I'm not a programmer, but I'd like to make a program that will open and
read a txt file and output to a mp3 file. I don't need to ever hear the
voice, but I'd like the program to direct
I've been google'ing around and have found a few tutorials about
converting pdfs to mp3 and converting typed text (in the source file)
to wave, but I'm looking for the ability to directly convert from txt
to mp3/ogg/wma.
Currently I'm running on WinXP, but I also have access to a Linux box.
Any help would be appreciated.
I have written a script to convert .pdf to .wma, which may get you
started:
#! /usr/bin/env python
# -*- coding: iso-latin-1 -*-
import os
import sys
basename, suffix = os.path.splitext(sys.argv[1])
suffix = suffix.lower()
pdfname = basename + ".pdf"
txtname = basename + ".txt"
wavname = basename + ".wav"
# pdf --> txt
if suffix == ".pdf":
os.system('pdftotext "%s"' % (pdfname, ))
suffix = ".txt"
# txt --> wav
if suffix == ".txt":
import re
data = open(txtname).read()
data = data.replace(". . .", "")
data = data.replace("<", "")
data = data.replace(">", "")
data = data.replace("Mr.", "Mister")
data = data.replace("«", '"')
data = data.replace("»", '"')
data = data.replace(',,', '"')
data = re.sub("\\d+\\s* ", "", data)
data = re.sub("\n\\s*\\d+\\.*\\s*\n", "\n", data)
data = re.sub("\n\\s*\n\\s*\n", "\n\n", data)
data = re.sub("-([a-z])", "\\1", data)
open(txtname, "w").write(data)
import pyTTS
tts = pyTTS.Create()
tts.SetRate(0)
tts.Volume = 100
tts.SetVoiceByName("ATT-DT-14-Klara16")
tts.SpeakToWave(wavname, data)
suffix = ".wav"
# split, wav --> mp3/ogg/wma
if suffix == ".wav":
import wave
os.mkdir(basename)
fi = wave.open(wavname, "r")
nchannels, sampwidth, framerate, nframes, comptype, compname = fi.getparams()
n = 0
while 1:
n += 1
data = fi.readframes(600 * framerate)
if not data: break
fo = wave.open("tmp.wav", "w")
fo.setparams((nchannels, sampwidth, framerate, 0, comptype, compname))
fo.writeframes(data)
fo.close()
# os.system('lame -m m --cbr -b 32 -q 0 -S tmp.wav "%s\\%02d.mp3" "%s"' % (basename, n))
# os.system('oggenc -q -1 -o "%s\\%02d.ogg" tmp.wav' % (basename, n))
os.system('cscript WMCmd.vbs -profile a20_1 -input tmp.wav -output "%s\\%02d.wma' % (basename, n))
fi.close()
os.remove("tmp.wav")
os.remove(wavname) ### This is just too big to keep around
