gnu/packages/speech.scm

   1 ;;; GNU Guix --- Functional package management for GNU
   2 ;;; Copyright © 2016 David Thompson <davet@gnu.org>
   3 ;;; Copyright © 2016, 2019 Marius Bakke <mbakke@fastmail.com>
   4 ;;; Copyright © 2017 Leo Famulari <leo@famulari.name>
   5 ;;; Copyright © 2018 Tobias Geerinckx-Rice <me@tobias.gr>
   6 ;;; Copyright © 2016 Kei Kebreau <kkebreau@posteo.net>
   7 ;;;
   8 ;;; This file is part of GNU Guix.
   9 ;;;
  10 ;;; GNU Guix is free software; you can redistribute it and/or modify it
  11 ;;; under the terms of the GNU General Public License as published by
  12 ;;; the Free Software Foundation; either version 3 of the License, or (at
  13 ;;; your option) any later version.
  14 ;;;
  15 ;;; GNU Guix is distributed in the hope that it will be useful, but
  16 ;;; WITHOUT ANY WARRANTY; without even the implied warranty of
  17 ;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18 ;;; GNU General Public License for more details.
  19 ;;;
  20 ;;; You should have received a copy of the GNU General Public License
  21 ;;; along with GNU Guix.  If not, see <http://www.gnu.org/licenses/>.
  22
  23 (define-module (gnu packages speech)
  24   #:use-module ((guix licenses) #:prefix license:)
  25   #:use-module (guix packages)
  26   #:use-module (guix download)
  27   #:use-module (guix utils)
  28   #:use-module (guix build-system gnu)
  29   #:use-module (gnu packages)
  30   #:use-module (gnu packages audio)
  31   #:use-module (gnu packages autotools)
  32   #:use-module (gnu packages compression)
  33   #:use-module (gnu packages gcc)
  34   #:use-module (gnu packages glib)
  35   #:use-module (gnu packages linux)
  36   #:use-module (gnu packages pkg-config)
  37   #:use-module (gnu packages pulseaudio)
  38   #:use-module (gnu packages python)
  39   #:use-module (gnu packages textutils))
  40
  41 (define-public espeak
  42   (package
  43     (name "espeak")
  44     (version "1.48.04")
  45     (source (origin
  46               (method url-fetch)
  47               (uri (string-append "mirror://sourceforge/espeak/espeak/"
  48                                   "espeak-" (version-major+minor version)
  49                                   "/espeak-" version "-source.zip"))
  50               (sha256
  51                (base32
  52                 "0n86gwh9pw0jqqpdz7mxggllfr8k0r7pc67ayy7w5z6z79kig6mz"))
  53               (modules '((guix build utils)))
  54               (snippet
  55                ;; remove prebuilt binaries
  56                '(begin
  57                   (delete-file-recursively "linux_32bit")
  58                   #t))))
  59     (build-system gnu-build-system)
  60     (arguments
  61      `(#:make-flags (list (string-append "PREFIX=" (assoc-ref %outputs "out"))
  62                           (string-append "DATADIR="
  63                                          (assoc-ref %outputs "out")
  64                                          "/share/espeak-data")
  65                           (string-append "LDFLAGS=-Wl,-rpath="
  66                                          (assoc-ref %outputs "out")
  67                                          "/lib")
  68                           "AUDIO=pulseaudio")
  69        #:tests? #f ; no check target
  70        #:phases
  71        (modify-phases %standard-phases
  72          (replace 'configure
  73            (lambda _
  74              (chdir "src")
  75              ;; We use version 19 of the PortAudio library, so we must copy the
  76              ;; corresponding file to be sure that espeak compiles correctly.
  77              (copy-file "portaudio19.h" "portaudio.h")
  78              (substitute* "Makefile"
  79                (("/bin/ln") "ln"))
  80              #t)))))
  81        (inputs
  82         `(("portaudio" ,portaudio)
  83           ("pulseaudio" ,pulseaudio)))
  84        (native-inputs `(("unzip" ,unzip)))
  85        (home-page "http://espeak.sourceforge.net/")
  86        (synopsis "Software speech synthesizer")
  87        (description "eSpeak is a software speech synthesizer for English and
  88 other languages.  eSpeak uses a \"formant synthesis\" method.  This allows many
  89 languages to be provided in a small size.  The speech is clear, and can be used
  90 at high speeds, but is not as natural or smooth as larger synthesizers which are
  91 based on human speech recordings.")
  92        (license license:gpl3+)))
  93
  94 (define-public espeak-ng
  95   (package
  96     (name "espeak-ng")
  97     (version "1.49.2")
  98     (home-page "https://github.com/espeak-ng/espeak-ng")
  99     (source (origin
 100               (method url-fetch)
 101               (uri (string-append home-page "/releases/download/" version
 102                                   "/espeak-ng-" version ".tar.gz"))
 103               (sha256
 104                (base32 "1d10x9rbvqi2zwcz65fxh04k0x0scnk7732l37laz6xra1ldhzng"))))
 105     (build-system gnu-build-system)
 106     (arguments
 107      `(#:configure-flags '("--disable-static")
 108        ;; Building in parallel triggers a race condition in 1.49.2.
 109        #:parallel-build? #f
 110        ;; XXX: Some tests require an audio device.
