notes

Log | Files | Refs

tts.txt (1219B)

      1 Install Piper on archlinux:
      2 
      3 yay -S piper-tts-bin
      4 
      5 Install voices:
      6 
      7 yay -S piper-voices-en-us
      8 yay -S piper-voices-en-gb
      9 yay -S piper-voices-zh-cn
     10 
     11 How to use it to play audio directly:
     12 
     13 cat test.txt | piper-tts \
     14         --model /usr/share/piper-voices/en/en_US/bryce/medium/en_US-bryce-medium.onnx \
     15         --output-raw | aplay -r 22050 -f S16_LE -t raw -
     16 
     17 ====================================================================
     18 
     19 Speech Dispatcher
     20 
     21 /etc/speech-dispatcher/modules/piper-tts-generic.conf
     22 
     23 Here you set up the default voice for reading PDFs which can
     24 integrate with Okular or Calibre
     25 
     26 Restart:
     27 
     28 systemctl --user status speech-dispatcher
     29 
     30 Test saying sth:
     31 
     32 spd-say "This is a test"
     33 
     34 Read a txt file:
     35 
     36 cat transcript.txt | spd-say -e
     37 
     38 ====================================================================
     39 Kokoro
     40 
     41 Spin up a dockerized kokoro: https://github.com/hwdsl2/docker-kokoro/
     42 
     43 and then try it:
     44 
     45 jq -n --rawfile t ~/transcript.txt \
     46      '{model:"tts-1", input:$t, voice:"af_heart", response_format:"wav"}' \
     47    | curl -s http://127.0.0.1:8880/v1/audio/speech \
     48        -H "Content-Type: application/json" \
     49        -d @- \
     50        --output /tmp/speech.wav \
     51    && ffplay -nodisp -autoexit /tmp/speech.wav