Need to convert PDF to Audio MP3

**SuperMike** · December 26th, 2009

My wife wants me to come up with a Linux script or program that can convert a PDF file into an audio MP3 that she can study while in her car or while working out.

Anyone got a Perl, Bash, or other script that can do this?

**Bonster** · December 26th, 2009

festival or gspeaker

**rennau80** · January 7th, 2010

hi, had the same problem, in the attachement is the file called pdf2mp3.py which does the conversion of a pdf or text/ascii file into a mp3 audio file.

i wanted to hear work-related publications while making sport during lunch brake. as there's only lot's of stupid commercial stuff in the web, i just wrote my own python-script/program that converts a pdf-file (or ascii/text-file) into an mp3 file. it should be very easy to make it work on a running ubuntu distribution
or other linux system, you just need to install the following packages:

Code:

 sudo apt-get install python poppler-utils festival festvox-rablpc16k lame

then you have to use the attached file called pdf2mp3.py and make it an executable via:

Code:

chmod +x pdf2mp3.py

and to be able to run that little program 'system-wide':

Code:

sudo cp pdf2mp3.py /usr/bin/

then you only have to call the script in your shell via:

Code:

pdf2mp3.py yourfilename.pdf

or via:

Code:

pdf2mp3.py yourfilename.txt

or via:

Code:

pdf2mp3.py yourfilename.dat

and an mp3 is being created!

PS: it is the english voice now. other voices can be downloaded here: http://cslu.cse.ogi.edu/tts/download/

**aiiaalla** · January 9th, 2010

i am having a problem........

it says.....

chmod: cannot access `pdf2mp3.py': No such file or directory

please advise..........

on following commands terminal shows...............

nahar@nahar-desktop:~$ sudo apt-get install python poppler-utils festival festvox-rablpc16k lame (command 1)
[sudo] password for nahar:
Reading package lists... Done
Building dependency tree
Reading state information... Done
python is already the newest version.
poppler-utils is already the newest version.
festival is already the newest version.
festvox-rablpc16k is already the newest version.
lame is already the newest version.
The following packages were automatically installed and are no longer required:
linux-headers-2.6.31-14 libnss3-dev libfltk1.1 libnspr4-dev
linux-headers-2.6.31-14-generic
Use 'apt-get autoremove' to remove them.
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
nahar@nahar-desktop:~$ chmod +x pdf2mp3.py (command 2)problem starts here
chmod: cannot access `pdf2mp3.py': No such file or directory
nahar@nahar-desktop:~$

**rennau80** · January 9th, 2010

seems that you downloaded the file pdf2mp3.py to a different folder. if you do a

Code:

ls pdf2mp3.py

it should list the file (i think that for your case it is simply not in your path). otherwise you need to know to which directory you've downloaded the file and do the chmod-command there.

**aiiaalla** · January 10th, 2010

nahar@nahar-desktop:~$ ls pdf2mp3.py
ls: cannot access pdf2mp3.py: No such file or directory
nahar@nahar-desktop:~$

the file is on the desktop........

**varocha** · September 29th, 2012

Server not found can you attach the file again thanks

**frytek** · October 10th, 2012

This one works in a console. Converts txt only but there are many methods of extracting txt from pdf.

I gave up the mbrola voices with espeak because at certain moment I realized they are less clear.

At first they seem to sound nicer and more natural, but later you come to a conclusion that pure espeak is definitely more clear and you can listen faster.

The problem with espeak+mbrola is that there are kind of micro-gaps between pronounced letters / words. The more I listened to it the less I liked it.

Here's the script. Save it in a file and chmod +x it.
The script accepst clear text only. No spaces in the filename.
Usage is:

script file.txt language_code - reads the file

en for English and so on.

or

script file.txt language_code l

encodes with lame. g encodes with gogo (faster. considerable difference if converting an unabridged book).