 111        #:tests? #f))
 112     (inputs
 113      `(("libcap" ,libcap)
 114        ("pcaudiolib" ,pcaudiolib)))
 115     (synopsis "Software speech synthesizer")
 116     (description
 117      "eSpeak NG is a software speech synthesizer for more than 100 languages.
 118 It is based on the eSpeak engine and supports spectral and Klatt formant
 119 synthesis, and the ability to use MBROLA voices.")
 120     (license license:gpl3+)))
 121
 122 (define-public mitlm
 123   (package
 124     (name "mitlm")
 125     (version "0.4.2")
 126     (source (origin
 127               (method url-fetch)
 128               (uri (string-append "https://github.com/mitlm/mitlm/releases/"
 129                                   "download/v" version "/"
 130                                   name "-" version ".tar.xz"))
 131               (sha256
 132                (base32
 133                 "09fv4fcpmw9g1j0zml0k5kk1lgjw2spr8gn51llbkaaph6v8d62a"))))
 134     (build-system gnu-build-system)
 135     (native-inputs
 136      `(("gfortran" ,gfortran)))
 137     (synopsis "The MIT Language Modeling toolkit")
 138     (description "The MIT Language Modeling (MITLM) toolkit is a set of
 139 tools designed for the efficient estimation of statistical n-gram language
 140 models involving iterative parameter estimation.  It achieves much of its
 141 efficiency through the use of a compact vector representation of n-grams.")
 142     (home-page "https://github.com/mitlm/mitlm")
 143     (license license:expat)))
 144
 145 (define-public speech-dispatcher
 146   (package
 147     (name "speech-dispatcher")
 148     (version "0.9.1")
 149     (source (origin
 150               (method url-fetch)
 151               (uri (string-append "https://github.com/brailcom/speechd/releases"
 152                                   "/download/" version "/speech-dispatcher-"
 153                                   version ".tar.gz"))
 154               (sha256
 155                (base32
 156                 "16bg52hnkrsrs7kgbzanb34b9zb6fqxwj0a9bmsxmj1skkil1h1p"))))
 157     (build-system gnu-build-system)
 158     (arguments
 159      `(#:configure-flags '("--disable-static"
 160
 161                            ;; Disable support for proprietary TTS engines.
 162                            "--with-ibmtts=no"
 163                            "--with-kali=no" "--with-baratinoo=no")))
 164     (native-inputs
 165      `(("intltool" ,intltool)
 166        ("pkg-config" ,pkg-config)))
 167     (inputs
 168      `(("dotconf" ,dotconf)
 169        ("espeak" ,espeak-ng)
 170        ("glib" ,glib)
 171        ("libltdl" ,libltdl)
 172        ("libsndfile" ,libsndfile)
 173        ("pulseaudio" ,pulseaudio)
 174        ("python" ,python)))
 175     (synopsis "Common interface to speech synthesizers")
 176     (description "The Speech Dispatcher project provides a high-level
 177 device independent layer for access to speech synthesis through a simple,
 178 stable and well documented interface.")
 179     (home-page "https://devel.freebsoft.org/speechd")
 180     ;; The software is distributed under GPL2+, but includes a number
 181     ;; of files covered by other licenses.  Note: in practice, this
 182     ;; is linked against dotconf, which is LGPL 2.1 only.
 183     (license (list license:gpl2+
 184                    license:fdl1.2+ ; Most files in doc/ are dual gpl2+/fdl1.2+.
 185                    license:lgpl2.1+
 186                    (license:non-copyleft
 187                     ;; festival_client.{c,h} carries an expat-style license.
 188                     "See src/modules/festival_client.c in the distribution.")
 189                    license:gpl3+)))) ; doc/texinfo.tex -- with TeX exception.
 190
 191 (define-public sonic
 192   (package
 193     (name "sonic")
 194     (version "0.2.0")
 195     (source (origin
 196              (method url-fetch)
 197              (uri (string-append "https://github.com/waywardgeek/sonic/archive/"
 198                                  "release-" version ".tar.gz"))
 199              (file-name (string-append name "-" version ".tar.gz"))
 200              (sha256
 201               (base32
 202                "11a0q9wkgbb9ymf52v7dvybfhj8hprgr67zs1xcng143fvjpr0n7"))))
 203     (build-system gnu-build-system)
 204     (arguments
 205       `(#:tests? #f ; No test suite.
 206         #:make-flags
 207          (list (string-append "DESTDIR=" (assoc-ref %outputs "out")))
 208         #:phases
 209         (modify-phases %standard-phases
 210           (delete 'configure)))) ; No ./configure script.
 211     (synopsis "Speed up or slow down speech")
 212     (description "Sonic implements a simple algorithm for speeding up or slowing
 213 down speech.  However, it's optimized for speed ups of over 2X, unlike previous
 214 algorithms for changing speech rate.  Sonic is a C library designed to be easily
 215 integrated into streaming voice applications such as text-to-speech (TTS) back
 216 ends.
 217
 218 The primary motivation behind Sonic is to enable the blind and visually impaired
 219 to improve their productivity with speech engines, like eSpeak.  Sonic can also
 220 be used by the sighted.")
 221     (home-page "https://github.com/waywardgeek/sonic")
 222     (license license:asl2.0)))