☰

分解文件，清理语法

tag-icon • tag-icon text-to-speech
分解文件，清理语法

pico2wave
希望将文本文件输入 pico2wave。

像这样...

pico2wave 'read a file'

我的整个代码

sudo apt update &&

dependencies=(calibre sox libttspico-utils)
for i in ${dependencies[@]}; do sudo apt install $i -y; done;

Name="Why-Did-John-Nash-Stop-His-Medication"
Address="https://www.researchgate.net/profile/Peter-Weiden/publication/7757848_Why_Did_John_Nash_Stop_His_Medication/links/5dff8196299bf10bc370696a/Why-Did-John-Nash-Stop-His-Medication.pdf"

wget --header="User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_0) AppleWebKit/600.1.17 (KHTML, like Gecko) Version/8.0 Safari/600.1.17" \
    $Address

ebook-convert $Name.pdf \
              $Name.txt

cat $Name.txt

pico2wave --lang=en-GB --wave=$Name.wav <  $Name.txt

play $Name.wav

答案1

漏洞：2015 年（8 年前的漏洞）

在此处输入图片描述

Bug #1468771

https://bugs.launchpad.net/ubuntu/+source/svox/+bug/1468771]

在此处输入图片描述

答案2

分解文件，清理语法

sudo apt update &&

dependencies=(calibre sox libttspico-utils)
for i in ${dependencies[@]}; do sudo apt install $i -y; done;


Name="Why-Did-John-Nash-Stop-His-Medication"

Address="https://www.researchgate.net/profile/Peter-Weiden/publication/7757848_Why_Did_John_Nash_Stop_His_Medication/links/5dff8196299bf10bc370696a/Why-Did-John-Nash-Stop-His-Medication.pdf"


wget --header="User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_0) AppleWebKit/600.1.17 (KHTML, like Gecko) Version/8.0 Safari/600.1.17" \
    $Address

ebook-convert $Name.pdf \
              $Name.txt


while IFS= read -r Line;
 do

  clear

  echo "$Line" > file.tmp;

  aspell -c file.tmp;

  diction --quiet file.tmp;

  String=$(cat file.tmp)

  clear

  echo "$String"

  pico2wave --lang=en-GB --wave=temp.wav "$String";

  aplay --quiet temp.wav

  clear

 done < "$Name.txt"

答案3

您可以使用下面的代码，通过 pico2wave 高效地将任意长度的文本转换为音频；该脚本将输入文本拆分为最大长度为 3500 个单词的片段，然后将每个片段转换为音频，最后将片段连接到最终的 audio.mp3 文件中。尽情享受吧！！（注意：您应该遵循 pyrubberband lib 的要求，特别是如果您想通过拉伸、音高变换等方式调整输出时……）

    import os
    import subprocess
    from pydub import AudioSegment
    import soundfile as sf
    import pyrubberband as pyrb
    import subprocess
    # Not the requirements of pyrb (sudo apt install libttspico-utils sox)
    # pip install numpy Audiosegment pysox sox sndfile

    def process_text_segment(text_segment, output_folder, segment_index):
        # Use pico2wave for audio generation
        wav_output_path = os.path.join(output_folder, f"audio{segment_index}.wav")
        wav_output_path1 = os.path.join(output_folder, f"audio{segment_index}.wav")
        subprocess.run(['pico2wave', '-w', wav_output_path, '-l', 'en-GB', text_segment])

        # Use pydub for audio processing
        audio = AudioSegment.from_wav(wav_output_path)
        duration = audio.duration_seconds

        # Additional processing steps if needed
        y, sr = sf.read(wav_output_path)
        # Play back at extra low speed
        y_stretch = pyrb.time_stretch(y, sr, 0.73)
        sf.write(wav_output_path1, y_stretch, sr, format='wav')
        # Play back extra low tones
        y, sr = sf.read(wav_output_path1)
        y_shift = pyrb.pitch_shift(y, sr, 1.5)
        sf.write(wav_output_path1, y_shift, sr, format='wav')
        sound = AudioSegment.from_wav(wav_output_path1)
        sound = sound + 12

        # Convert the processed audio to MP3
        mp3_output_path = os.path.join(output_folder, f"audio{segment_index}.mp3")
        sound.export(mp3_output_path, format="mp3")

        # Clean up temporary WAV files
        os.remove(wav_output_path)
        os.remove(wav_output_path1)

        return mp3_output_path

    def concatenate_audio_segments(output_folder, segment_count, mp3_output_path):
        # Concatenate audio segments
        segments = [AudioSegment.from_mp3(os.path.join(output_folder, f"audio{i}.mp3")) for i in range(1, segment_count + 1)]
        final_audio = sum(segments)

        # Export the final concatenated audio to MP3
        final_audio.export(mp3_output_path, format="mp3")

        # Clean up intermediate audio segments
    for i in range(1, segment_count + 1):
            os.remove(os.path.join(output_folder, f"audio{i}.mp3"))

    def process_large_text(input_file_path, output_folder):
        with open(input_file_path, "r") as file:
            text_content = file.read()

        # Split text content into segments of maximum 3500 words
        word_limit = 3500
        text_segments = [text_content[i:i+word_limit] for i in range(0, len(text_content), word_limit)]

        segment_count = len(text_segments)

        for i, text_segment in enumerate(text_segments, start=1):
            process_text_segment(text_segment, output_folder, i)

        # Concatenate audio segments
        mp3_output_path = os.path.join(output_folder, "audio.mp3")
        concatenate_audio_segments(output_folder, segment_count, mp3_output_path)

        print(f"Audio generated and saved to {mp3_output_path}")

    # Example usage:
    input_file_path = "path/to/your/input/file.txt"
    output_folder = "path/to/your/output/folder"
    process_large_text(input_file_path, output_folder)

相关内容