Code:

#!/bin/sh
# txt2mp3 - convert text files to mp3 audio files (aka audiobooks)
# v0.1
#
# (c) 2011 http://www.gnu.org/copyleft/gpl.html
#
# OBS.: install some pre-requisites first, with
#       sudo apt-get install espeak lame gogo


# espeak parameters
speed=200	#160
pitch=70	#50
paragraph=40	#0
gap=0
lang=$2



if [ "$1" = "" ]; then
    clear
    echo "No arguments. The usage is:"
    echo "$0 file.txt lang_code [g | l] "
    echo "g is for gogo, l for lame encoding."
    echo "Without g or l text will be played only."
    echo "Gogo is more than 2 times faster than lame "
    echo "and with speech encoding there are no "
    echo "substantial quality differences."
    echo "Language code is en, pl, ru and so on."
    echo "Refer to espeak manual."
    echo "Look inside the file to change speed, pitch etc."
    echo "because these are kind of fixed preferences."
    exit

fi

if [ "$2" = "" ]; then
    echo "Not enough arguments. The usage is:"
    echo "$0 file.txt lang_code [g | l]."
    echo "g is for gogo, l for lame encoding."
    echo "Without g or l text will be played only."
    echo "Gogo is more than 2 times faster than lame "
    echo "and with speech encoding there are no "
    echo "substantial quality differences."
    echo "Language code is en, pl, ru and so on."
    echo "Refer to espeak manual."
    echo "Look inside the file to change speed, pitch etc."
    echo "because these are kind of fixed preferences."
    exit


    exit
fi


if [ "$2" = g ]; then
    echo "The second argument should be the language: en, pl and so on."
    echo "The optional 'g' or 'l' goes on third position."
    exit
fi




# encoding
encode="$3"    # If g or l in command line is ommited, the text will be played. 

TXT_FILE="$1"
BASENAME=`echo "$TXT_FILE" | sed 's/\(.*\)\(\....$\)/\1/g'`

echo "Processing ${TXT_FILE} with TTS"


if [ "$encode" = l ] ; then
	espeak -f "$1" -gap $gap -s $speed -l $paragraph -p $pitch -v$lang --stdout | \
	lame --verbose  --preset cbr 32  - "${BASENAME}_espeak-l.mp3"
	echo "...done! Voice saved as ${BASENAME}_espeak-l.mp3"
elif [ "$encode" = g ] ; then
	espeak -f "$1" -gap $gap -s $speed -l $paragraph -p $pitch -v$lang --stdout | \
	gogo stdin -b 32 -m m -emp 5 "${BASENAME}_espeak-g.mp3"
	echo "...done! Voice saved as ${BASENAME}_espeak-g.mp3"
else
	espeak -f "$1" -s $speed -l $paragraph -p $pitch -v$lang
	echo "...done! Add 'g' or 'l'  to command line to encode. "
fi

I also cut the file afterwards into 15 minutes chunks cause my phone much better handles them that way (better scrolling, quicker search for a chapter if necessary).

I use mp3splt with -0 0.2 which makes an overlap of 2 sec. betwen chunks.

**frytek** · October 10th, 2012

I also found some code that does the actual conversion but I don't use it as usually I convert ebooks to txt and to mp3, (not pdf, doc or odt). And and I do it from Callibre.
Anyway, if someone would like to use this, here it goes:

install some pre-requisites first, with
# sudo apt-get install espeak lame xpdf-utils odt2txt antiword

Code:

# if it isn't a TXT file, convert it first
if [ "$ext" != "txt" ] ; then
    TMP_FILE="/tmp/espeakfile-$$.txt"

    # PDF
    if [ "$ext" = "pdf" ] ; then
        echo "converting from PDF to TXT"
        pdftotext "${TXT_FILE}" "${TMP_FILE}"
    fi

    # ODT
    if [ "$ext" = "odt" ] ; then
        echo "converting from ODT to TXT"
        odt2txt --subst=all "${TXT_FILE}" > "${TMP_FILE}"
    fi

    # DOC
    if [ "$ext" = "doc" ] ; then
        echo "converting from DOC to TXT"
        antiword "${TXT_FILE}" > "${TMP_FILE}"
    fi

    TXT_FILE="${TMP_FILE}"
fi