Reply to post

SSML audio tag support

Author
Post
Essentials Only Full Version
david.dupont
User
  • Total Posts : 20
  • Reward points: 0
  • Joined: 2015/04/10 03:37:04
  • Status: offline
2015/06/03 09:39:52 (permalink)

SSML audio tag support

Hi

I have just find after a research about SSML for silent break (thanks for this one) the audio tag .
<audio src="welcome.wav"> Hello </audio>
 
I use a Cereproc 3.1 voice but 3.2 voice (still not French) should support it.
If the welcome.wav is not found, Hello is use for TTS.
after my test i only manage to have the Hello
 
I have not test it with Google or Samsung TTS for now.
And I'm not sure where to put the file.wav
 
For those of we, that can't have this tag, is this possible to use it in yours softwares to play a .wav ?
if it found, and the TTS if not, That minimise the impact of older version.
 
 
source:
http://www.w3.org/TR/speech-synthesis/#S3.3.1
Quote  #1

6 Replies Related Threads
    Admin
    Administrator
    • Total Posts : 275
    • Reward points: 0
    • Joined: 2010/11/22 00:00:00
    • Location: USA
    • Status: offline
    Re: SSML audio tag support 2015/06/03 15:16:08 (permalink)
    Hi,
    to generate silence, it's not necessary to use a wav file - SSML has a specific command for silence, where you can just provide the number of milliseconds.
    However you could a wav file too for this purpose, or maybe adorn your text file with per-recorded sound effects... The example you give:
     
    <audio src="welcome.wav"> Hello </audio>
     
    is not very good, because the TTS voice software does not know where on your memory card or in the system the welcome.wav file is. It would be better to give the full path to that file, e.g. something like:
     
    <audio src="/sdcard/MySounds/welcome.wav" />
     
    I'm really not an expert on SSML, did not use it much yet.
     
    Greg
    Quote  #2
    david.dupont
    User
    • Total Posts : 20
    • Reward points: 0
    • Joined: 2015/04/10 03:37:04
    • Status: offline
    Re: SSML audio tag support 2015/06/03 16:22:39 (permalink)
    Hi Greg.
     
    I was not clear, I don't use "audio" to have silent. I saw another post on this forum about "silent".
    I didn't know about SSML at this time, and then I made a research
     
    The example come from the link from my last post. Your right that a full path is better but that not what I think is more important. It can be a path directory on the option menu.
    Where I find it's clever, is that if the wav is not allowable then use the TTS on the text, but if the wav is played, skip the text.
     
    I send a email to Cereproc, and had a answer in less than 1 hour (impressive).
    Audio tag don't work for android version, but he send me a pdf about specific tag for Cereproc TTS.
    I will make some test tomorrow.
     
    Do you think, you could add this directly in your software ?
    A user can use the rule, for "***" replace "<audio src="/sdcard/MySounds/myWav.wav" />"
    or replace "Chapter" by "<audio src="/sdcard/MySounds/Chapter.wav" />Chapter "
     
    Thanks
     
    Quote  #3
    Admin
    Administrator
    • Total Posts : 275
    • Reward points: 0
    • Joined: 2010/11/22 00:00:00
    • Location: USA
    • Status: offline
    Re: SSML audio tag support 2015/06/03 17:35:36 (permalink)
    If you have a TTS voice for which <audio...> tags work, you could do it right now, with the Edit Speech feature of @Voice app. I tried this right now with Ivona, Acapela, Google, Samsung and eSpeak voices. None supports the <audio> tag - some of them just ignore it, others read aloud all the characters of the command.
     
    If you would like my app to play a specific audio file, it is doable, but not in the middle of a sentence. I have to send complete sentences to TTS engine. This would also require modification of the app code, so it could not be done with the current version. Realistically it could be a rule like this:
     
    IF the sentence contains [this_substring] or [matches_this_RegEx_expression]
    then PLAY [this_sound_file] before OR after reading aloud this sentence.
     
    Greg
    Quote  #4
    david.dupont
    User
    • Total Posts : 20
    • Reward points: 0
    • Joined: 2015/04/10 03:37:04
    • Status: offline
    Re: SSML audio tag support 2015/06/03 17:41:36 (permalink)
    I don't have any TTS voice wih support of <audio> :(
     
    Read a sound in the middle of a sentence is not what i'm looking for.
    Before or after, will do the trick perfectly.
     
    Read a book will be about  20% cooler
    Quote  #5
    thiago.eec
    User
    • Total Posts : 3
    • Reward points: 0
    • Joined: 2015/06/16 14:32:25
    • Status: offline
    Re: SSML audio tag support 2015/06/16 15:00:07 (permalink)
    Hi, Greg.
     
    If you are using it for ePub3 books, you could use this css code for the chapter headings:
    h1 {
     -epub-pause: 50ms 25ms;
     -epub-cue: url('audio/ping.mp3') none;
     -epub-rest: 10ms 0ms
    }

    This would set the audio file (ping.mp3) to be played everytime right before the new chapter starts.
    Offcourse this only works if the Reading System has support for ePub3 (including the properties used).
    post edited by thiago.eec - 2015/06/16 15:04:53
    Quote  #6
    Admin
    Administrator
    • Total Posts : 275
    • Reward points: 0
    • Joined: 2010/11/22 00:00:00
    • Location: USA
    • Status: offline
    Re: SSML audio tag support 2015/06/17 15:06:21 (permalink)
    Thank you, Thiago! The "reading system" in this case is my app, @Voice, so I would have to implement these commands. I guess I'll have to take a closer look at the ePub3 standard and implement what's necessary.
     
    Greg
    Quote  #7
    Jump to:
    © 2024 APG vNext Commercial Version 5.1