of references etc. when speaking the text.
The link is: https://chrome.google.com/webstore/detail/paper-to-audio/djncfliejhhejjgbhcopflpnlaeicnco?authuser=1
It may have a few bugs which I'm still trying to sort out. Hope this helps.
My app Speech Central can handle this use case well. In some cases it is free (on Android and iOS if you read one book in a month or for blind people). It is also available for Windows and Mac as a single fairly affordable payment.
It should handle some specifics of academic PDFs that most apps can’t like skipping headers and footers or skipping quotations. On top of that it can skip even footnotes, which no other app can.
pdftotextgenerally does a good job in the first part. There are plenty of TTS softwares but I don't use them...Is there a better pdf to text converter than pdftotext?askubuntu.com/questions/52040/… (not only linux)